I am absolutely flummoxed by this review. Mine is configured with nearly identical specs (except that I have 64GB RAM) and is one of the worst purchases I've ever made.
Enabling deep sleep works fine; waking from it is impossible. WiFi stops working. The touch pad becomes hypersensitive. I've resorted to turning the whole thing completely off after every use. But the battery still loses power even when the machine is completely powered down.
The machine also freezes if I try to open a jpg or a PDF from the file explorer. Just completely freezes. Fan spins up like crazy, but as I mentioned, I have 64GB of RAM in this thing. It should be able to open a single one-page scanned document from my own scanner. If there's anything wrong with the file, this is the computer that's making the file poorly.
The internet connection also just turns off after about half an hour of use. I have a Belkin dock with Ethernet plugged into it, which is then plugged into my Framework's USB-C port. This dock supports my work Mac all work day every day with no issues at all, so it's not the dock's problem. But the Framework? Internet just stops working after half an hour. And it manages to kill both wired and WiFi when it dies. I can turn them back on, but this is a ridiculous problem to have.
This last point is pretty minor, but the hinges are floppy. The whole screen shakes around when I type on it.
I have no idea how this author's experience is so much better than mine unless they just haven't used it much yet, but I regret absolutely everything about the purchase of this machine and feel like I was sold a bill of goods by all the glowing positive reviews all over the internet.
We definitely want to help you resolve these issues. Most of them sound like specific compatibility challenges on Ubuntu 21.10. To be clear, I agree that we need to get better with new Ubuntu releases. We've been providing hardware to the team at Fedora and have full support out of the box with Fedora 35, but don't have a setup like that with Ubuntu yet.
I am running 21.04. All these problems are occurring on 21.04.
In that case, we definitely want to understand what is occurring, because that is pretty unusual. I would recommend either commenting in the 21.04 thread or creating a new thread in our Linux sub-forum on this, so that both us and members of the community can help debug this: https://community.frame.work/t/ubuntu-21-04-on-the-framework...
I will try to do this as soon as I have the time.
But time is the main reason I bought the machine in the first place. Reviews about Ubuntu support were glowing. Linux support was touted in the website significantly.
I'm really not trying to do anything extraordinary with the machine, so I thought it would be better just to spend the money and get one that works than it has been trying to do things like run Lubuntu on my old Chromebook Pixel.
I'm doing basic things like research things in a web browser, scan documents with xsane, or occasionally record audio in Audacity. I'm not trying to do anything extraordinary or complicated.
I just wanted a machine that would do those things without tracking me everywhere (Windows) and where I could fix pieces rather than having to replace the whole device if something broke or wore out.
I paid more than twice as much as I ever have for a computer so that I would NOT have to help you develop the product, do quality control, or troubleshoot it. I wanted it to be ready to use for my run-of-the-mill needs.
And it wasn't, and that's why I wrote this review here.
When faced with the same issue (optimizing for time wrt computer setup), I moved from a thinkpad with Arch Linux to an M1 MacBook + headless Ubuntu server under my desk.
Truly best of both worlds in my experience- incredible battery life and everything ‘just works’ for normal workflows on the MacBook, and can ssh into the headless Ubuntu box directly or from anywhere with internet. Depending on your dev needs, this anecdote may be useful: I’ve exclusively used vim and VS Code with a remote server extension to target the Ubuntu box for builds and I haven’t touched the setup for 6-8 months.
Also, if you’re willing to spend some time learning, getting into Nix ends up saving time in the long run as your nix shell definitions should provide replicable user / build environments on top of whatever Linux flavor you happen to be running at the time, for the headless dev box
> I paid more than twice as much as I ever have for a computer so that I would NOT have to help you develop the product, do quality control, or troubleshoot it. I wanted it to be ready to use for my run-of-the-mill needs.
TBH, then I think you should have bought an Apple or Windows computer, and put up with whatever the downsides of such things are (such as the tracking you mention).
Not saying that the extremely poor performance you describe is your fault or that the manufacturer should be absolved or anything, just that this is precisely the point of paying for conventional computers and putting up with their bullshit (real or perceived).
My threat model does not support Apple or Microsoft products.
I'm glad yours does; that does sound convenient!
Never buy v1 of anything- you basically pay for the privilege of "beta testing"
> time is the main reason
If time is your valuable commodity then buying a first gen "revolutionary" product was never going to be a good move. In your case, buying "known good" was the direction you should have gone and let others cut themselves and bleed out on the edge. Sorry for your bad experience, chalk it up to lessons learned. It's ever the case with tech: the vision is far, far ahead of reality.
If it hadn't been for the glowing reviews of how seamlessly it supported Linux, and the suggestions if not outright claims on the website that it was designed with Linux support in mind, I wouldn't have.
It doesn't sound ridiculous to me that someone might decide to target the Linux market really well even though others haven't. I had hoped that was what had happened. Maybe I wanted to believe it more than I should have believed it.
But that's the motivation behind my posting here, to save anyone from making the same mistake.
I have a $1700 paperweight. Oh well. But that doesn't mean I want anyone else to make the same mistake.
I want Framework to succeed (and, in fact, had strongly considered becoming a customer) so its out of love that I say: it's this reply just as much as the original post that convinces me never to buy a Framework laptop. OP describes numerous major issues with the laptop and calls it one of the worst purchases he has ever made (and it wasn't a small purchase) and you...direct him to an internet forum for help troubleshooting??? If I were the original poster, I'd be furious.
Given all the very positive press out there about Framework products, and the forum where this conversation is happening, I would have expected an answer more along the lines of: "your experience is unacceptable to us. Let us send you a box to return the laptop for an exchange or a full refund."
I realize that this might be more appropriate for a private communication with he customer, rather than a public HN post. Maybe you've done that. I have no way of knowing. But from what I can see, it seems like OP is being told that its on him to figure out how to fix a laptop that he just paid $1,600(?) for and that, from the sound of it, never worked properly.
I get that this product is aimed at folks who aren't afraid to get their hands dirty. But I would not have expected that to mean a customer would be relegated to the forums to address such major issues. The fact that some of your employees may help him do this on the forums is very cold comfort. It is still going to burn up a lot of the customer's time.
Edit -- Just to refine my thoughts here: it's not a bad thing that Framework has an active forum that can help a customer resolve issues like this, nor is it a bad thing to invite a customer to participate there if they want to. On the contrary: that's all to the good. At the same time, though, I'd expect a company like Framework to evince a greater sense of their own ownership of the customer's problems, rather than exclusively suggesting ways that the customer can find a fix for themselves.
I'd counter with: someone complaining on HN that hasn't so much as sent an email to the manufacturer is kind of acting in bad faith. I COMPLETELY agree that a brand new laptop should work out of the box, but it's also kind of on you to let the manufacturer know so they can actually do something about it. Now you could make the argument: but a forum? They should be opening a support ticket!
But that's kind of framework's go to market: an open ecosystem. Which includes troubleshooting in public so other users can see what the problem and solution is. Vs. the magic bullet being locked in some internal KB. It also would do no good to run through troubleshooting here, clogging up a thread on HN trying to solve one users issue.
So - you say: immediately return the laptop! I say: nothing he's described immediately indicates a hardware problem; it mostly looks like software configuration issues. So why on earth would they replace the hardware as step 1?
Fair points. But doesn't that just mean the response should be something along the line of: "Hey, those issues sound very abnormal to us to please reach out to us here [support link] so we can fix it. If you prefer, you might also consider posting on [forum] so others can learn from this. Either way, don't worry. We'll absolutely get it sorted out."
I don't mean to tell them exactly how they should communicate with their customers, but communicating some sense of ownership over the problem would have gone a long way.
Of course, it's also entirely possible that I'm just not the target market for these laptops. I certainly have no problem with fixing my own tech problems as a general matter, but I wouldn't want that to be my only recourse on my daily driver. If that's just not what Framework is selling, then I'm glad they've communicated that!
I would assume that OP would have the option of returning if he so desired, without that needing to be pointed out in a public thread.
One would think. But, as a prospective customer, the posts on this thread give me some information about what kind of response I can expect if I have my own problems with their product. (Especially if they crop up after the close of the return window, etc.)
They have a 30-day guarantee. If OP is having such severe problems, then they should return it and get a refund.
Hey. I don't have a Framework and I'm not in the market soon, so I'm not particularly invested in the underlying specifics here, but...
As a general matter, I have seen this kind of sentiment a lot, and I don't think it's particularly helpful. This subthread OP clearly feels let down and potentially betrayed. Why would you respond by suggesting that the conversation goes private onto a forum Framework controls? Shifting the forum of the discussion to a place where one party has more power seems inherently questionable.
I say that even though I have no reason to believe you all would do anything inappropriate, and my general view of your project is highly positive.
I suspect HN isn't a very good place for a back-and-forth on debugging issues like this. If nothing else, HN has a reply depth limit that will be hit pretty quick.
> by suggesting that the conversation goes private onto a forum Framework controls
AFAIK it will not go private, it will still be available for everyone to follow. HN is not exactly convenient place for user support (unless it's Google or Apple).
I agree with the sentiment, but they cannot provide assistance on HN either, it's not a good fit.
At this point, when a client of mine that is technically litterate has such a bad experience, I pay them a visit.
They would probably learn a ton by having the setup at hand.
really curios if there is ANY overlap of your target audience and fedora users.
i mean, for a top-down approach i would probably also choose redhat because they seem to have structures for a streamlined b2b communication and commitment, but on the other hand i absolutely know nobody who uses it apart from the odd researcher and enterprise-it guys.
> really curios if there is ANY overlap of your target audience and fedora users.
I believe it's the other way around: no other Linux distribution has a higher overlap with the target market of the Framework Laptop than Fedora.
Because unlike Ubuntu, Fedora Workstation has been specifically design for developers, hobbyists, and early adopters.
Personally, I only recommend and install Ubuntu for non-technical people, who don't need or care about always having the most modern OS features.
I always used Ubuntu on my hardware and about 90% Debian derived distros for customers vs a 10% of Red Hat derived ones, some CentOS and some Amazon Linux 2. I really don't like rpm. I do adapt to the different organization of the configuration files but I prefer the Debian one (maybe because of habit.) Most importantly I feel it's easier to google for solutions to Ubuntu problems. Of course I'm also more used to do that, I'm probably at 20 vs 1 Debian/Ubuntu vs RH machines now.
I have to say I agree. I feel like a lot of people seem to think it is a suit and tie enterprise type boring distribution. Its really not. It has similarities to the enterprise distros, but it certainly doesn't feel like RHEL.
Its where the new fun stuff goes first... not so new that it will bite you, but still new enough to be fun to play with. That seems perfectly in line with the marketing I see from framework.
I’m a developer who’s pretty much always used Ubuntu, and haven't bothered digging too deep into other distros. What are the modern features I’m missing?
To put it simply: most things that you are enjoying in Ubuntu now first showed up in Fedora months or even years ago.
The packaging system is what originally brought me to debian and ubuntu; RPM caught up a while back, but apt was better for decades. Stability and easy access to good tooling is what I like about Ubuntu; how is Fedora better?
Fedora is a maintainers dream, in my (limited) opinion.
Things like mock and fedpkg make modifying upstream/bundling your own stuff up super easy. Then add in COPR (like PPAs) and it's a feedback-loop friendly setup
What does Fedora currently have that Ubuntu is missing?
Fedora just gets everything sooner. Newer kernels, gnome, packages. But it's still rock solid because of Red Hat/IBM.
Well, that's just bullshit.
You either get stuff sooner or get stuff more stable, you can't have 'rock solid' stuff sooner.
Depends what you mean by stable... In my experience, recent Fedora releases don't break. They also have the best laptop compatibility. But if 3rd party software (or your own software) targets specific libs, you need to keep up. Ubuntu is typically about one release cycle behind.
Ubuntu feels to me like it’s been feature complete for about ten years and all we really get in the new versions are updated upstream apps and drivers, a stability crapshoot (sometimes better, sometimes worse), and an extension of support. Is Fedora doing anything different?
Fedora is where most of the innovations in Linux happen, because it's an upstream of CentOS Stream, which is an upstream of RHEL.
Manjaro! It gives you the best of Arch without the tinkerers overhead... I've used RHEL, Debian/Ubuntu, Slackware, et al, and Manjaro, for a laptop, is hands-down the best developer distro out there IMHO.
I switched from Arch (btw!) To Fedora last year and it has been an awesome experience. Highly recommended! 10 years of Arch has thought me alot and I have enjoyed i3 and sway, but now I needed an OS that just works.
Nice to hear that Framework plays well with Fedora, it will definitely be the next laptop I buy (if they ever start shipping to Norway).
did you ever try Manjaro?
Yes, some years ago during a distro jump. Went back to Arch pretty fast, like all the other times I went through a distro-curious phase. Fedora 35 has been the only distro where I have not been crawling back to arch within weeks.
Do you think Fedora is an unpopular distro?
no. i imagine the install- and user-base is huge due to the footprint in edu and enterprise.
but ime it is unpopular for personal use esp. in the enthusiast/dev/ops/hacker demographic that i consider to be a sizeable portion of the framework target market.
I feel like the exact opposite of what you're saying is true. Education and enterprise stick to Ubuntu, which is practically the Windows of Linux distributions. Hugely popular and mainstream.
> The machine also freezes if I try to open a jpg or a PDF from the file explorer. Just completely freezes. Fan spins up like crazy, but as I mentioned, I have 64GB of RAM in this thing. It should be able to open a single one-page scanned document from my own scanner. If there's anything wrong with the file, this is the computer that's making the file poorly.
This almost certainly has nothing whatsoever to do with the framework itself. That sounds like a software bug all day long. Something about your OS install is broken or similar.
The only other possibility seems like it'd be if the storage drive is broken. Which could then possibly be why you're experiencing so many other issues, if things like code pages are just failing to load or getting corrupted in the same way your JPG isn't able to load.
But it seems like starting with a fsck or even a re-install seems like a good idea.
> This almost certainly has nothing whatsoever to do with the framework itself.
I mean... Kind of? It's pretty clear that they're driver issues. Hardware manufacturers have to make in-roads with OS teams to get these things delivered flawlessly, otherwise they have to wait for the community to build fixes. System76 and Lenovo do their best to get ahead of driver issues from what I've seen. I'm assuming a lot of this has to do with the fact that OP is running 21.10 rather than an LTS release.
But opening a jpeg or PDF doesn't really involve any drivers, so driver issues wouldn't explain why those specifically are failing to open. If opening anything caused random freezes, that'd suggest something like the GPU driver might be failing. But otherwise opening JPEGs & PDFs are just generic CPU work, there's no drivers for those.
To be clear, they open fine if I open a compatible program and then open them from within the program.
They just freeze the system if I try to open them from the file explorer.
I've not seen anything like this before.
It sounds like it's failing to lookup what program to launch from the file type. That's all but certainly either a corrupt OS install or a bad storage drive then, and wouldn't be indicative of any general problem with Framework laptop's linux support (or lack thereof).
Honestly, that sounds like a kernel panic. Two guesses: bizarre GPU/driver issue triggered by some kind of preview/desktop environment thing (less likely), or your filesystem is corrupt and/or your SSD has trouble (more likely).
Try `pv /dev/nvme0n1 > /dev/null` (or just cat, but pv will give you a nice progress bar if you install it). If that errors out or your machine freezes, you have a storage problem and probably need to replace your SSD. If that works, try booting in single user mode or from a USB boot drive and running fsck.
You can also try SSHing in from another machine while trying to open the file locally. If the SSH session dies instantly, it's probably a kernel panic. If it survives, you can look at logs and top see what happened (maybe the issue is the desktop environment locks up?)
Sleep issues, the dock stuff, those sound like legitimate driver/support issues that would affect anyone. But opening PDFs crashing your machine really sounds like either your hardware or install is specifically broken and needs fixing, or you're running into a really weird bug that is not hardware-specific.
Do you have a separate partition for /home and other for / ?
I do, and this enables me to delete and reinstall any Linux version without changing my home directory.
It seems reinstalling the operating system is the first thing you should try.
Does that work with flatpaks and snaps that store some app data in /home?
It should. The idea is, you never format or delete what is in /home.
You give them access to that partition
try Fedora on it and see if these issues still persist. It can't hurt and may aid in troubleshooting.
I'm with you. My configuration is more-or-less maxed out and at over $2500 it's one of the more regretful large purchases I've ever made. Current issues:
- As has been noted time and time again, battery life is atrocious. As in unusable for even the travel time required for a three hour flight: factoring in the poor sleep, an hour at the airport, and time on the aircraft the battery is dead. Crazy for 2021/2022 - my eight year old Macbook does better.
- Wifi. The Intel Wifi modules provided with the Framework are very poorly supported in Linux. I spent six hours the other night running additional ethernet drops because I just couldn't stand my workday routinely being interrupted by repeatedly disassociating from my AP (fixed with reboot). Even unloading and reloading the associated kernel modules wouldn't bring it back.
- USB-C acts... Strange. I have two 4K displays using Displayport over USB-C and the Framework repeatedly fails to initialize them from boot. I have to do a strange dance (the steps of which I'm still figuring out) involving powering off the Framework, unplugging/re-plugging everything, and then rebooting until it magically works again.
- Fan. The fan is crazy. I've settled on disabling turbo mode in Linux.
- Build quality. I'm convinced if I drop this thing it's all over.
All in all I'm mad at myself for spending over $3k on a productivity configuration centered around what is pretty much beta hardware. I should have known better.
Anyway, there's hope for some of these issues as they seem to be releasing BIOS updates pretty regularly (currently running 3.07) but even then I need to USB boot to use their EFI update tool. Intel wifi should get better over time and worst case I can swap it out for Atheros or something. That said I think the battery, build, and fan issues are fundamental hardware design choices.
- Re battery: Yup, battery life during sleep sucks, it seems like this is mostly an Intel 11th gen issue as S3 sleep went away and folks on Dell, Lenovo, etc., 11th gen hardware have similar issues.
- Re Fan: it does seem to like to ramp up a bit too much even when components are not really that hot yet.
- Re USB-C: have not used that, so cannot comment
- Re build quality: yes, it is a bit flimsier than I would like (though have definitely used laptops that are way worse), but it is also lighter than competitors that are less so. I am mostly okay with that trade-off, others may not be.
As you say, the company has been responsive and helpful on the community forums in a way the big players never have been, and the BIOS updates are fixing genuine issues and complaints with real release notes. It's not all great for sure, but there's lots of things that are a lot better than competitors.
I mean at the end of the day, you’re buying a laptop from framework, if the laptops work does work that’s on them for choosing to use an intel chip with known issues.
Laptops are sold as an appliance for the most part, they have to own it’s quality.
Actually more than that, this should first be an excellent laptop and then the repairability and customisable it should be layered on top of that without compromising it (too much), as opposed to the other way around.
That's not really fair... no one is going to buy a new laptop without the most modern processor. The problem really is that Intel on mobile is awful and has only been getting worse as they try to compete. I guess you could say Framework dropped the ball by not using AMD.
Thank you for clarifying that the Intel wifi issues are a Linux kernel thing, not a Framework thing. I tried to communicate that in my comment but obviously I could have been clearer. I'm fairly confident they'll get fixed in Linux eventually but it's still yet-another-issue that's frustrating for a laptop so heavily embraced by the Linux community.
I understand the sleep issues but battery life is miserable generally - the sleep issues only exacerbate an already pretty-much-terrible situation. Fortunately I don't really need significant battery life but if I did I'd be looking to unload this thing at this point.
How are the wifi issues not Framework's responsibility? They are the ones who defined the hardware specs. I would assume they tested different configurations and picked the best options.
Which makes me wonder - maybe the Intel wifi board has a lot of problems but is still the best Linux option...
Their design aim was to make a laptop that is upgradeable and serviceable first. Not 100% Linux support first. This makes sense because software is much easier to improve after the fact than hardware.
And I'm curious to know which WiFi chipset you would rather they have used, as most alternatives either have worse Linux support or aren't as advanced as Intel's. Would you rather they used Broadcom, The bane of WiFi and Linux?
> which WiFi chipset you would rather they have used
That's what my last sentence was trying to express. Something can be shitty and also the best option.
My bad, I took that to be sarcasm.
I mean, Dell and Lenovo sell laptops with Linux pre-installed with Intel AX* WiFi boards that almost certainly suffer the same issue.
I’m not trying to give Framework (or any other laptop maker) a free pass here, just saying that it’s not a unique problem to this laptop.
I was thinking that there was a good Linux option out there. That may not be true.
One of the benefits that a company like Framework has is that they don't have to support everything. They can be like Apple and tailor their hardware choices to the ones best supported by the software.
To me, Asahi Linux is the most exciting thing going on in Linux-land right now. They have a very small number of hardware configurations to support. In theory, they should be able to get it really dialed in and polished to the point where it just works. I don't care if they go for KDE or Gnome, SystemD or SysVinit, or anything else in particular. If it works right from the start and is extraordinarily stable, I'll learn to love the choices they make.
The issues are the same as always: drivers and suspend. Drivers are made worse by the intersection of newer hardware and lagging distros, suspend lately made worse by Intel’s changes in 11th gen and Microsoft pushing its vision of “modern standby”.
Asahi is more of a project for driver and related development and upstreaming it. They do plan to release a distro, but the current plan is for it to just be Arch with the bleeding edge of their work applied (so perhaps not the “dialed-in” you’re looking for since Arch is far from that). IMO, Asahi is just solving the same issue with anything else: lack of drivers for some hardware. It just so happens that the hardware is very specific and popular. My feeling is that I don’t want to invest in hardware whose manufacturer is unwilling to invest at all in my use-case (Linux).
New throwaway since this is a rant at an entitled software developer community…
Do you want open computing?
I’m really tired of the IT community that just wants high paying jobs but none of the skills that earn them.
As a hardware designer; do better, software “engineers”.
If the community is going to claim the title, it damn well better start engineering better than these awful websites and hacky operating system kernels and GUIs that glitch.
Or stfu over others efforts which are a bit beyond memorizing textbook math and recalling them via your text editor with algorithmic help.
Look at all the non-contributors certain they can do better.
FWIW the workaround in the bug report 100% resolves the WiFi issues for me (it just means losing 802.11ax support, but my AP does not have that anyway). It was definitely the most frustrating issue for me; made game streaming via Moonlight impossible.
Anecdotally, my battery life is plenty good when actually using it (I've yet to feel like I need to reach for the charger at all unless I started very low), but everyone's use case is different.
Yep, I'd actually forgotten about that workaround but I couldn't risk another entire day of lost productivity so I spent my evening running Ethernet. No regrets!
Thanks for the feedback!
Andrew noted in a sibling comment what the issue was with the AX210 in Linux. Intel has informed us that the patch is going into kernel 5.17. In retrospect had we known how many issues we'd see with AX210 relative to AX201, we would have gone with the latter as the default option on the DIY Edition. Going forward though, we do expect compatibility and stability to improve (on 5.17 and later).
On the USB-C issues, it would be great to know the model numbers of the displays and cables that are having this issue. We want to continue to build up our library of peripherals that have unknown issues in order to resolve them.
On build quality, we've seen exceeding few field failures on mechanical issues, and a couple of reports of people dropping their system (or as often happens, blaming their cat for this) from high enough to dent a corner of the chassis and then being able resolve that by ordering a replacement cover part from us.
Hi! Thanks for taking the time to reply to my overwhelmingly negative review. I suppose what I'm saying, generally, is that I should be more selective in my tools. You're a startup. I'm on my third. I should know that when I'm often barely hanging on from the stress and workload of building my own stuff I should select "safe" and "proven" tools. In this case that should have probably been a Macbook (again) because Linux on a laptop is still practically a part-time job itself.
EDIT: Speaking of startups, if you didn't know already you have a lot of "street cred" out there. When wrapping an in-person presentation or meeting many of the questions I get are about my Framework!
Where should I report on the displays and cables with the USB-C issues?
Regarding drops I've actually never dropped a laptop in the > 20 years I've had them. So kind of a non-concern but the comment was referring to the general "feel" which may not really mean anything outside of perception.
Would be interesting to see how much battery life there is on Windows. For ThinkPads it’s always much better on Windows than on Linux.
You have to go to bios and disable secure boot, then the battery doesn’t get drained anymore when the lid is closed. You also have to disable ps2 mouse emulation in bios, the mouse hypersensitivity goes away after that. I also enabled large text in accessibility so the text size is perfect and super high res with the 3:2 aspect screen, it’s fantastic.
I just bought the framework and running ubuntu 21.10. I had some of the problems you mentioned but the forum has solutions to many. It’s a new product, and specially with Linux, I knew there would be some problems. You have to be willing to solve these problems o/w you should stay away from buying new hardware from a startup.
> You have to go to bios
IMO, a computer should work decently without needing to tweak the BIOS. If a feature is "experimental" and is likely to cause problems, it should be disabled by default.
I imagine it does work decently in Windows…it’s pretty standard to have to flip a few setting in the BIOS when installing Linux on a machine that came with Windows, best case is to have clear instructions on what settings need to be changed.
That said this is also the very first model of a device that’s intended to be DIY repairable…tweaking some BIOS settings shouldn’t be scary for someone buying what is effectively the beta version of a laptop that comes with a screwdriver so you can change out its parts.
For what it's worth, on most computers I've daily driven with Linux, tweaking the bios has been the norm. At minimum, I've had to disable secure boot on most distros.
I agree with the idea that it _should_ work decently without - but as of now, that is not the status quo.
Oh, please. I've had to disable Secure Boot on every Thinkpad I've ever had in order to run Linux on it without drama. It takes two seconds.
This is a computer you have to put together yourself (if you want to choose your hardware config). It comes with a screwdriver in the box. The logo is a gear. These are not-so-subtle hints that this machine is for tinkerers.
> disable secure boot, then the battery doesn’t get drained anymore when the lid is closed
Do you have a link to further elaboration and testing on this? Seems strange, and searching the forums is a bit difficult since the two terms get mentioned together a lot.
Secure boot currently results in the Linux kernel disabling hibernation for security reasons IIRC.
That is true, I had (perhaps incorrectly) assumed the previous commenter was talking about it affecting power consumption in “normal” suspend rather than hibernate. (I actually can’t use hibernate anyway because I use ZFS, but that’s a different story.)
Thank you for these directions! I'll give them a shot as soon as I get a chance and hopefully they'll make a difference!
awesome. i would be keen to get a little status report in a week if you could? others will then possibly be convinced to support the development and growth of Framework if they are searching HN through Algolia
Or they'll be convinced to stay away, depending on the outcome...
NB. Hopefully the 'no hibernate with secure boot turned on' is going to get sorted eventually - I believe Matthew Garrett has been working on it.
At the moment, the inevitable battery drain while asleep is quite annoying on modern laptops running Linux :(
They work fine under Windows. Ergo the problem isn't the laptop, its the driver support under Linux.
In my experience current gen laptops except for certain ThinkPads and Clevos will have support issues in Linux. You get a much better experience using a model a year or two old because the community has had time to address any issues.
The design goal of the Framework wasn't out of the box 100% Linux support, it was reparability. While they encourage Linux development on the platform, what hardware works best to meet their design goal may or may not already have Linux support.
An awful lot of the reviews and hype have been touting Linux support, and this review in particular says it "just works."
It doesn't, and I very much wish the company and reviewers had made that clearer.
>It doesn't, and I very much wish the company and reviewers had made that clearer.
The reviewers definitely need to do a better job of thorough testing before making claims. But in my experience it is common for things to inexplicably not work when before they did and vice versa in Linux with no good explanation as to why. It's part of the reason why "the year of the Linux desktop" is always next year IMO.
As for Framework themselves, I checked the product page and this is all they had to say about Linux support:
"Available in configurations with Windows 10 Home or Windows 10 Pro, we’ve also tested for compatibility with common Linux distributions and will be publishing guides on using them."
I don't see those guides in the support section yet so it seems they don't support them yet. If a vendor doesn't explicitly list an OS as supported then I don't consider it supported. So like 99% of Linux situations I would assume YMMV.
I suspect that many of the reviewers were running 21.04 as opposed to 21.10 which has been acknowledged still has support issues.
I've also been running 21.04, though, and had all the issues I mentioned above in that environment.
In my case it just worked perfectly; though I have only tried with Gentoo and Guix
I love linux distros but let me just say that anytime I hear "linux" and "just works" in the same sentence, I knowingly snicker to myself at the road that lies in front of that poor sucker. I "just works" with plenty of learning opportunities included for free.
There was a time before current sleep issues where this was true.
You could buy some laptops, install a decent distro and be on your way.
Nowadays sleep issues are affecting most recent CPUs, I had a pretty poor experience on any laptop.
I picked a Surface 2 because there was a community effort (linux-surface) and I still have the same sleep / freezing issues I had on day one, despite my model being covered in green checkmarks on the support page of the project.
Unfortunately OSS is people working for free in their spare time and you can't expect much.
Honestly I wish a company would come up and make a paid linux distro with proper support for one model.
I'd be happy to pay for an operative system that works and it's not Mac OS or Windows, as they both suck pretty bad in their latest iterations.
My ideal UX is the configuration of KDE I've used for years or my Sway setup, not window managers without tiling support and limited proprietary file managers which are designed for noobs.
>Unfortunately OSS is people working for free in their spare time and you can't expect much.
The majority of Linux dev work is actually done by paid devs at tech companies who contribute to Linux. Its been this way for years now. The reason it doesn't seem that way is because almost all of that contribution is going to enterprise code with the rest going to supporting whatever technologies those companies directly use. Google isn't spending time on making x86 laptop sleep states work, they are improving Android. Same with Amazon, they are contributing code that makes Linux more useful in AWS.
The consumer side is much much worse supported than the enterprise side. And there are actually paid versions of Linux that have paid development and support behind them. But again, it isn't aimed at PCs so PC specific features aren't getting fixed.
The only software company I can think of that is putting serious resources behind improving Linux on the PC is Valve.
> You get a much better experience using a model a year or two old because the community has had time to address any issues.
It's what I do. I bought a used LG Gram 17" with 24 GB of RAM about 18 months ago I think. Ultra lightweight (I think it's the lightest 17" laptop that exists) and I've got zero issue under Linux. It may not be as good as the Mac M1 I also own but it is lighter, has more RAM, works fine with Linux and cost me only 400 EUR (granted, that was a good deal back then).
For a laptop running Linux I never buy the latest of the latest.
I've pre-ordered my framework with the expectation that Linux support will rapidly improve because lots of hackers will buy them, same as what happened with classic thinkpads.
FWIW, if you want a laptop that is specifically designed to work with Linux, Star Labs does that.
I am surprised by the review also. While not mine a friend owns a Framework laptop and I have spent many hours trying to help them get it working as well as the Dell Latitude it replaced.
WiFi and power draw issues are the biggest two problems they continue to experience.
Also while the screen is very nice it is let down by the resolution. It isn’t high enough to used at 2x scaling but non-integer scaling really hits battery and performance. I feel that was a terrible oversight.
It is a shame as the laptops overall build quality is very good. Hopefully these common issues can be sorted properly as I would be interested in picking one up once it has rocked solid Linux support. Also a higher (or lower) resolution panel option would be nice. After all it is the Framework laptop, surely they should have different panel options :)
If I have to turn on scaling I always think I screwed up. I spent too much on pixel density that I can't even use, so now I have to make everything aliased just to read the text.
It’s probably the USB-C dock - I have the same issues with a Lenovo laptop and Ubuntu 20 LTS but works fine without dock connected.
On my phone now but there was a bug report I found to match this - apparently 5.14 kernel might work better but I haven’t been able to install it yet from the 3rd party repo.
The rest of the problems persist whether or not it's connected to the dock.
But the dock also doesn't have Ethernet problems with any of my other devices.
I have been using mine (64GB with Ubuntu 21.04, Gnome Flashback) for the past several months as my primary work laptop. I have not experienced any of the issues you mention. I like it so far. In fact, hardly notice it. Except maybe when I do sudo/login. I can use my fingerprint instead of password :). My earlier laptop was a 2012 MacBook Pro which gave up last year. I do have an M1 which is used exclusively for iOS dev.
I did have to tweak few things mentioned in their community page/forum to get everything working.
Two issues (which hasn't bothered me yet)
- battery drain during suspend
- gnome not supporting different scaling with multiple monitors (fixed in later version of gnome as I understand)
As a counterpoint, I've been running Fedora on my DIY Framework since mid-October without any issues (besides needing to install a Respin to get the required WiFi driver). Even upgraded from Fedora 34 to 35 just fine. Enabled deep sleep and it's working as intended, issues with WiFi have always been due to Spectrum's terrible non-fiber service, etc.
What OS did you install? It could be related to their choice to install a cutting edge version of Ubuntu versus Ubuntu LTS (for patches that may resolve some of what you're experiencing).
First Ubuntu 21.10 like the author. Then reverted to 21.04 because the Framework forums recommended that to get the WiFi card working at all (it didn't on 21.10).
With the RAM exception, my Framework and the author's appear to be apples to apples.
That's why this review confuses me so much; we appear to be using almost the exact same machine.
I have the same configuration as the author and get similar results (except I don’t have hibernation setup yet)
I had touchpad issues but there was a fix for it in the forums
With windows (I setup dual boot) everything works done. I had to tinker with Linux for everything to work how I wanted it. Which is what I’ve had to do with other Linux desktops too
BTW: You can buy the DIY version w/o the Intel WiFi board and bring your own that's more compatible.
Absolutely! I’d just already bought that one (not knowing the Linux driver issues) and, while considered buying another board, was fortunately able to find a workaround for the issue that fully resolves for me (just disables 802.11ax, which I don’t need presently anyway).
That's surprising to hear. I've had one for a few months now. I pulled the nvme drive from my old laptop, put it in the framework laptop, and it booted up and has been running great. I haven't had any issues opening a jpg/pdf, waking up from sleep, or wifi dying.
Which isn't to disregard what you're experiencing, but maybe try updating the firmware? I've seen at least one email they sent out about new firmware that people should probably install if they haven't yet. And maybe try out a liveusb just to see if there's something weird in your installation?
Have you contacted their tech support to see how they stand by their claims? (Not asking to be passive-aggressive; genuinely curious to hear about their after-sales.)
I see their team working alongside the community in the forums trying to troubleshoot my issues and issues like them, but I haven't found solutions there and haven't contacted them directly.
For that kind of money, getting that kind of battery life would be a return from me..
I'd love to get off the apple train, but I can't go back after the battery life on my M1.. I can do two full 8 hour workdays without a charger on this thing... With VMs..
I have no issues with mine, except for the deep sleep one. The current prevailing theory on the forums suggests it's got something to do with PCIe Gen4 SSD firmware. The issue is widespread enough that I'm hopeful we'll see a fix soon enough.
Otherwise I'm extremely happy with mine!
I have also had wireless issues, but since I've only installed NixOS I can't tell if it's because of the OS or some hardware issues.
I also thing (again very minor) that the hinge is too loose. If I have it on my lap and type on it the hinge slowly opens up. I wish there was a stiffer hinge.
What sort of issues? I'm also running NixOS on mine.
are you using NetworkManager? I'm using i3 so I use nmtui and I absolutely hate it. I much prefer iwctl, but that's not one of the supported options unfortunately...
The main issues are just difficulty connecting to networks. When connected it usually works fine, sometimes DNS will randomly fail though.
> The touch pad becomes hypersensitive.
I haven't seen a lot of the problems you're seeing, but I definitely have seen this one. Not only is the touch pad hard to use because it's very sensitive, just using the builtin keyboard can accidentally cause the touch pad to be activated. I've read up on ways of fixing this in Linux, but I haven't quite found the trick to make the touch pad behave as well as the Mac touch pad, which is what I'm used to. I haven't even really found a fix yet, so I type very carefully, avoiding resting my hands on the laptop, when I'm not using an external keyboard.
Palm detection in Linux is a definite weak area. I have tried everything and still get spurious clicks when typing on pretty much every laptop.
I feel like this happens with any item. For example, I've been supremely happy with all the Apple laptops I've ever owned, but have seen others with a litany of problems. Could it be that you got a dud with some sort of hardware problem? Try contacting their support and maybe they'll let you exchange it or something.
Stupid question: Could it be that your hardware is faulty? Power supply, memory, or a defective Mainboard?
Maybe? All the hardware came from Framework (i.e., I haven't subbed in any parts), though, so it would point to a quality control issue at least.
Do those things work well on other laptops with Linux and the same CPU?
The author does not address the screen resolution (and just says the screen is nice) but I'm hearing that the screen res is a bit to high to work with without scaling and a tad to low to do 2x scaling. And people don't like fractional scaling, at least in Gnome (it seems to look weird). Is this still an issue?
I tend to prefer 1080p for this reason on 12-14" screens. My Thinkpad X13 gen 2 has a 13.3", 1920x1200 screen, windows sets it to 150%, which seems ok, although Ubuntu works well for me without scaling still. The Frame work laptop has an even higher res (2256x1504). I do hear people scaling the fonts and that seems to be a nice solution... I remember from back in the old days that I never really liked the look of this.
What do people think of this? I'm hoping that in the future you can choose the screen (and easily get replacements as well, I hear they are working on that).
Edit: I hope these issues are being addressed by established DE's, I'm assuming they won't be an issue on Canonical's (hypothetical ;)) Flutter based DE that's (obviously ;)) coming and on System76's Rust based new DE for Pop OS .
I run Mint with 2x scaling on mine and I like it a lot. The screen's not quite as sharp as the one on the old XPS it replaced, but it's still quite nice.
I run a middling-width monospace font, and I can't quite fit two 80-column windows side by side at 2x without reducing the font size, but if you want that then you want smaller text anyway. (If you run a narrow font it shouldn't be an issue.) Other UI elements are not unreasonably large in my opinion, but YMMV.
I like my fonts on the quite small side; ~14" @ 2256x1504 sounds ideal to me.
That aside, I've been playing around with fractional scaling in Wayland on everything from a 6" Pinephone to 32"@4k in various configurations and it's been mostly painless. Mostly terminal, web browsing, Steam games (latter not on the Pinephone obv). I wouldn't be surprised if there can be issues with xwayland that I'm yet to experience.
>I've been playing around with fractional scaling in Wayland
Even with Intel graphics? I have stumbled on a number of showstopping bugs and moved back to X, which is also pretty horrible...
Oh, actually not yet. That's surprising as in general I think Intel graphics tend to have the best compatibility out there. Currently main workstation on amdgpu on a 5xxxG - I'm still not sure if I've gotten the hardware transcoding etc working properly but that's all unrelated to fractional scaling, I assume.
I’ve been using Fedora 35 with Wayland and fractional scaling set to 150% on my Framework laptop and it works great for me.
As an aside, I’m increasingly convinced that Fedora is probably the best Linux distro around right now for both enthusiasts and beginners.
Great stability. Close to the edge of progress. And they focus on delivering real unique value on top of what others have created as opposed to reskinning and hacking Gnome.
I’ve never used Fedora but I’m keen on moving over my personal desktop the next time I have a couple of days to mess around with it.
I don't understand why 1440p hasn't become the standard on laptops. Its always 1080p or 4K which is useless on such a small display.
1440p is hugely common on gaming laptops. Alienware M15, Razer Blade 14/15, Lenovo Legion 5, etc... all offer displays in the QHD or WQXGA range.
The reason you typically see 1080p or 4k on ultrabooks is because 1080p is how you get the low entry price, and 4k is what sounds better on marketing and looks the best for text-related things (4k is not at all useless - the sharpness it provides to text is noticeable). The balance that QHD provides isn't very desirable in that market usually, although there are exceptions like the Framework laptop or the Surface Laptop 4. Usually those exceptions then also come with more unique aspect ratio displays like 3:2, though, so they aren't exactly 1440p/QHD. But they are in that density.
It's not useless.
Resolution is independent of size. Just because it doesn't work perfectly on Linux high DPI is much nicer and easier to read.
The decapsulate on of resolution and size is btw already quite old. Games have this as well were they dynamically change the internal resolution but not the screen resolution.
And you might not care about it but:
- text is much smoother
- images from DSLR have higher resolution than 4k for ages and you can see the difference
The only arguments against 4k on smaller screens should be power consumption and not scaling issues. But for this we should focus on dynamic refrehsrates and similar power saving mechanism and again NOT complaining about 4k.
MacOS is doing this flawless for years. Windows can do it and under Linux it starts to be usable based on comments of this article.
MacOS doesn't scale. You can change font sizes, you can change resolution, but scaling via DPI isn't present.
If somebody understands the situation better on MacOS I'd be interested, but based on my experience, Windows is the only OS that gets this as close as possible to right. It's a single setting that affects all applications and I can set monitors independently and it will scale on the fly (even if it does look a little weird as you drag an app across and it dynamically shifts)
Every now and then I'll come across a app that doesn't quite deal with high DPI scaling correctly, but it is the exception to the rule.
Linux also works pretty well, but as with everything Linux, it's almost always "it depends and well, not quite" (not usually multi-monitor aware, lots of per app settings)
I'm not really sure what you think 'scaling' means or what you think this Stack Exchange answer means, but macOS (not 'MacOS') does allow scaling the UI of both integral and external monitors, and independently.
It renders a high-resolution image to a buffer, and then scales that when rendering to the actual screen buffer. The buffer image is at least as large as the actual screen buffer.
You can see for yourself that it isn't changing the hardware resolution of external monitors when scaling by taking a look at the actual resolution on your monitors OSD if it has done. It'll always be the native resolution.
It does not do 1.5 or 1.33 or 1.25 scaling, or does it blurred. Windows can handle any ratio.
Again I don't really know what you think you mean by 'blurred'? How do you think this is working?
If I tried all the possible scalings for my external monitor, the grey line around each window, to give a specific example of, is always a crisp physical one-pixel wide.
macOS creates a buffer to render to that is the same size as the physical output you're choosing to use, and is able to render with pixel precision in that buffer.
Would be easier if you said what scalings have you tried and what is your external monitor native resolution.
4k external monitor - here's HN at all the scalings https://imgur.com/a/QTrI6uJ. For example, look at the white box around the Y, that's always rendered precisely on hard physical pixel boundaries. It's one pixel for some scalings, then moves to two pixels for larger scalings. It's never a blur from a larger or smaller image that is then interpolated. It renders at the native resolution, knowing physical pixel boundaries, by the scaling of the render commands (the x and y to render at, before the render command runs) not an intermediate rendered image. (Remember to download the images so you aren't using browser interpolation to scale them.)
This looks pretty good. Though scaling in browsers has been mostly independent of corresponding OS features for quite a while. E.g. the issues mentioned above with Linux and fractional scaling do not affect browsers as much. The problem is usually with native UI elements, such as the settings app itself.
> In this article for Sierra you seem to be only able to pick integer factors
Maybe? But that's ancient history.
I don't know what the ratios are - the UI doesn't expose them.
What does it expose? In Sierra as you can see from the screenshot it simply shows scaled resolutions. So you can just divide native resolution by the scaled to get the factor.
It looks like it treats them as 0.5 / 0.67 / 0.78 / 0.87 / 1.0 scaling (1080 / 1440 / 1692 / 1890 / 2160 for physical resolution of 2160).
But this is all a bit arbitrary - I think the macOS API deals with an abstract 'points' unit.
Great to know. Might make me reconsider that M1 Mac Mini option.
> It does not do 1.5 or 1.33 or 1.25 scaling, or does it blurred. Windows can handle any ratio.
Non-integer scaling (of a pixel source, vectors are different but not relevant here) either requires blending or crisp unequal-sizing artifacts.
In Mac OS, because that's how its scaling is designed. Windows does not have this problem, because it tells programs to render themselves at 1.33 or whatever and provides facilities to do it without pain in the ass. As the result scaling in Windows is painless with any displays at any resolutions. In fact, my main setup has two 1.25x DPI displays and an old 1x DPI display with zero blending or sizing artifacts excluding brief moments when you move a window from 1.25x display to 1x one until it lands.
What? MacOS does scale. MacOS was doing this correctly years before Windows!
First of all, DPI is not the metric we're interested in. That is a unit for print quality, and it has no relevance for an OS, for Photoshop, or anything we do on a screen. The operating system does not know how big your screen is. Inches nor dots are relevant; only pixels are. It would be more accurate to talk about the ratio of monitor pixels to apparent pixels. I guess Microsoft finally got the memo in Windows 11 (or was it 10?) and changed to percentages.
In MacOS, a 4K monitor set to 1920x1080 will still be 3840x2160, but it displays 4 pixels for every 1 apparent pixel. Everything scales correctly, except for legacy applications. The UI elements, text, etc. appear to be a sharper version of 1080p. In Windows land, this is the same thing as "200%." As far as I can tell, Windows does the same thing but uses scaling percentages instead of apparent resolution.
Your only problem with Windows appears to be that it named the scaling setting "DPI" until Windows 7 or so to handle scaling. But as the parent mentioned, Windows is definitely superior to MacOS, as it provides arbitrary scaling ratios, and renders text perfectly at any ratio.
To go deeper into nitpicking, DPI is not even that misleading, if you take "inch" in DPI to be equal to 72 points as used in font sizes. Then a 72pt font on the screen with DPI X will be rendered with height of X pixels.
DPI is a valid metric for anything that the human eye sees. There's no difference in that regard between text printed on paper, and text displayed on the screen. The smaller the dots, the better it looks.
I'm not sure what you mean. My MacOS does the scaling and it looks good.
How is MacOS scaling different?
The problem of 4k is efficiency. 1440p is already 2x at 14 inches, which at normal viewing distances is about as sharp as people can see. Going up to 4k does not really improve visual quality, but it dramatically increases the pixel count that must be rendered, which burns through a lot more power. 4k laptops have worse battery life and worse display performance than 1440p laptops, but they offer no upside for those downsides.
There is a reason macbooks, still the gold standard for hi dpi rendering, have never had 4k panels.
Just be because apple does something doesn't mean they are right.
There have been plenty of counter examples.
Anyway, my smartphone has 1000x2000 resolution on a 6" display. The text looks perfect.
They can make very high resolution displays with very efficient energy consumption.
Feel free not to need it but my DSLR and my smartphone both take 4k and higher resolution pictures for years.
Yes I want such a display on my 14" laptop.
I do like the MacBook pro screen and it's much better than 1080p. True.
> Going up to 4k does not really improve visual quality
Speak for yourself. The 4k display in my XPS is my favorite output device I have ever owned. I wish my rMBP had that kind of pixel density.
> I don't understand why 1440p hasn't become the standard on laptops.
Because that's really low.
My 14 inch laptop is 1964 rows and I wouldn't really want anything lower than that.
I’m not on Ubuntu but using Wayland on nixos. Fractional scaling does not look or work well at all in my experience. I decided on 1.0 scaling and pushed it all over to font sizing (around 1.5). This works pretty well for most apps but it is a little small sometimes. The biggest issue is having to manually scale up some apps that don’t keep my preferences.
Gnome's accessibility feature "larger text" works a bit better than font scaling IMEx. With font scaling often some alignment and spacing seems odd. And you can toggle "larger text" with one click, if you make the accessibility menu permanent.
This is what I do as well. I find it works better than fractional scaling.
I've run Gnome+Wayland at 125% on a 32" 4k screen for over a year and noticed no issues. But that is just a sample size of 1.
But a 1080p screen would be a no for me. At 2256x1504 on 13.5" screen I think Framework laptop would probably be fine at 100% ie no scaling
Isn't scaling from 3072x1728 to 3840x2160 (so 125%) very blurry? A one-pixel line at 3072x1728 would be spread over slight more than line pixel wide on 4k.
To get nice integer scaling one could render at 15360x8640 (16k) which is 5x3072x1728 and 4x3840x2160 but I doubt that's what Wayland is doing. No common graphics card could handle that.
The font hinting relies on the resolution too. By scaling the rendered pixmap 125% the fonts would look bad too. Rendering the fonts in a larger size at the native screen resolution should look better.
Who said anything about scaling rendered pixmaps? Proper UI scaling is handled at the desktop and UI toolkit level, instead of brute force changing desktop resolution or scaling rendered windows. That's why it's not a simple problem and support varies not only based on operating system, but also UI libraries used by applications, their combination, application developers updating their UI libraries and using the new APIs. Whether the UI toolkit exposes it to application developer by having the graphic api operate at pixel independent units or having the developer manually calculate sizes based on the scaling factor or relative to something like default font size varies. Basic geometric shapes and text (assuming vector instead of bitmap fonts) can be rendered at any size and resolution. As for more photo like images downscaling them by fractional amount typically works quite well, so if they are big enough for integer 2x scaling they also work with 1.25 scale.
Here is an explanation of fractional scaling of GNOME in Wayland which is what the parent is using at 125%:
For scaling 125%, the client windows are rendered at 2x their size and sampled down. On the parent's system, a full-screen window would be rendered at 6144x3456 and scaled down to 3840x2160 (0.625%).
This gives a blurry result and is taxing on the graphics chip.
Qt apps and Firefox support scaling themselves which gives a much crisper result for fonts and line-art.
Isn't macOS using the same approach? I remember it even having a warning at certain scale that it will be more computationally intensive. And blurriness from downscaling nearest bigger integer size is nowhere near as bad as blurriness from software which use xwayland and gets rendered at 1x scale and then upscaled.
Is it really all that taxing? I mean, we're not talking about a 3D game here.
As soon as you open up a legacy X application it looks blurry. The main offender was Chrome/Electron but I think this has been fixed last year.
The screen resolution is my least favourite aspect of the Framework, for sure. Not a deal breaker, but every now and then my eye catches a poorly aliased bit of text and gets sad.
I've basically resigned myself to running Sway in 1.5x scale mode, which means that things get rendered at 2x and scaled down. Sway does about as well as I think any desktop environment could be expected to, but it's never going to be the same as rendering at true screen resolution. Alas, running in 1x or 2x and using font settings to handle it breaks down the moment you connect to an external display that really is 1x or 2x.
That said, I love the rest of the laptop (well, maybe except the battery life), so my hope is that at some point we can buy a replacement screen at a more useful resolution, at which point I'll be first in line to buy and install it.
In my experience, fractional scaling in Gnome Wayland works at least as well as on Windows. I use 150% on my 4K 27" monitor and have experienced basically zero problems in the last few months.
> I do hear people scaling the fonts and that seems to be a nice solution... I remember from back in the old days that I never really liked the look of this.
I'm doing this, in particular on a 24" 4k screen. It's true that it may look a bit "weird", in that fonts are huge compared to other UI elements.
However, for me that's a win. I don't care to have huge buttons or what have you, I mostly use shortcuts. So I get nice, sharp fonts and also get to have smaller UI elements which leave more screen space for the text.
Personally, I've found fractional scaling on Gnome makes the window furniture too large for my taste. I've settled on boosting the default font size in gnome-tweaks to 125% instead, which makes font rendering match the DPI of my monitor but leaves window bars / tool bars etc at a more sensible size.
I am on gnome 41+wayland and quite satisfied with fractional scaling. That is with firstname.lastname@example.orgX+2560x1440@1X. Fractional scaling related issues don't seem to be more often than programs not supporting wayland and scaling at all (in which case those programs are slightly blurry but reasonable size). There are certain settings which need to be enabled, they might be enabled by default on some distros. I might not notice some of the issues due to bad scaling at 4k still being comparable to lower resolution screen, unless things break badly. Issues might be more obvious with ~2K at 1.25-1.5 scale. Things may be slightly simpler in single monitor setups without mixed scaling. How much of effort configuring stuff is acceptable depends on person and what kind of software they commonly use.
* Gnome the desktop -> org.gnome.mutter/experimental-features "scale-monitor-framebuffer". Only integer scales available without this. Following assume that this is enabled.
* gnome apps: work great out of the box
* QT apps: Qt5 requires QT_QPA_PLATFORM=wayland environment variable to force wayland backend. Enabled by default in Qt6. Used to be quite bad ~1-2 years ago, but is now usable.
* Firefox: MOZ_ENABLE_WAYLAND=1
* chromium and electron: --ozone-platform=wayland seems fine for chrome, but some electron apps went crazy when moving between windows with different scale. Maybe the electron app bundled an older chrome version. No issues (except less sharp text) when not enabling the experimental wayland backend.
I've been wondering about this for a while. Why do we need scale numbers and can't use dpi? Essentially if wayland (or desktops shells) would specify sizes in a "real world unit" changing the size of windows, text etc would just mean changing the dpi. In particular it would be trivial to have different size/resolution monitors display the elements at the same size. Considering if elemtents are vector graphics we don't even need fractional scaling for desktop elements.
Now we still would need to scale images etc, but that is much less an issue than fractional scaling text or lines etc.
The only program I’ve seen do it “properly” is Darktable, which has a dots-per-inch setting for things which have to be compatible with real-world linear measurements (WYSIWYG stuff for the most part), a dots-per-degree setting for things which depend on how far away you are sitting from your screen (UI element size), and a font size. I suspect the reason people don’t regularly do this is the same as the reason CSS no longer has pixels: it’s difficult to design (in the declarative CSS model) with more than one independent length unit.
(I just went to check my several-years-old impressions against Darktable’s almost-undocumented config file, and... it no longer seems to have these settings? I’m not sure.)
Unfortunately, the “UI scale” thing seems to be baked into Wayland nowadays; I have no idea how to get Evince to realize that an A4 sheet at 100% should be 210 cm wide whatever my screen resolution and “UI scale”.
> it’s difficult to design (in the declarative CSS model) with more than one independent length unit.
Modern CSS is quite close to Turing complete, so an entirely constraint-driven design (which is what's needed when more than one unit is being chosen by the user) ought to be feasible. You need at least "real world units" for stuff like the minimum size of touch targets, and "angular units" (like CSS pixels) for the general scaling of widgets. Plus it would be nice to add a separate scaling factor for text, since users differ widely in what they find most readable. So that makes for three independent settings, in the general case.
> You need at least "real world units" for stuff like the minimum size of touch targets [...].
That touch targets need absolute lengths is a good point.
> So that makes for three independent settings, in the general case.
Yes, and you’ll notice that’s what Darktable does (or did) :) I’m not unsympathetic, just have never seen it done in a moderately complex situation.
> Modern CSS is quite close to Turing complete, so an entirely constraint-driven design (which is what's needed when more than one unit is being chosen by the user) ought to be feasible.
I suspect it’s not a question of theoretical expressive power so much as ergonomics: allowing easy description of (what people think about as) simple things. It probably doesn’t have to be precisely; if a full constraint-based / linear programming approach is the way to go, I expect designers will adapt, it’s just that last I checked those could be quite tricky to use: small changes could lead to drastic rearrangements of the layout, and as soon as you try to add a couple of breakpoints the whole thing becomes NP-complete.
It’s like, say, LR(k) parsers. LR(k) parsers are nice, fast, and good at expressing our intuitions about ambiguity (unlike ordered choice). They are expressive enough for almost anything you might want to say. But nobody really wants to work with them, the failures are tricky and hard to predict, and they lack many things you want for modularity (say, closure under unions).
It seems to me that the situation with truly resolution-independent layouts is similar in this respect: we can technically do them, to some extent, but they are a pain, so nobody does. And it’s not so much a problem of finding the perfect language as a problem of figuring out what it is that we want to say that seems so simple in our heads.
> and as soon as you try to add a couple of breakpoints the whole thing becomes NP-complete.
Sure, that's a given with the constraint satisfaction approach but NP-complete problems work just fine when N is small enough.
It's incredibly hard to change course if the UI toolkits weren't designed for it in the first place.
Android is still about the only OS with reliable density independence. It's also the only one where things like DPI were an initial part of the UI system, and not something bolted on later.
Windows has been trying to retrofit density independence for nearly a decade now, and it's still pretty hit or miss. Apple didn't bother trying at all, and has been more successful at pretending it works (so scale rendering by whole numbers only, then fractionally scale the resulting image to the actual display size). Web browsers mostly went the Apple route, but we all also regularly expect pinch/zoom scaling to fix any glaring issues anyway and just accept it's kinda janky. Linux... well, Linux is still struggling with the basics of putting stuff on screen (Wayland vs. X11). I wouldn't hold my breath for that ecosystem to get this UX refinement working well anytime soon. Especially not across the many UI widget systems, window managers with differing border decorations, DEs, etc...
Windows has had the necessary bits for DPI independence since the 90s - it's why stuff like "dialog units" is a thing.
The problem, as usual, is getting the apps to actually use it.
"dialog units" seems to be about the size of the system font, not about DPI independence. Things like WM_DPICHANGED seem to only really show up in Windows 8+, with some stuff like LOGPIXELSX showing up as "early" as Windows 2000 but doesn't look very complete.
The size of the system font was basically a proxy for DPI. If you set scale to higher than 100% in, say, Win2K, it would also increase that size.
(This doesn't handle multi-monitor setups well, since font size would be the same for them all. But multiple monitors were not exactly common back then.)
WM_DPICHANGED is a late addition, yes, but it's only necessary for the apps to dynamically reflow if the user changes that setting. Before it showed up, you'd have to restart the app for it to register the new value - but that does not preclude DPI independence as such.
The real problem was that almost everything in Win32 outside of CreateDialog uses device pixels to position. You were supposed to use GetDialogBaseUnits and/or MapDialogRect manually to get the correct values, and very few apps did.
For another early example, Visual Basic (VB6, pre-.NET, so we're talking 1998) measured everything in twips.
> The size of the system font was basically a proxy for DPI.
Not entirely. The font size can still be changed independently of the DPI, and not everything should be scaled as a result. Things like dividing lines and padding probably shouldn't be scaled when the font size is increased.
You still need a distinct DIP value independent of scaled font size. So the value of 'dp' vs. 'sp' in Android terms, which it looks like Windows' now also has that distinction.
And yeah this is ignoring that things like the visual winforms tool in Visual Studio really just wanted to operate in fixed positions, and just getting things to work when the window was resized wasn't always intuitive or obvious. So even if the APIs were there, the ecosystem definitely didn't care about it and Microsoft's own tools definitely didn't help.
If I remember correctly, the system font size couldn't be changed independently of the DPI, at least not in a supported way. You could change it yourself on a per-window basis, but a well-behaved app is not supposed to do that.
Fixed layouts is orthogonal to all that - so long as those fixed layouts are defined in DPI-independent units, they will scale just fine. FWIW in WinForms (and also Delphi VCL, where this originated) you'd normally use anchors to design layouts that scale when the window is resized. Granted, it's nowhere near as flexible as proper layout containers - which WinForms (2.0+) also has, and which the visual form designer also supports, although it can be a pain to work with them.
Oh, and LOGPIXELSX/Y showed up much earlier than Win2K. If you're getting that from MSDN, keep in mind that 2K is basically as early as the docs go, even if the actual API is much older. For example, it says the same for CreateWindow, even though that one goes all the way back to Windows 1.0.
I use Pop_OS on my Framework and loved 1.5x scaling except for screen tearing I get (YouTube for example). I reverted to 1x and increased icon and font sizes to semi compensate.
Do you still get the tearing when you use Wayland?
Even if you're on 1x scaling, you always have the option to adjust the zoom in your browser, and the text size in your terminal/IDE/text editor.
I love high res smaller screens. I have the X13 with 2560x1600. I leave the screen at 100% (perfect size for window widgets, etc.) and scale fonts by 1.5. Works great and everything is super sharp. I can also scale down the fonts if I want and don't lose the sharpness due to the increased dpi.
Instead of scaling the whole of the UI, I find that scaling via fonts works well enough when integer based scaling is not an option.
I've got a Thinkpad with a 14" screen at 2256x1504. Running Wayland and SwayWM, I could never get integer or fractional scaling to work just right. In the end, turning off system-wide scaling and configuring apps that don't scale automatically was what did the trick.
How long ago did you try last? I recently started using Sway and haven't hit any issues yet.
Sorry, I wasn't clear - just mentioning that I'm using Sway, but it's not the problem. The problem was with apps running in XWayland, e.g. Gimp, Firefox, Spotify, Slack, gnome apps, etc
Ah, gotcha. The Wayland situation in Firefox has been changing for every major release BTW. Been running it in native Wayland since recently and so far it’s mostly been good and stable. Still the occasional crash or popup glitch but since ~v94 it already feels worth it.
I used to have issues with two different screens that needed different scaling but as of 20.04 it's mostly fine to use fractional scaling on one monitor and not the other.
I've still had some issues with some applications not scaling their widgets properly, but it's mostly ok.
Anecdotal personal experience:
I’m on PopOS running Wayland with fractional scaling to 200%. It looks fantastic imo and have only had some issues with Guake terminal when plugged in to an external monitor, which I believe may be a Guake-Wayland-specific bug.
200% is not fractional scaling, though?
Oops, you’re right. 150% is what I have set, typo on my end!
> I have a machine with better specs than a comparable MacBook Pro M1 for less.
The “for less” bit might be true but is the better specs part?
In a comparison the M1 in a MacBook Air seems to handily beat the Intel Core i7 1165G7 chip on most metrics.
Personally I’m not that CPU bound generally and I like Intel Linux compatibility so I’d still make the trade off.
I was in the market for a new laptop and right now I just could not find anything that would beat the M1 MacBooks on a CPU-battery-price combination, especially if you're OK 16GB of RAM - it gets a bit more competitive if you're looking at 32GB models, but even if you can compete on CPU and memory you loose hard on battery life.
I just couldn't bring myself to use macOS daily, so I punted the laptop change to next year. I wish Apple would just sell their hardware like anyone else, working with OS vendors to get it supported.
> I just couldn't bring myself to use macOS daily
I have to use it for work, and it really is a drag compared to Linux.
same. i have a maxed mbp 16" from work and it's amazing how much that thing stutters during normal usage.
also, sometimes even a simple `ls` from terminal takes a few seconds. why? i have absolutely no idea.
> why? i have absolutely no idea.
If you don't have a stable internet connection, that could be the reason - macOS sends a hash of every executable to Apple's servers before it is opened. This caused a major issue at the end of 2020, when these servers stopped working and all macs stopped working unless disconnected from the internet.
macOS has also been updated so that syspolicyd bypasses VPNs and system firewalls like Little Snitch, so you can't easily block these connections now.
Your own links show that likely isn't parent's issue. It only sends a hash on the first run of an executable. I'm not saying the problem you're talking about isn't a problem or concerning, but it's very likely not the problem they're talking about.
I hate when people say "mine works," but here's an `ls` of my homedir showing it's not universally slow. I currently have an absolute garbage network connection.
> ls -G 0.00s user 0.00s system 64% cpu 0.010 total
There are also many, many other reasons it could be; some macOS specific and others that aren't--most importantly what they're seeing isn't universal. macOS often ships with very old GPL2 tools that can cause various problems (many people brew install updated GPL3 versions), people often have configurations that can slow down `ls` by multiple factors (colors, sorting, etc can each cause multiple queries to disk or require the listing to complete before displaying output), customizations causing a slow prompt, a slow or corrupt disk, listing a slow network drive, etc.
The VPN bypass was very quickly removed from macOS over a year ago . So it would only be relevant if they were using a very old version of Big Sur.
I'll jump in here and say that it probably _is_ notarization. The issues arrise when osx thinks it can get a connection to ocsp but actually because of real world consequences it can't. This can cause a delay of upto 5 seconds while it times out.
Some specific examples,
No internet connection: instant fail over
Blocked OCSP firewall or whatever: instant fail over
Slow internet but still able to reach: slow start: 1+ seconds
Bad internet, not able to reach: 3-5 second delay waiting
Normal internet, OSCP reachable: <1 second delay
Disabled trustd: Nothing will start, single user mode and trustd restore required
I've experienced all of these and is one of the reasons I have a shiney new Framework laptop sitting waiting to be migrated over to.
Also the "only on first run" also isn't true. It periodically checks for certificate revocation (as it should) and therefore will cause issues at sporadic intervals.
And the kicker of course is that all this is via plain ol' http, so everyone knows what developer's programs you're starting via the hash.
On my 2019 work machine (MBP) all filesystem operations were slow. On my 2021 one they aren't. Same specs, same OS. Maybe there's some bug in the filesystem? Not sure if you can fix by clean-reinstalling or if it's tied to the hardware.
My 2019 (intel) work MBP is painfully slow and stuttery compared to my personal (M1) laptop. I wonder if it's partly the fault of the software that my company has pre-loaded on there.
i really think it's the monitoring software that is installed. the thing is, i complained to internal support and they said "i was the only one complaining" so ¯\_(ツ)_/¯
I’ve owned macs for years and never once had any issue like that. I’d dig deeper than “it must be macOS”
Genuine question: what are you doing at work that would make Linux more desirable? Unless you write applications for Linux, I'm have a hard time thinking of a niche that would make Linux better.
I do recognize that you could also just like Linux more from a usability perspective, which is your opinion to hold!
I write code.
Mind elaborating a little bit?
I've tried to push myself to use various Linux distros (most recently Pop and Ubuntu), but I always end up back on MacOS.
It's just not set up the way I want, with things like focus follows mouse, and it's harder to customize than Linux with a good window manager. I also prefer how Linux package management works; that it's an integral part of the system rather than something bolted on. I do like the screen and keyboard on the Mac, with the exception of the lack of volume buttons.
I also like the open nature of Linux. It's fun to be able to look into the source code for pretty much anything.
Apple has kept the M1 Macs open for other operating systems, which is why a project like Asahi Linux can exist: https://asahilinux.org
Apple doesn't provide Linux drivers for their components, but neither does "anyone else", sadly.
But it means that Linux on M1 Macs might actually happen in the near future.
> Apple doesn't provide Linux drivers for their components, but neither does "anyone else", sadly.
There are vendor-provided Linux drivers for pretty much everything?
Everything in Intel and AMD CPUs are supported in mainline. nVidia provides Linux drivers.
I don't have a clear view about how many devices nowadays are reverse-engineered, and how many are provided by someone with datasheet (it happens quite often that mainline drivers are not provided by vendor directly, but by a third party that got access to datasheet), but it really feels like most vendors do work with OS vendors to get their hardware supported, and that Apple is totally an exception in the hardware world.
I am aware that for GPUs specifically, the driver situation on Linux is pretty good (minus the Nvidia/Wayland friction). But a laptop is more than just a GPU and a CPU.
When I was shopping around a few years ago, the fingerprint reader even on Dell's Linux-focused developer laptops was dead. I liked the Razer Blade, but many components didn't support Linux; looks like the trackpad on some models still doesn't work.
If "any vendor" actively supported Linux, we wouldn't need companies like Framework or system76 that pick compatible components. I'm not saying that Apple is particularly helpful, just that they're hardly alone in not caring about Linux.
Apple has requested and obtained semi-custom hardware from vendors before. Their Boradcom wireless cards (around 2006-2012 IIRC) are almost produced only for them, and getting it run on Linux was a long endeavor.
Also, like Java, they have maintained their own nVidia drivers up to a certain point.
Today that gap is smaller since HW manufacturers are more open about their firmwares and drivers, and Linux is more widely adopted.
> Apple has requested and obtained semi-custom hardware from vendors before. Their Boradcom wireless cards (around 2006-2012 IIRC) are almost produced only for them, and getting it run on Linux was a long endeavor.
I just sent in a patch series to Linux a few days ago to get all custom Broadcom cards for T2 and M1 Macs working (~2017-present), with the right firmware selection logic. Took a week or so to develop and test on both ARM64 and x86. (There was some prior art but certainly nothing that would've taken more than a few more weeks to work out from scratch; besides, I found a Broadcom source dump for Android that contained most of the interesting knowledge excluding the Apple-only bits, since some of it is relevant to newer non-Apple chips too).
Once you have a team of the right people and enough motivation, getting reverse engineered hardware working on Linux doesn't take as long as everyone thinks :-)
I remember many people have bruised themselves by banging their heads to their peripherals and desks while trying to reverse these BCM chips and trying to make them work with Linux, however back in the day BCM was hostile towards Linux, and NDIS_wrapper was a thing...
OTOH, kudos for all the hard work you're doing on the kernel. Maybe one day I'd have enough spare time to work on that marvelous beast.
> Once you have a team of the right people and enough motivation, getting reverse engineered hardware working on Linux doesn't take as long as everyone thinks :-)
As long as the hardware (or the vendor) is not intentionally malicious towards Linux or being reverse engineered. I've seen my fair share. :-)
> I remember many people have bruised themselves by banging their heads to their peripherals and desks while trying to reverse these BCM chips and trying to make them work with Linux, however back in the day BCM was hostile towards Linux, and NDIS_wrapper was a thing...
The b43 days were definitely special, and that was a full from scratch reverse engineered driver plus open firmware, which is a crazy accomplishment. These days though, we just accept that we're going to have to run vendor firmware, either because there's signing involved or because it's a huge amount of work to attempt to reverse engineer and reimplement that bit. Firmwares have grown from dozens of kilobytes to megabytes and it's just not viable to rewrite all that any more, if it's even possible.
OTOH, we also have better tools for and experience reverse engineering now than we did back then. Ghidra has made "proper" disassembly and decompilation accessible to anyone, and we can run proprietary OSes in VMs to study how they interact with hardware.
> As long as the hardware (or the vendor) is not intentionally malicious towards Linux or being reverse engineered. I've seen my fair share. :-)
Well, it doesn't happen that often outside of closed platforms (i.e. needing an exploit to do anything) and Nvidia ;-)
Apple still maintains their own Java internally. Very minor tweaks and with their own CA since they self sign certain internally starting 2-3 years ago.
But the packages are still distributed by ORACLE, right?
Not the internal Java. That’s not for customers. That’s for apples internal servers.
AMD and Intel provide drivers for their components. Realtek provides drivers for their components. Even Nvidia provides Linux drivers for their components, in proprietary form. Several laptop vendors even sell their hardware with Linux preinstalled.
Apple not actively hindering Asahi Linux isn't exactly a great accomplishment. They chose not to write Linux drivers for their platform and that's their choice, but that puts them behind even Nvidia as far as supporting Linux goes.
FWIW, Nvidia actively sabotaged the Nouveau project with the firmware signing story; Apple aren't doing that (their firmware is signed but loaded pre-boot and they provide OS images we can use to set all that up, so it's seamless for Linux).
Great progress on Asahi. What's the firmware situation like in general ? Since firmware won't be redistributable, would users need to periodically 'fetch' it from Apple CDNs or OS images ? Also, in the past broadcom has been quite problematic for wifi. How is it on the new macs ?
There's two classes of firmware: the bits loaded pre OS boot by the OS loader (which runs before our Linux loader), and the bits loaded by Linux. Both classes come from macOS recovery images and our installer takes care of fetching that from Apple's CDN. For the pre-boot firmware, it just gets installed into the Preboot partition the same way it does for macOS, and we don't have to worry about it after that. For the firmware loaded by Linux (current list: WiFi, video decoder), I have a design for a vendor firmware mechanism where the installer copies a tarball to the EFI boot partition, and then distros will have a script that extracts that on first boot or when it changes.
Since Apple do not keep their firmware ABI stable across the board, we'll have specific firmware versions (=specific macOS versions they come from) "blessed" for Linux compatibility. For updates (bugfixes or security issues, chiefly) you'd re-run our installer which will eventually have an "upgrade my firmware please" mode. That would run from recovery mode, same as the original install. All this stuff is per-OS, so it does not affect any adjacent macOS install which will use its own firmware - you can upgrade macOS whenever you want without breaking Linux.
Wifi works fine (modulo some firmware support details that'll get ironed out); people have been doing performance and stability tests and there is no major issue. I've heard stories about bad radio performance on Macs in the past, but those are almost certainly due to using the wrong firmware or config. That's what my patch set fixes: we use all the same firmwares and configs that Apple does, so we get the same results.
> AMD and Intel provide drivers for their components.
ATI/AMD intentionally redesigned their GPU silicon to make it's possible to develop open source drivers without exposing HDCP secrets, and even the open source drivers can use HDCP if they want to.
This is not known very widely, but this is worthy of great praise if you ask me.
I don't think any GPU vendor has ever had HDCP secrets exposed to the OS. That'd be terrible security; you'd need mandatory OS secure boot to keep the keys secret. This is the case for Nvidia, for Apple M1s, etc; the OS never has to touch plaintext HDCP keys, and this has nothing to do with open source drivers.
Can you clarify what you mean exactly? I have a hard time believing AMD ever ran HDCP keys through the Windows kernel.
> I don't think any GPU vendor has ever had HDCP secrets exposed to the OS. That'd be terrible security; you'd need mandatory OS secure boot to keep the keys secret.
Yes, I'd not argue with that, I also never wanted or implied that I shall be able to get said keys from the card.
AMD/ATI back in the day said that, they have intentionally coupled video decoders, related logic and HDCP to same IP block for performance and silicon real-estate reasons, but in order to keep their promises about open drivers and accessible hardware, starting with (then) next generation silicon, they'll move all HDCP stuff to another IP. In turn, they will allow unfettered access to all GPU including 3D and video related features sans HDCP. However, with that architecture, even open drivers can use HDCP related hardware and functions without compromising the security and licensing around it, as a result allowing identical experience and freedom (from a developer standpoint) regardless of the driver used.
This is something to praise for, IMHO.
I still don't really understand how this is relevant for the drivers. Even if HDCP is part of another IP block, that doesn't mean open drivers can't use it. That's just an implementation detail. You can use HDCP on any platform the same way the proprietary drivers would (if you implement the right knobs to control it). Do you have a reference with more detail?
Ultimately the open driver should always be able to do all the same things the closed driver does and get the same result. If there are any secrets in the closed driver that "can't" be exposed, that's a security flaw in the design.
In their previous setup, as per their word, it was impossible to access video decoders and related machinery without exposing some HDCP stuff, incl. the key related storage and other secret stuff.
So, they were unable to provide the access they wanted to provide (for video decoders, at least), on the previous (then current) architecture.
That's... very weird. What would that expose? If you can read the keys at all from the OS, directly or indirectly, that's a security flaw. AFAIK AMD GPUs have supported video decode on Linux across all generations, so what was unavailable?
Considering the history of ATI with video, and how they enabled GL shaders to work with video acceleration, post processing and effects before anyone was able to dream this thing (Readeon 8500 was able to use shaders to post-process and apply effects to the video), I think their silicon layout was very convoluted and unconventional for today's standards.
Considering HDCP came later, it might have just been slapped as a side module and some of their math magic might be offloaded to GPU even. I remember having old ATI drivers having a very cinematic look when video was accelerated by the GPU itself (it was surreally beautiful), but they had to throw all this machinery out from the driver when they started to modernize their stack (and they broke video playback for months due to not supporting proper overlay formats).
Similarly, I think closed source drivers were programming some of the cards during initialization, because I remember they're talking about moving all this init step to the BIOS itself, and opening some doors on the BIOS to tweak the card during the init (clocks, fans, temp reads, etc. IIRC) for allowing OS and open drivers to play nicer and with less effort with the cards in the long run.
The details are hazy, because these transitions were made almost a decade ago if not longer. However, while NVIDIA was playing the deaf, and Intel was just starting to provide good open drivers for their IGPs, AMDs move as a prime GPU vendor brought a lot of good light to them.
If you want, I can try to dig the history and create a timeline for you, but it might not be an immediate endeavor.
Oh, I know that AMD has done a lot to help out the open ecosystem (and even interop with the closed parts - heck, I got their closed AMDGPU-PRO Linux drivers to work on the PS4 without any patches, even though that's not a GPU variant offered elsewhere, because the amdgpu userspace/kernel interface is just so well designed). I recommend AMD for Linux users; lately, they've been doing good work while Intel drivers bitrot, especially on older hardware.
I'm just quite confused as to how HDCP factors into all that. HDCP really is just a hardware block that's part of the video output which you enable or disable; there is a bunch of nonsense associated with PAVP and other such stuff on Windows, but none of that should really matter to Linux, nor should whatever way HDCP is implemented affect other features of the open drivers. AMD certainly do deserve credit for their work on Linux support, I just find your specific HDCP story quite strange :-)
But it is not all about GPUs. Broadcom is notoriously bad on many items and is used a lot for wifi and BLE and webcams, as is that touch pad vendor whose name I forgot that is in many laptops.
I installed Linux on a 2015 MBP recently and there is no real way to get the webcam working.
I'd agree if we were talking about desktop machines, which are basically a box around a GPU, CPU, and some networking. The situation there is infinitely better than on the M1 Mac Mini.
But if I go and buy a random Windows _laptop_ and try to slap Linux on it, I might have all sorts of issues with the trackpad, webcam, fingerprint sensors, never mind sleep/wake.
That's why Linux laptop vendors are a thing and deserve our praise (just like the companies you've mentioned)! All I'm saying is that Apple is hardly the only company left that doesn't actively support Linux.
People unfamiliar with computer history beyond the past decade are downvoting you.
Apple may not contribute much more than a bare platform with M1, but that isn't unusual in the scope of computer history. What's important is that they are not actively enforcing exclusivity or deterring the attempt. They've even shown tacit support in recent EFI patches by making changes that preserve the current mechanism other OSes can be loaded.
If you're ok with 16 GB, how do you use the large CPU power? Just curious about various usage patterns other people have.
Does it make compiling stuff, running unit tests, etc. faster?
The new i9-12900HK processors are supposed to be more powerful than M1 Max. I am personally waiting till these show up in laptops.
Being able to outperform an M1 Max isn't the issue. It's doing it while also being as efficient as it.
No one doubted that Intel could make a CPU as powerful. But everyone knows that it isn't gonna be cool and quiet.
While M1s definitely win in the efficiency category, the added efficiency is really not that useful for most people on a day to day basis.
Cellphones have moved away from replaceable batteries specifically because of fast charge. Most phones with 20 watt charging will reach ~50% battery in 30 mins, so even with battery degradation, you really don't notice the effect in day to day life.
Same thing with laptops, there really are very few situations where you don't have access to a charging port for ~6-8 hours.
Given that, Id personally rather take the hit on efficiency in return for way better compatibility, ability to run a full linux natively , and not deal with all the extra crap that Apple throws in the M1. And, If I really care about longevity, a second battery is easily an option for laptops like Frameworks.
The power requirements to achieve that though are insane. Battery life will be small; heatsinks will be massive.
Either that... or massive throttling due to manufacturers putting them in thin laptops without adequate heat dissipation.
Which lose to the M1 handily on the battery and price aspects.
at the expense of twice the power consumption.
Raw CPU perf obviously favours the M1, but if you're a dev, then a laptop that can take up to 64Gb of commodity RAM & as much fast storage as you can cram into it is pretty compelling.
Compiling LLVM eats a lot of memory for instance. 16Gb is not really enough to take full advantage of the number of CPUs in a modern laptop.
Is this a laptop for everyone? Not necessarily, but for certain classes of users it's manna from heaven.
You realize the framework laptop’s cpus are limited to 4 cores, right?
No $2000 laptop in 2020 should have only quad core cpus.
The M1 16gb will run circles around the i7-1185g7 (highest spec chip framework sells) compiling llvm with 64gb of ram.
Only using a shitload of idle VMs will have any apparent benefit over M1 MacBook pros.
The cpu is just too limiting for the 64gb of ram argument.
>The M1 16gb will run circles around the i7-1185g7
I have an M1 MBP and an 1185g7 based Linux laptop. Single core benchmarks, the i7 comes out slightly ahead. Multi core benchmarks, the M1 comes out slightly ahead. There are no circles being run around in performance - only in battery efficiency. For day to day developer usage, testing things literally side by side, I find absolutely no discernible difference in speed.
The M1 does win for battery life, but the battery still lasts for a full work day in my i7 laptop. Honestly at home I always reach for the Linux laptop, because I don't really need the battery life. But for going out and about away from power, the M1 MBP has advantages. Sadly the software (personally) lets it down, I'd probably use the MBP more if it has better Linux support in the future.
Prolonged compiling tasks will thermally throttle the 1185g7 chip long before the M1.
It will run circles around the 1185g7 compiling large codebases.
Passmarks across hundreds of benches show a 25% single and 40% multithreaded performance boost over the 1185g7.
I’ve got an i9 8/16 laptop chip and the 13” M1 is trading blows with my laptop despite 2-3x the thermal headroom and considerably less cores/threads across all workloads.
The i9 Apple laptops were notorious for not having a cooling system that could match the heat output of the CPU & throttling almost instantly.
Modern Intel laptop CPUs are no where near the performance / W of the M1, but they are considerably better than those that Apple stuffed into their laptops a few years ago with inadequate cooling I believe.
That was a firmware issue pre 2019 model that was corrected in a few weeks that got overblown. The throttling problem doesn’t exist anymore.
The only problem with the first generation framework laptop is intel’s quad core CPU because of thermal and power problems associated with their process nodes.
Hex core is really the lowest I will go now for anything over $1000. If this was a Ryzen 5800u, I’d have bought one already.
On my 1185g7 laptop, the browser benchmark Speedometer 2.0 scores 116. On my M1, I get 288. That's a bit more than "slightly ahead". The M1 absolutely smokes Tiger Lake and even Tiger Lake-H.
FWIW, Cinebench multi-core benchmark is around 20% faster for me on the M1.
I doubt that instruction is relevant. It comes down to an abundance of microarchitectural resources on the M1. A 12th-gen Intel CPU gets 310 on this benchmark while having no particular accelerators. They just added tons of caches and buffers, like the M1 has.
Sure: 4 cores, but 8 hyperthreads.
LLVM builds eat memory in my experience & end up swapping if you don’t play games with the build system to reduce the parallelism of memory intensive parts of the build.
If you’ve got a VM or two sitting around then obviously that makes things worse.
Yeah I totally agree with that sentiment.
It’s often tempting to dismiss a product that doesn’t meet your own personal use cases but it’s a bit naive.
It’s actually fun to see products get popular that I know I wouldn’t buy myself. It doesn’t make them bad it just makes them not for me right now.
Is this really common? I mean using a laptop for such compute intensive tasks? I mostly use my macbook air m1 as a client to more powerful machines. I always have a few ssh sessions opened, a few tramp emacs buffers, some browser tab to jupyter notebooks running elsewhere...
I think it is quite common. I develop financial systems, and I compile apps regularly for running tests locally and large parts of the system is running locally using docker compose. This has the benefit of fast iteration for unit and system tests. Of course the tests and deployment is also run by a CI server but that is just for QA purposes so people dont have to remember to run their tests locally and to keep a clean main branch. I cannot imagine how long our feedback loop would be if we had to wait for the CI server for every change, or if we had a common dev server that everyone interacted with through a «dirty» development branch. Basically I think it is just different use cases. My experience with jupyter is that you use it for data analysis which is really more IO bound than cpu bound, also big data is often not possible to keep locally (because of the size and also in europe GDPR)
Note that I didn't necessarily refer to CI servers when talking about remote machines. Sometimes it's just a better specced workstation in the office, sometimes it's my own headless machine learning rig at home. Sure different engineers and different fields have different workflows but I believe it's still easier to build a good workstation from components than finding a laptop that excels in both portability and specs.
Interesting approach. I have some coleagues that actually remote desktops to their powerful desktop computer. I never did that because I fear all the tiny issues I have to deal with, but they tell me it works quite well. Also just offloading the heavy lifting through ssh or other types of build servers seems super interesting (I think some build systems have built in support for a common server, not sure if it is true, but it could be very helpful if it works well)
Yes, I specced my machines for local docker-compose of testing. It's pretty valuable, although different engineers use it different amounts.
Presumably, the people who were using desktop workstations pre-pandemic have started working from home with laptops - but their ways of working and tooling all center around doing things locally.
Kernel compiles are not particularly memory hungry. LLVM compiles on the other hand are very memory hungry, especially during linking.
I’m sure there are other 'workstation-ish' jobs that people would happily run on their laptop if it had > 16Gb memory. 64Gb is quite the leap.
As he explicitly points out, the benchmark amounts to cross-compiling for the Intel CPU vs compiling native code for the Mac CPU, so it's not exactly even. I have no dog in this hunt, I don't have an M1 Mac and I don't care, but it felt wrong to let that stand without correction.
1) There are 64Gb M1 MacBooks.
2) If you need 64Gb just sometimes, swapping on M1 MBP is insanely fast.
64GB ram for my lenovo legion was 250EUR, i could not configure a m1 with 64gb for less then 2900EUR
i hate it that they bundled the ram with the m1 max i wanted to order the 64gb ram option but the difference between 32gb pro and 64gb max is 650EUR or more, and i just dont want to pay that to have more RAM.
Your laptop has DDR4 memory. I’ve checked prices for DDR5 memory 64Gb - around ~€500. So the difference with Apple is not so dramatic.
it cost 250EUR to go from 0GB to 64GB on lenovo, but it cost 650EUR to go from 32GB pro to 64GB max macbook pro 16,
so it is dramatic for me.
Im buying a macbook for 4-5 years.
You want to ignore the difference between DDR4 and DDR5, but the market doesn't ignore it and DDR5 is more expensive (I'm talking about non-Apple memory sticks).
Also, you want to ignore 400 Gb/s provided by Apple's 64Gb memory, but... you know.
After using a dead-quiet laptop for a few weeks, it's hard even to think about Intel airplane+heater offerings.
IMHO it's a bit troublesome that it ships with a 1165G7, while competition available during the same time (e.g. T14 gen 2) is already on a noticable faster ryzen 5000 (through also stuck at only 16GiB of RAM).
I understand that upgrading the CPU is not yet quit viable at this point in time for framework but still, it noticable lacks behind the competition when it comes to CPU perf for many (all) tasks.
And things will only get harder with the Ryzen 6000 laptop CPUs coming late this year (on the German marked maybe only early next year).
And similar is true for the competition from newer Intel processors.
This kinda has me stuck, 16GiB of RAM isn't enough, but the additional perf of the newer CPU is well wanted, but then USB ports being non easily repairable on a T14 is a problem to as I somehow tend to brake them. And only 2 USB-C ports on the T14 are a proble, too.
he was not talking only about CPU but about extensibility!
> Graphics card - Can this be made upgradable in the future? Would seriously consider building one for gaming if they were.
Oh man if they found a way to have upgradable GPU on a laptop it would be next level! I think it can be done through M.2 slot (the reason why you can get M.2 to PCIe adapters is because M.2 provides a PCIe interface , some eGPU solutions use M.2 , IIRC it's just a 4 lanes connection through the M.2 connector but better than nothing) - so in order to make this possible on a laptop you would basically have to take small (mobile class) GPUs and solder them on a board that can be inserted in an M.2 slot... I think this is not rocket science but probably those GPUs would be expensive unless the concept really takes off and they are mass produced.
EDIT - I just remembered that in the past there were some gaming laptops with upgradeable GPU, I think they were made maybe by Asus? And if I remember correctly it was a proprietary connector (but I might be wrong)
They did in the past made laptops with upgradable GPU, I have one Asus that still is working (from 2009).
However so few people ever bothered to upgrade them, hence why OEMs have moved on.
Even on desktops, every time I came around to upgrade components back in the day, it was about time to upgrade everything.
That's my experience too. Honestly I don't think people are going to upgrade their Framework laptops much. And while they're popular now, it's still likely they'll stop producing parts (assuming the standard for each component isn't completely free to copy). Then you're out of luck.
But the real huge win is replacing broken components easily. That alone is making the Framework extremely tempting for me.
It was dells area 51m recently. I own one, and it was marketed as upgradeable, though it turns out its only upgradable in the same class e.g. mine can only be swapped with a 2060 2070 or 2080.
Just use an external GPU?
Or make it upgradeable like an external GPU.
Using a mobile GPU chip and putting it into a small external GPU case could be easy.
The only reason why current external GPU cases are so big is that the form factor of GPUs is big and the big power consumption.
Using a mobile GPU could make this into a small cube of 15*15 or so.
Would you settle for an external eGPU over USBC?
This is what I have been looking into for a portable gaming setup.
My framework setup already has so much raw power in terms of RAM and CPU. A solid GPU is the next step :-)
Would do a GPU pass through setup with virtio/QEMU and a Linux host… if only eGPU setups weren’t so damned expensive! And that’s not even factoring in the sheer lack of availability of 30x series cards.
I have a custom built small form factor PC in a Dan Case A4 . The case is a lot smaller than a Razer Core. And it's a whole PC, not just a GPU. Fits in a cabin bag or backpack and I have traveled with it.
You can go even smaller at 3.9L in a Velka 3 , if you use a single fan graphics card.
The Dan Case is absolutely gorgeous. What hardware are you running in yours?
I’d love to get a setup like this going with a 30x series card, but I’m not sure how difficult getting one at listing price is these days.
No chance buying a GPU at listing price today.
Buyer beware: for the Dan Case, it's only compatible with 2.25 slot (45mm) width cards at max. Most 3080s are 2.5 slot (50mm) wide. The EVGA 3080 XC3 Black is a notable exception, and is a very satisfying tight fit . There's also a 3D printed GPU spacer to increase width of the case (sacrificing portability) .
Upgrading a GPU internally in a laptop chassis is tricky because you need the power/heat/space to work for the range of possible products. And I like a CPU where you might design for a 35W CPU and that gives you dozens of options from 15W to 35W, the GPU range needs to encompass 100W (!) if the purpose is meant to allow configuring as a gaming machine. This will obviously affect cooling/power supply/bulk for all who buy the laptop and don’t want the gaming config.
External GPUs have their own difficulties but I think it’s a less challenging design problem than “configurable GPUs [up to enthusiast/gaming/pro cards]”
My HP ZBook from 2014 has upgradable CPU and GPU. I guess that after a few years the problem would be to find a new part that fits into the slots or somebody selling an old compatible one at a reasonable price.
> were some gaming laptops with upgradeable GPU
You are thinking about MXM 
Idea was neat but Regular Joe doesn't upgrade GPU in the laptop. And Non-regular Joes who does demand what it should be cheap and performant.
I will buy a Framework laptop in a heartbeat if they came with RTX GPUs.
It would be doubly great if the GPU were upgradable as you say.
The Schenker vision 14 comes with 3050ti and it is pretty repairable
I'm using a Framework with Kubuntu and I have mixed feelings about it. I like the 3:2 screen ratio, repairability & the expansion cards. However, there are many downsides:
- The battery life is ATROCIOUS, even after TLP etc. It's about the same as my 4-year-old(!) Thinkpad. Standby battery drain is the worst, easily 20-30% overnight, which I find unacceptable. I can't use this laptop as intended and always have to switch it off completely. Also because of this:
- BIOS 3.06 ships with a bug that could result in the Framework not switching on if the battery is drained to 0%. The new beta BIOS might fix that, but errors like this show this is not a mature product.
- There are no stable LVFS firmware upgrades for Linux yet and some users also report overwritten bootloaders after upgrading. Linux compatibility is definitely not there yet.
- Many small annoyances on Linux: Fingerprint reader is not working out of the box, screen tearing, Bluetooth regressions etc. on certain kernels.
- My Framework didn't turn on for almost 2 weeks. I tried different RAM/SSDs to no avail... then, it suddenly worked again with the original components. No problems after that, very strange.
- The speakers are worse than the one in my smartphone. On most surfaces it sounds muffled and just not right.
- My CPU fan made strange noises. I could fix that thanks to the great repairability though.
- The fan can get very loud. Fortunately, it happens very rarely. Most of the time it's silent when browsing the internet or doing web dev work.
- Build quality is clearly a step down from my old Thinkpad X1 Yoga. The hinge doesn't feel as strong, some keys are mushy/creaking and I'm skeptical my Framework will survive as many falls as my old laptop.
Don't get me wrong: it's impressive for a first iteration product and a lot of modern laptops can't compete with it (despite the Framework being far more extensible). It ticks many boxes and offers a package that is hard to find these days (Lenovo & Co. seem to love soldered-on RAM, decreased keyboard travel, fewer ports).
After many glowing reviews I just expected a bit more. Coming from a 4-yo premium Thinkpad I'm not sure the Framework is an upgrade. It's more fragile, (currently) has worse Linux support & no next-day business support. I would definitely wait for the next generation of laptops.
Standby battery drain in deep sleep in Linux is something we're investigating. We have seen enough reports of it to know there is an issue there, but one that does not occur in Windows.
We do strongly recommend updating to 3.07. The issue in 3.06 was a regression that we've released 3.07 to resolve. We'll be removing the Beta label on the release shortly since we've seen large update on the release.
Beyond that, we very much appreciate the feedback. We're always looking to improve what we're building, and real user feedback helps us do that.
Maybe need mem_sleep_default=deep in the boot parameters in the grub config - this was needed on my XPS to stop battery drain during suspend.
Too late to edit, but that should be "we've seen large uptake on the release".
I have to disagree on the build quality, I found it really only second to MacBooks and Dell XPS. Ultimately I returned it for an M1 MacBook because of the battery. With an M1 I charge it once a week and it sips power on standby. The Framework was dropping 30-40% battery overnight on sleep.
That is most likey a Linux problem in general. Linux is known to have worse energy management. TFA explains the owner had to enable deep sleep and hibernation:
> One of the initial caveats with installing Ubuntu is that deep sleep and hibernate are not enabled by default. Deep sleep is easy to enable. Hibernate takes a bit more effort, but it’s straightforward. Once configured it works like a charm. I noticed about a 2% drain in 3 days of hibernation. If this power loss is linear, it can go for weeks in hibernation.
My battery results were with deep sleep. I was unable to get hibernation working although I didn’t put a lot of time into trying it.
> Standby battery drain is the worst, easily 20-30% overnight, which I find unacceptable.
I guess that’s the “modern standby” that is replacing s3 sleep these days.
Unfortunately this is happening with S3 sleep and numerous battery tuning tricks activated.
Don't buy the Framework as your main laptop if you plan to travel/use it unplugged a lot - I consider it a broken device for this intended usage. Put down the lid overnight and you'll have around 2-4h SOT left (and that's with a fully charged brand-new battery).
No other business laptop in the last 15 years felt so limiting in this regard, battery life is a constant worry. My Framework won't leave the dock and I'll hope manufacturers close the gap to the M1 soon (I don't care about raw power, but energy efficiency).
No, it isn't even. Mine drains that fast with the s3 "deep" sleep enabled.
I'm glad to find that it's not just me. I was wondering what I've done wrong to wake up with 80% battery after closing the lid and unplugging it. After two swollen batteries in a MacBook Pro, I've made a habit of not leaving a laptop on the charger 24/7, but it sucks to throw this thing in a backpack and head to a coffee shop only to discover that I don't have a full charge.
I'll also add that I'm surprised at how long it takes to wake up from deep sleep. It's like a full 15-30 seconds of tapping ctrl to check if it's awake yet before my screen locker appears. I guess using a Mac for the last ten years spoiled me, because they usually wake up almost instantly.
I haven' yet tried to set up hibernate with a swapfile, but I wonder if that would be the thing I'm missing with regards to battery drain. It would be cool to have it suspend for an hour and then go into hibernation automatically, but I'm not sure if it would help. Might be to glitchy upon wake.
I realize it may sound like I'm complaining, so I want to clarify that I love this machine. The screen resolution is fine. It took me a little while to get it where I like it, but via .Xresources
Everything looks nice in i3.
The build quality is good, it feels nice in my hands, and I am genuinely enjoying using it. It's been a while, and I forgot how much I loved tinkering with a Linux setup and making everything work how I want it.
I don't have a Framework, but I had this kind of problem with my Dell XP. Turns out there is a USB port that will drain the battery if anything is plugged in there - and this is even documented but not very visible in the docs. I had to move my device to a different port to alleviate this. Maybe something similar is going on here?
I don't have mine plugged into anything while it's sleeping.
The only thing I can think of could be if the port pieces, which are essentially form-fitted dongles plugged into USB-C ports in the laptop, are causing the problem even if nothing is plugged into them?
I'm running Xubuntu 21.10 on my Framework and am relatively happy. Mine stays docked most of the time without much travel so haven't noticed the battery issues.
> - There are no stable LVFS firmware upgrades for Linux yet and some users also report overwritten bootloaders after upgrading. Linux compatibility is definitely not there yet.
At the moment, you can't upgrade BIOS on Linux systems without using Windows which is a major shortcoming. No LVFS support is a big miss right now for Linux users
I've measured a drain of ~0.9% per hour on mine when the lid is closed (deep sleep, not hibernating) and I've only USB-C expansion cards slotted in (or no cards at all). When I've two USB-C cards and and 2 USB-A cards it goes up to ~2%/hr. So your large drain might (~3%/hr) might be mostly from whichever cards you've got in.
> This configuration at $1,600 USD is an incredible value compared to a MacBook and other comparable laptops.
I know expensive laptops are the norm in SF, but does anyone else pay that much for laptops they may drop or looked at any time?
(My current workhorse is a T-series thinkpad I got for free)
I'm self employed. The laptop is my main "tool". If I was in construction this would be comparable to a van load of power tools and then some.
I spend more time using it than any other tool I have.
It's also a taxable expense.
So for me the £2000+ I paid for my laptop to max it out to the extreme is money well spent.
Did I mention it's a taxable expense?
where you guys are reselling? highly interested!
I paid $2000 for my laptop, and not a day goes by that I'm not grateful to past-me for that decision. Having a snappy machine is immensely pleasant.
I have a ten year old X220 with an SSD upgrade.... I don't have to wait for anything really, either. Don't get me wrong, I know there is a difference, but... is it a 2000$ difference?
This is about snappiness only, I have no trouble finding workloads stressing this old i5 (e.g. 2k/4k videos), where an M1 would fly through. Personally, I just think snappiness alone is a bit forced argument to spent 2k$.
Build quality is probably night and day. An old thinkpad is going to be clunky and made of cheap plastic. A modern $2000 ultra book is going to be slim and made of metal or other high quality materials, it will be pleasant to type on, to look at, and it will have a reasonable trackpad size by modern standards.
When this is your main tool, it’s much more pleasant to use something with a nice build quality.
I don't know the model he mentioned specifically, but I'd put some old T-series thinkpads against almost anything you can buy now for build quality.
Thinkpads typically have magnesium cases, not plastic, and the older models have rollcages so they can take a beating like no other. I have an m1 air and while it feels sturdy, I have no doubt that X220 can withstand much more abuse than my macbook. Also, I have no doubt the X220 is more pleasant to type on, because older thinkpads have keyboards that have no equal in the laptop space. I would say my macbook has a satisfactory keyboard, but it is clearly a step down from the T series thinkpad it replaced. Still though, the macbook is for me overall the better package than any thinkpad that was in my budget range.
Keyboard is unmatched indeed, but my nipple mechanic is worn now and the trackpad is atrocious. I am not defending this machine, it is 10 years old, the CPU is a major limitation now. I bought it refurbished for 240€ in 2015, so I had a "single use" laptop for a trip... Became my main machine for years to follow. The low price point is one of the best features, as I am not constantly worried about it. Also the original spill resistance saved my ass two times. Oh yeah and SSD and RAM upgrade took me literally 5 minutes. Ridiculous. Almost disappointing as a DIY project.
Tho, I suspect it will perform much, much better cleaned, repaired and thermal-paste reapplied. But I wont open it, as long as I have no alternative set up.
(Also M1 + Linux, would be a no-brainer for me - best bang for the buck right now. I just fucking hate Apple OSs and the ecosystem. Oh and the keyboard too - feels as pleasant an ATM's number pad... Following Asahi closely, still.)
Wow, that seems like a steal for 240€. If you don’t mind me asking, why do you need a single use laptop to travel? Also regarding the MacBook keyboards, they’ve completely redesigned them. They were the worst keyboard I’ve ever used from 2016-2019, but the latest models have a brand new keyboard and it’s my favorite laptop keyboard ever, although that’s obviously just personal preference.
"Single-use" may have been the wrong expression. I meant, I needed something that could break, be stolen, or tempered with at the airport, without me losing trust/access to my main work/digital life tool, and without me getting too sad about it.
I try the Macbooks' keyboards every time I see one in the store. Yes, it has been even worse, but the new ones feel clickity-clackity, like an ATM pad, to me too. No travel, and annoying high pitched noises. Feels weird to rest your fingers on those keys. Not even close to a Thinkpad keyboard, old or new. I could get used to it, but it is not a great input device. Not at all. Trackpad rocks of course, but so does the nipple, if you type a lot.
Ah okay. Fair enough, I could see how the keyboard is very polarizing. I’m tempted to pick up an old thinkpad with how much people rave about them..
Magnesium? I must be confused then, because the “thinkpad” I was imagining was made of plastic. I’ve definitely felt a magnesium laptop from Microsoft’s surface brand before and that build quality was great.
Clunky sure, but cheap plastic? Old thinkpads can take a beating and keep going.
I guess different people just have different needs.
I didn’t mean cheap plastic in terms of durability, but in terms of how it feels quality wise. They can definitely take a beating
In terms of durability, the 10 year old ThinkPad will exceed just about any modern laptop. Size/weight is different story though. Just depends on priorities.
Jeez, I explicitly stated my comment was about "snappiness" only. Sure you can move the goalpost, if it makes you happy, but I am not trying to argue a 10 year old Thinkpad, or any laptop that age, is nicer than a recent Macbook. You win, tiger.
Meh, I just forgot you said that while I was replying
My laptop is a X1 Carbon Gen 3 (released in 2015). Bought it used on ebay and I'm still fairly happy with it. I will say, the lack of hardware transcoding is the biggest pain point these days. When I was still in school and had zoom calls, those were a nightmare and ramped up to a consistent 90% CPU usage. Other than that, Rust compiles times are slow, but nothing else is a huge issue.
Next machine needs to have hardware transcoding though. That's become so important to me. Needing to give up all of my processing power just to be on a video call is awful, but it served well for years for $400.
I live in a country where the average salary is $20k and the M1 MacBook Air STARTS at $1330. Devs here still often use MacBooks or similarly priced higher-end machines. So yeah, good laptops are worth it even in much poorer countries.
My last laptop purchase was 11 years ago. I got recall-upgraded to a new laptop 6 years ago. Once amortised, it's a really small price to pay for something I use this much.
Since it's a business expense, it's even cheaper.
I’ll bet your T-series cost at least $1600 2021 dollars when it was new.
3 years later that same laptop is on eBay for $500 with almost perfect Linux compatibility though. It’s tough to beat used thinkpads on eBay. I never buy my personal laptops new.
Fair point, those things are expensive when new.
But then again, I would never buy those when just released.
I never paid more than $400, but I also use Linux and don't play games or edit video, so I never really felt the need for a modern powerful computer.
Same here. My laptop cost $400 in 2016 and still serves my needs perfectly.
My last laptop was $3500.
I dropped it 4 days after purchase.
My next laptop was $2000.
I dropped it a few months after purchase.
I don't live in SF. So I guess the answer to your question is yes.
If you purchased your laptops with a credit card you should check to see if you have accidental damage protection.
My credit card has that and I've used it to replace a phone with a cracked screen. It was literally a 5 minute phone call to make the claim and all of my money was refunded the next day.
Ah well I had AppleCare+ on both, replaced the first one then bought the current one a year later when the M1 came out because I wanted the battery life for travel. The $3500 computer I sold for a surprisingly large loss even though it was basically in mint condition after Apple replaced everything – turns out that although Apple product resale value is pretty high, the 2nd hand market for a $3500 computer is pretty small. Lesson learned.
- retina screen
- great touchpad (and gestures)
- all day battery
- no noise at all (optional but for me a big plus)
- hinge that stands the test of time
- aluminum (of very resistant) body
i'll buy a laptop that checks these off under $1600, but so far m1 air is the only one I found, and it's cheaper than that
But most of these are meaningless if you don't work full days on the run. Most of us sits in an office with a big monitor and a mouse and keyboard most of the time.
And I don't get the "better productivity" talk either. During the most productive period of my life (by a huge margin, now I understand) I only had a $200 laptop. It was literally sold as "the cheapest laptop you will ever find".
(I give you the noise though, I hate fan noise)
Edit: do hinges ever break? Unless you sit on it, I have never seen this happen even with the cheapest laptops.
Whether you can get away with a cheap laptop really depends on what you’re doing.
Like for someone who works in Photoshop/Illustrator/Sketch/Figma even a little will need a great screen and almost without exception you’re not going to be getting that in a $200 laptop. This admittedly can be worked around with a “cheap” $300 IPS external monitor, but at that point you’re spending more on the monitor than your laptop which feels upside-down. (I know this from experience — at one point I had to do PS work on a $500 Gateway laptop and it was miserable because a quarter of the document’s details weren’t even visible).
Or if you’re compiling code a lot, you’re going to want more oomph than a $200 laptop can provide, because otherwise you’re going to be twiddling your thumbs and getting distracted and breaking flow waiting for code to compile. For me this is particularly impactful, and any reduction in compile times is easily felt.
As for hinges, on even a number of “mid tier” laptops, they tend to get loose and wobbly over time. I’ve seen a number of IdeaPads owned by friends and family suffer this fate.
You can still do a lot of great work and have decent screen by just focusing on second hand stuff. I never bought a personnal laptop as new and right now my main laptop is one I bought from my previous company when leaving. The bonus is usually once you get them they are already fully supported and stable on Linux.
It only costed me 375 euros. It is not a high end model, but a mid range (Lenovo E580) from a few years ago that was specced with the fastest i7 cpu and with as many ram (32GB) as supported. It is not Macbook M1 whatever fast/nice but for me it is still pretty much current. I cringe when people say they cannot work with 1080p screen but the reality is you are perefectly comfortable on them as long as you don't know better. This is something I have learned with different domains, not just computing but also sport equipment, music hardware. You don't necessarily miss the new tech until you tried it and you are not necessarily less productive, comfortable, creative or competitive on older tech, only expectations changes. But once you tested the new drug sure it is hard to come back.
I won’t argue that there’s high value in used machines. I will say that I probably wouldn’t buy a used consumer grade laptop… their specs are often mediocre even when new and their build quality often doesn’t hold up to the rigors of time. I would only give used business and workstation laptops serious consideration.
Personal anecdote: at one point in the early 2010s I was cash strapped and needed a reasonably capable machine. First I tried that $500 Gateway I mentioned, which was terrible to use across almost all dimensions. I ended up returning that machine and instead putting it toward a used Dell Precision M4400 workstation laptop, which was expensive when it was new in 2008 but I got it along with a bunch of accessories off of Craigslist for $350. That thing was immensely better than the cheap 4 years newer Gateway in every way, and much much more usable for work. That machine held out for me for several years until I could afford something better.
As far as screens go, resolution is less important to me than panel quality. I’ll take a 1280x800 screen with decent color performance over a 1920x1080 screen with garbage color. In the case of the cheap Gateway, the screen’s unusability had more to do with it being a terrible bottom of the barrel TN display than it having a 1366x768 resolution. While the Precision’s panel was higher rez (1920x1200) it also handled color much much better which is what made the difference for me.
I hear you
I was very surprised after I bought a nice big monitor and great mouse that I ignored them most of the time, so now I just work off the laptop, it works best for me, which I never would have guessed in my desktop years
That said, I've been very productive with a creaky noisy cheap lenovo with ubuntu so in my case it's not just about productivity but also comfort
Re: do hinges ever break?
they tend to, when you use the laptop 10 hours per day
That's terrible ergonomics and you will grow a hunch. One day you will notice in the mirror and your consequent attempts to straighten your neck will fail. Laptops are such a legacy format, we only endure because we're denied power-/useful tablets, which would be much more flexible ergonomics-wise.
So much this, working off a laptop while having a proper keyboard and mouse near by seems crazy to me.
I have a standing desk and take plenty of breaks, which helps
Standing desk doesn't help with the viewing angle and neck strain. Your upper body should be exactly the same as sitting well adjusted.
I haven't used mouse since I switched to mbp - and I use autocad/rhino a lot.
> And I don't get the "better productivity" talk either.
Among other reasons some people are concerned with energy use in general.
I think you’re underplaying two things:
- not just a great touchpad, literally the best one in the game undisputed
- all day battery life. And not PC laptops that claim 14 hours of battery but see 5 in reality, I’m talking actually 14 hours of battery in real world use.
- decent speakers
I don't live in SF but paid about that much (bit less) for my current laptop. It's more than I've ever paid for a laptop before but I wanted something that would last a good while both in terms of build quality and specs/upgradability. I got a Dell Latitude but the latest Thinkpad T-series was running about the same price with similar specs. I live in the UK.
I have several laptops that are multiple times that. It's common for people earning $200k+ per year to spend 5-10k per year on computing hardware.
My desktop was like $14k new. It has been my daily driver for over 4 years, so less than $300/month (and still going down). That's a rounding error in relation to the money I have earned using it.
My private laptops are around 1000 to 1500. But I also have my desktop PC.
My company laptop is 3k$ and I actually carry that around more often. True I care less about it but I still don't feel like I'm dropping devices while I'm moving.
I dropt my Thinkpad primarily at home when it was sitting next to the couch :')
Yes, since I use it about 8 hours a day for a couple years, the cost per hour is pretty low and the better screen, keyboard and speed is worth it.
However getting accidental damage protection, like from Lenovo, does help the peace of mind.
I'm still amazed that people prefer to work on a laptop over a desktop.
I mean, I still use a laptop in the evenings, but for work a desktop is to me a must.
I think it depends what kind of work you do and if you travel a lot. Pre/post pandemia are also different for a lot of people. A lot of people are now working 100% at home in a dedicated office and could only use a desktop while they used to move from meeting to meeting with their laptop prior to the pandemic. Others were travelling a lot from conf to conf and customers. Some still do a lot of customers on site work. There are many use-cases.
Anyway a laptop becomes a desktop by plugging it's usb-c port to your docking station so there is very little downside appart from being subject to throttling when doing heavy cpu stuff to using one instead of a desktop. Which not everybody is subject too, mostly only people compiling stuff on their laptop.
I personnally spend most of my time in my office at home with dual screen but I like to move to the living room when attending conferences, in my rooftop when it is not to sunny or in the kitchen to have a chat with my gf while grabbing a tea when she is at home and I am waiting for a job/whatever to complete so a laptop is still nice to have even when working 100% from home.
For my personal machine, I’m fine with a desktop because it’s unlikely I’ll need that power portably, and if I ever do I can set things up to remote into my personal desktop with my personal laptop.
For my work machine, a laptop is preferred because then I can do my work anywhere without fuss even without an internet connection.
Besides, with an M1 Pro, plugging my work laptop into the Thunderbolt hub at my desk provides a near-desktop experience for the overwhelming majority of its usage. For day to day work the difference between my M1 Pro laptop and 5950X desktop isn’t staggering enough to justify giving up portability.
As someone who shifts between 3 different desks (home, main office, project office) and sometimes as to go and see clients, having a laptop is pretty necessary. However once I plug my laptop into its docking station connected to my monitors, keyboard, mouse etc. I really don't notice much of a difference between a laptop and desktop.
But I agree that working all day on a laptop in laptop mode is right out.
I have desktops at home and at office that i keep partially in sync (my home desktop is also for gaming (steam linux), no need for that in office).
I have a laptop for those rare cases when I need to work where i am not at either of those locations.
With so capable phones in pocket, I rarely find myself needing to pull out a laptop to the point that I don't think I'd buy replacement if this one broke.
But to each his own.
I use a desktop into which I SSH or RDP from my laptop. Giving my flexibility and performance. But I still have an extra monitor in my office and home which I almost always use.
All the laptops I've bought have been $1-2k. It's not really that much, sub $1k laptops are pretty shit.
There’s a lot of solid laptops in the $600-1000 range. Sub $600 is generally pretty shit but there’s some sales and exceptions
When you spend so much time in from of the laptop, even 2k doesn’t seem excessive.
A $6k laptop seems reasonable if you are using it for earning SWE incomes ($200k+/year).
I charge 1000 Euro per day when I'm consulting. That would be in Berlin. For that money, I need a laptop that is fast and works. I spend most waking hours using it so it also needs to be comfortable to use. So, I value quality and performance.
That being said, my expensive 2017 Macbook Pro broke a few weeks ago and I picked up a very nicely discounted Samsung Galaxy book for about 700 Euro because the only macs available were obsolete and the new ones had 4 week delivery times. It's a 16GB, i5 with a surprisingly decent Iris XE graphics system on a chip thing and 512GB ssd. Nothing fancy but good enough for the money and great value at that price actually.
I put Manjaro on it and was up and running in a few hours. I only had 4 hours to configure it in the evening as I had a customer meeting the next morning and needed a working development environment. So, I was pretty pleased all of that worked out fine. That means that laptop earned itself back in under a day. I even managed to revive some of my 32 bit Steam library and get playable fps (with that intel XE graphics!). And unlike the mac, it does not have thermal throttling issues. It was always unusable for gaming.
I'll probably buy one of the new fancy macs when delivery times improve a bit and when we get some clarity on the inevitable early adopter issues (which bits are going to break this time?). But I'm not in a hurry. This thing works well enough and I'm kind of liking Manjaro so far. Lots of rough edges but nothing I can't deal with. Definitely not for users not comfortable using a command line.
I'd totally spend 3000-4000 Euro on a laptop regardless of the OS if it is good value for money. My last mac book pro was 3500 Euro. Money is not the issue for me when I literally spend a lot of my billable time waiting for this thing to do stuff for minutes on end many times a day and interrupting my flow (which is priceless). But I expect performance for that money as well as a decent keyboard (so definitely not my Macbook Pro) and a nice screen. Apple fixed all of those things with their latest iteration but the configuration I'd want puts me close to 4000-5000 Euro this time. I'm actually considering doing that. But it's a lot of money and Apple tax this time. And the build quality of the last one was terrible, which makes me more hesitant. The worst I've ever seen from Apple. Hence, the cheapo emergency replacement.
So, this laptop looks like a great deal. Twice the memory (or more if you upgrade), ssd, and better CPU than what I picked up for just 700 Euro a few weeks ago. Close to good enough for me. I'd like a proper screen though. Not having retina/HDPI feels like going back to the stone age a bit. Yes, Linux needs a bit of work on that front but Wayland seems capable of this at least. But other than that, not bad. If they do a more expensive version with a better screen, I'd consider getting one.
Finding decent laptops is actually a problem. There are not a lot of premium laptops that are nice enough for my criteria. I don't want a gaming laptop monstrosity that sounds like a vacuum cleaner. I don't care for Thinkpads with a nipple and shitty touchpad. Dell has support issues though their XPS is pretty nice. But I want something premium, with a nice screen, keyboard and touchpad that doesn't feel like a huge compromise. This Samsung is quite nice on this front actually considering the price. The touchpad could be a bit bigger. But it does multi touch and it works (after lots of fiddling with settings). The 1080p screen feels like a huge leap backwards after having used a Retina screen for the last few years. Not loving that.
I'm surprised by comments about how it didn't work ok for some people. It worked almost flawless for me, maybe because I chose Fedora 35?
One small thing I loved about the laptop was how I was able to (albeit slowly) charge it with a pixel 3 charger.
I as well am very surprised by how many people seem to have had issues with it.
I have used Manjaro and Pop OS on it, and encountered basically 0 issues on either one. The only issue with Pop OS is Gnome’s support for fractional scaling is pretty terrible(at least at the time I tried it).
I use Cinnamon on Manjaro and pretty much everything works flawlessly. The only thing I had any issue with was the fingerprint scanner, though I mostly just gave up on it at the first sign of issue as it wasn’t a feature I really cared about having.
I haven’t had any issues with battery life, and I just carry an Anker portable battery with it so it’s not really a concern anyways, and having that gets me charging for laptop, phone, or anything else that can use USB C power.
Even still, I’ve observed as high as like 8 hours of battery life while being actively used and that’s more than enough for me.
Might be worth noting I don’t primarily use a laptop though. It’s almost exclusively for when I’m out somewhere and an issue comes up or something.
Heads up to anyone who is running into the right click issue on Ubuntu relating to the framework laptop (right click acts a bit funky), I added the following lines to /usr/share/X11/xorg.conf.d/40-libinput.conf
This fixed the issue for me. Your mileage may vary though.
Framework is remarkable. Intel MacBook Pro power in MacBook Air footprint.
The build quality is far better than first gen devices, if not second.
Given enough time, it might be able to get many people off macOS.
The screen is amazingly crisp if you can overlook the fun resolution. Hard to look past 3x2.
Why did I return it?
I realized I needed to begin with a retail and turnkey Ubuntu experience where everything just worked and the things that mattered to me worked out of the box.
Framework is not quite an optimized retail Ubuntu experience yet like a Dell XPS with mature drivers largely ready to go.
- Battery life isn’t optimized and maximized out of the box, you will have to tpm. I have limited time for this at present. I’m sure there’s lots of interpretations that is not a a big deal. I’d rather be solving problems with the laptop than be solving problems in the laptop. I want max battery life without investing hours up front.
- Fingerprint reader required manual setup
- Wifi can have hiccups on the latest Ubuntu, and without trying into optimizing battery and kernels upset instead of getting things done.
- Touchpad is so so. Good hardware, maybe more tweaking in Linux needed. Very used to macOS too.
- HDMI port draws extra battery life when not in use so you have to to keep it removed.
The components otherwise seem high quality and well put together.
Again it’s not that these issues can’t be overcome, or that they won’t be out of the box in the future, I simply don’t have the time for either at present.
It was a joy to use for browsing in Ubuntu 21.04 as stuff is broken in 21.10.
I got mine a few months ago, and overall I'm rather pleased with the purchase, however I don't think I'm quite as satisfied as the author of this article. For the most part, I think it's a really great machine, but there are some oddities that have caused me trouble. I've been running Arch Linux + Plasma on it, and have gone between X11 and Wayland (more on that later).
- My touchpad is really flaky. Sometimes it works perfectly, but most of the time simply clicking just... doesn't work. Moving the cursor, tap to click, scrolling work fine, but the actual pressing the touchpad to an audible click won't result in a click on the machine. This seems to be a hardware issue, it can sometimes be remedied by pressing down across the whole touchpad, but not always, and it seems to be worse when it's cold in the room. Also, once in a blue moon it won't initialize quite right when waking from sleep causing it to be hypersensitive; putting the device back to sleep and rewaking it fixes that.
- They mentioned they were fond of the battery life, which surprises me as that seems to be the devices biggest failing. I'm somewhat new to Linux laptops, but even after setting up hibernate and deep sleep and all that, I simply can't leave it on sleep over night or it will drain the entire battery. The battery life while using it is fine, not as good as my old MacBook, but I've had to start shutting it down when not in use.
- The speakers really aren't great. A big part of that is the fact that they're downward facing, which seems a really bizarre design decision to me.
- Fractional scaling. This is more of an issue with Linux itself. I actually really like the display as a whole, particularly the 3:2 ratio. However, 1x is too small, and I found 2x to be too large. I've been using 150% scaling, which worked well for the most part, except some applications wouldn't obey it (cough Steam). I then tried switching to Wayland, which seems to have fixed Steam's issues (and the trackpad scrolling notably improved oddly), but xwayland programs have really blurry text.. There's always a bit of a tradeoff, although I'm still experimenting with this. I also find this display gives me more eye strain at later hours than other screens I use. YMMV.
Overall, I would say I am satisfied with my purchase however. I love the keyboard, love the modularity, I like the customizability. I don't regret the purchase, but I do think there are some things potential buyers should be aware of.
We've finally root caused physical clicks not registering on some units, and found that it was a batch of dome switches with a coating that was more likely to corrode over time depending on environmental conditions. We've recently switched coatings to prevent this from happening in the future, and we're also sending replacement parts to folks who write in with this issue. One of the benefits of this product being easy to repair is that we can just send you a replacement module instead of needing to take your system back or send you a full system!
Seconding your experience, except my battery life has been quite good somehow. I’m not sure if that’s due to one of the settings I changed when setting it up, OS (I’m using Pop!_OS), or just luck of the draw.
My touchpad will also just stop clicking sometimes. It’s usually an exercise of finding the right spot to push to reset the hardware clicker itself. I have the fix that Framework sent out about a loose touchpad cable, so it shouldn’t be that. It also occasionally will just stop registering the mouse at all, either from the touchpad or from an external mouse. Closing the lid and reopening it will fix this.
Fractional scaling was a little dicey, but I found 150% hits the sweet spot for me, I don’t have any issues aside from some occasional jagged screen regions while scrolling, which I don’t mind.
Overall, I’ve been happy with mine. It’s not perfect, and it occasionally hiccups in weird ways (like the mouse thing), but that was my expectation going in. As a daily driver and dev machine, it works well and I would recommend the DIY edition to anyone with a little tolerance for issues in their laptop.
Yep it works well. As is usual with linux, it took a frustrating day or two to set up but I now play Steam windows games with Proton on NixOS using an NVidia eGPU. I'm using X, not wayland though, but it works both on external displays and on the Framework's "internal" display.
It's an Intel Xe GPU, so it's probably integral to the CPU package.
I have a batch one Framework laptop with the i7-1185G7. Overall, I absolutely love it, but it didn't come easy.
1. Ubuntu 21.10 had horrid support of the hardware. Eventually switching to Fedora 35 fixed most bugs
2. The DisplayPort expansion card I received was faulty and sometimes caused horrible performance.
3. EMI shields were incorrectly placed on my expansion slots[^1]. Eventually I simply removed them to get full performance back.
In the end, after months of tweaking and talking to support, I got a new DisplayPort card, switched to Fedora 35 and everything works smoothly.
I'm sad about the speakers if they are as flat as they said, even more if they are not upgradable parts. MacBook speakers are really good for their purpose and I even consider them to be one of the importants things I trully love on macbooks (with the screen and the touchpad).
Anyway, I'm still looking to buy a Framework laptop soon but I'd love to be able to test one before buying it.
Most modern laptops use DSP to get the speakers to sound the way they do. If you switch that off they have the same ‘flat’ sound described in this review. Wonder if this can be solved in software ?
This is one of those things where Linux is a decade behind everything else because nobody has bothered to solve this problem yet. Yes, most laptops require DSP for the speakers to sound the way they're supposed to, and Linux has no way of handling this. The Windows driver provided by the OEM would normally do this, AIUI.
While you can certainly manually add DSP to your system (PulseEffects, some random JACK or PipeWire flow, etc. - I do exactly this for headphone calibration on my desktop), what we need is a standardized mechanism for storing speaker profiles and having the audio subsystem seamlessly and transparently apply the right EQ and crossover profile for your device (there are usually more than 2 speakers). This is already a thing on e.g. Android, but not desktop Linux.
This is required for MacBooks too. For Asahi Linux, once the underlying sound hardware works, we'll be spending some time working out a sane way to do this; I consider it a mandatory feature in this day and age. And once that's all done, everyone will be able to use it to get proper sound on any laptop as long as they can obtain or make an EQ profile.
For the Macs, our plan is to measure impulse responses of what macOS does (we have the ability to snoop on its hardware accesses, so we can just play a test sound and capture the raw samples going to the speakers) and use them on Linux. This should give identical audio at moderate volumes. I'll have to check whether it also does some kind of compression or volume-dependent EQ worth replicating or emulating too, at higher volumes.
You can handle it with Pulse (qpaeq), but it's pretty janky.
That's doing it manually; there are a million ways of doing that, but the point is it should be transparent and automated and select the right EQ profile for you based on your hardware.
> For the Macs, our plan is to measure impulse responses of what macOS does (we have the ability to snoop on its hardware accesses, so we can just play a test sound and capture the raw samples going to the speakers)
They do some sort of crosstalk cancellation to make stereo sound better, it'd be interesting to look at those impulse responses and see that happening!
Yes, the terrible sound on my Lenovo x1c is the only driver-related issue I have with this laptop on Linux. In all other aspects it is either same or superior experience compared to Windows.
The X1 Carbon 7th edition has pretty good sound with Linux. It does require quite recent kernel versions to get the good driver.
The speakers are replaceable (and therefore upgradeable), but like a sibling comment noted, most of the magic in speakers in portable consumer electronics products comes down to DSPs and smart amps. We started with a CODEC that had built in smart amp and DSP functionality, but weren't able to get the necessary support from the supplier to enable it. It ended up not mattering anyway, because we had to switch to a different CODEC without that functionality due to the global silicon crunch.
This is absolutely something we want to improve in the future. We're actively investigating other audio parts we can use that have better capability, but equally importantly, supply availability and software support.
Thank you for your detailed answer on this topic.
It’s wonderful when companies listen to (potential) customers.
You are doing an amazing (and important) work.
I wish you all the best and you are on my top list for my next laptop (although, my current one is far from dead).
And no mention of the touchpad quality either, in the review.
When I purchased mbp m1 pro I didn't expect the speakers to be that good as most of the mbp reviews were about processor, ports, and screen. I don't think I now need an external music system for indoor listening.
To be fair, it’s not unique to the M1 pro. All macs of the last decade at least have pretty good speakers (for the size, that is). Just what’s needed to don’t need anything else to listen music on the go when you have no other solution or to watch movies in the bed.
Not of the last decade, their speakers have really only been top notch since the 2016 redesign, and even more so starting in 2019 with the 16”. I’ve had a 2012 MacBook, a 2015, 2016, 2019, and now 2021.
Ideally speakers should be flat. If you want effects they should be done in software.
I suspect the confusion is between flat as "linear response across ear bandwidth" and "total lack of depth", eg, the later isn't about being linear, just low quality speakers with very inconsistent frequency response.
Yeah, there's clearly confusion here. The article and GGP are using "flat" to mean "bad". The GP is using it to mean "flat frequency response", i.e. good. Unfortunately, all small speakers sound awful uncorrected; there's no way to make a flat-as-in-good laptop speaker. The question is how would those speakers sound with software DSP for correction, which is standard on other OSes but not Linux.
You're misunderstanding the situation.
Who listen anything on speakers anyway?
When travelling I will always use headset, when hooked to docking it uses my home studio monitor speakers and if I want to listen to watch a movie in the bed or a show in my kitchen it is paired to my Bose bluetooth speaker.
I didn’t realize it had a hardwire switch for the webcam, that’s a nice feature they could put on the front page. I see it now on their “learn more” page.
It would be nice if the expansion cards came with a plastic cap to protect the usb c plug, I wouldn’t want to toss the spare hdmi card in my bag as is, but in the spirit of open hardware maybe one will pop up on thigaverse.
Yep! It's a hardware cut-off on both the camera and microphones.
> It would be nice if the expansion cards came with a plastic cap to protect the usb c plug,...
Have your pick of these . There are expansion card holders/cases, which might also be an option.
I'm evaluating reports of Framework+Linux-distros' sleep/wake/hibernate impacts upon video resolution switching, WiFi stability, and trackpad stability, to determine when to pull the trigger on purchasing a fully tricked-out model. I can afford to run an Apple M-class laptop (holding out for next batch) and a Framework at the same time, and make the effort to transition off the macOS ecosystem (any consumer-grade and office apps I want not available on Linux, I'm happy to run under Windows within Linux).
The straw that broke my camel's back was the latest macOS update that deleted all Safari tabs. Around the same time, iOS update did something similar: deleted all back history on all Safari tabs. After the many other problems I put up with in macOS, I figured if I'm going to be spelunking around the system to fix Apple software problems as frequently as I have found myself doing in the past decade, I might as well work with others on the source, as Linux is getting awfully close to "good enough" quality these days for my daily driver, at least enough for me to give it a fair shot with sustained effort on peer hardware as my Apple gear.
I also recently(~1 month) purchased one of these, (DIY edition + windows). So far my only complaint is the screen size. I was really wanting a 17" touch screen and a keyboard with a num pad. My first thought when I was opening + working with this was that the whole experience felt like a labor of love.
I love the concept of the Framework and it fits my use-case but I really wish they would at least add some form of screen configurability/options to their tech roadmap.
- I need a mate screen option, glossy displays in ambient light give me headaches.
- Some folks will also probably want touchscreens (uber glossy), more power to them
- Would also be very cool to bring back the old tech of transflective displays like the one in the Pixel Qi display. For folks that want to work outside (me again!) and are not fussed about color accuracy.
- TrackPoint pointing stick (red thing in the middle of the keyboard)
- Magnesium top display cover &
- USB-C charging
- Smartcard reader
- Ports including USB-A and -C, SD, HDMI, Ethernet, and audio jack
- Security lock slot
- Thunderbolt 3
- Docking station (latest Ultra) that works
& Parts swap
It's massively upgradable. 8 captive screws, and the bottom case pops out, the keyboard pops out too. ~10 screws, and the motherboard and everything else is out. People make frakenpads because the parts stayed so similar over the years. All the screws are Phillips. No hunting for weird Torx bits.
For me the appeal is that when a feature that I don't currently have becomes available it's straight forward for me to upgrade.
When the next iteration of motherboards come out, I won't be buying a new laptop, ideally I will go here https://frame.work/marketplace and just order the one that I want and follow the video guides to install it.
The potential to customize this is huge, larger than any other laptop that I'm familiar with. The ability to self maintain, to fix every piece of the hardware if necessary just makes this thing a win for me.
It remains to be seen how much the promise of upgradability turns into a reality.
I typically buy replace my laptops or desktops after around 6 years. That is a long time to expect things to remain compatible with a chassis.
My issues with Thinkpads:
- They seem to be getting worse for repairability and build quality with each new generation. Everyone who raves seems to be running older generations (yours is comparatively not too old).
- Trying to buy new is an exercise in frustration because of their ridiculous pricing. Either get gouged outside a sale paying a huge premium over competitors, or wait for the right sale and get a competitive (but not cheap) price.
many Thinkpad models in similar sizes are also also limited to 16GiB of RAM.
And while the are up and downs with the quality there seen to be a general downward trend.
Like my previous Thinkpad didn't had any major defects after the years but with my current one after not quite the years the mousepad broke (the coating is damaged getting slowly worse starting as just an annoyance to now frequently registering phantom middle clicks).
Similar while using custom secure boot platform keys have been working with Thinkpad T series for years, for the lastish T14 series a non distinguishable subset of them gets bricked by it.
I've been a ThinkPad user in the past and was using a ThinkPad as my last laptop before starting Framework. In many ways, the Framework Laptop is a modern take on where ThinkPads came from. The T480 is relatively recent, but even there, the current generation of it (T14 Gen 2) loses one of the SO-DIMMs and the swappable battery.
If the author is here, please add to your review: trackpad; fans noise; heating; Bluetooth stability.
Honestly, the framework was one of the better purchases I have made. There were definitely some Bluetooth hiccups, but all other functionality worked perfectly out of the box on Arch. The wiki included a few key tips.
One thing to note, is that I had a bad time with tlp. I would suggest using auto-cpufreq if you want better battery life.
Lots of people seem to complain about Ubuntu support on this machine. There seems to be a consensus that Fedora is the option that works best for that machine.
I noted this elsewhere, but we've been providing the team at Fedora with hardware, which has helped make Fedora 35 work great out of the box. We've also been doing that with some smaller distros like Elementary. That isn't a relationship we have with Ubuntu at the moment, and it shows, though Ubuntu 21.04 works well: https://community.frame.work/t/ubuntu-21-04-on-the-framework...
> I have a machine with better specs than a comparable MacBook Pro M1 for less.
…except for display quality, performance and battery life.
At least Framework placed the Power button NOT in the keyboard top-right corner (oh don't let me start on this).
But all other keys... does nobody ever uses Home/End/PgUp/PgDown/Insert/ContextMenu keys these days?
Ask HN: Nobody ever uses Home/End/PgUp/PgDown/Insert/ContextMenu keys these days?
I understand what the current laptop keyboard layouts are made by alien^W someone who is clearly not using the computer as a work tool, but we are close to a decade of this ...nonsense.
All Fn+ keys REQUIRE using both hands. It is all fine until you are somewhere where you need to use one hand to hold your notebook and ... you can no longer press any Fn+ combo.
I have a pretty big hands and I can't stretch my fingers enough to reach both bottom-left Fn and any key on the right side of keyboard. On a 14" laptop. Which could had all the keys (compare X220 and T440 for example) but it doesn't.
Thinkpads are still have PrtScr on ContextMenu, which means I would 100% lose contents of the clipboard some hours later. Who needs PrtScr /there/ anyway? I want to hear the reasoning, like "we noticed (through our telemetry? how?) what our users are making screenshots all the day so we moved the PrtScr in the place of ContextMenu button"? You can replace the Space with PrtScr then, it would be pressed even more.
16" notebook? You will still have a sub 12" keyboard on it 
14" notebook? You won't even have the chance to see it's keyboard - you can see photo of it's back (like this is somehow important in the age where is nothing on a laptop back) but you can't see it's keyboard.
You may want to consider a Mac. Most of its software doesn't use any Fn+ combos, or Home/End/PgUp/PgDown/Insert/ContextMenu keys. They have other shortcuts. It's a pain to learn them in the beginning, but in the long term it gives you an idea that maybe having many single-purpose keys on a laptop keyboard is not a best idea, ergonomically.
Sorry but I do use navigation keys when writing text or code.
> They have other shortcuts
Thanks, I prefer two dedicated hardware keys to delete text, instead of one and Fn+Delete.
I tried it for a few weeks but had to ultimately send it back. Tried Arch and Ubuntu, but even as a developer who's very comfortable using Linux for servers and whatnot (moreso than consumer windows, honestly), I just couldn't deal with the battle to accomplish anything on desktop Linux.
Want fingerprint support (which is supposedly first-class)? Spend an hour fiddling with drivers and configuration files until it works.
Want to download discord? Spend an hour trying out the different methods of installing it until you find the one that works acceptably for your installation.
and so on. It was not uncommon to wake it up from sleep only to have it kernel. That's not acceptable when I'm trying to take a timed quiz in class.
And yet I still think it's a great machine. It's just not for me- someone who's purchasing as an enthusiast or looking for a "project laptop" would probably love it. I need a laptop with high availability though, and I just don't think desktop Linux is ready for that.
This is so over the top enthusiastic I hope he is getting paid.
You can swap battery, harddrive, memory, network chip yay.
That was possible on nearly all laptops some years back.
At the enterprise level both Dell and Lenovo have models
that does most of this. (Unless recently discontinued)
Quite a few laptops had "bays" where you could swap out
extra harddrive, cd rom, extra battery,
Then a while back computer had PCMCIA/PC Card, that let
you add all sorts of very useful hardware.
All in all this is far less user configurable than a two
bay, to pcmia slots. (with battery, memory easily upgraded).
I would love someone to come out with a laptop like that again.
Further modules for the framework are expensive, made by a single
company, few variants. If the company fails you are stuck.
On the old style laptops, ram, memory, cd player, extra hard drive,
batteries, PCMCIA/PC Cards were all usually available from several
sources. Bays might be the most proprietary.
> That was possible on nearly all laptops some years back.
Not necessarily in this size class/with this chassis design. "Ultrabooks" as a class haven't been nearly as user-serviceable as you're suggesting.
> made by a single company, few variants; Bays might be the most proprietary
Eh. I'm not terribly worried about this given that they're TB3 or TB4 (to wit: PCIe) and provided with reference designs.
I have to admit, though, that I couldn't easily find IBM's old expansion bay spec or proprietary connector... or for that matter, HP's, or Dell's, or... well, yes. Bays sure are proprietary. To extoll the virtues of somebody else's bays and in the same breath deny these is at best misplaced nostalgia. :)
Quote: "If you need to do native iOS or macOS development then a MacBook is your obvious choice."
No, it's not. I do cross-platform development for over a decade now and all I have are virtual machines. Over 2 dozens of them and I never owned any Apple hardware. Why pay for overpriced hardware when virtualization covers everything for a 100th of the price?
Honestly I think you should be able to test stuff in real world scenarios if possible.
I would never do iOS development seriously without native devices.
Virtualization brings you far, true, but it's still different.
I would even create a hardware lap if I would do end-user device Software development.
And the quote is still true. If you do iOS MacOS development it IS the obvious choice obviously.
Yours might be cheaper and might have other advantages but it is NOT obvious.
If I remember correctly, Apple's EULA prohibits users from running virtualized macOS on a non-Apple hardware.
Except I am not an user, I'm a developer, hence no EULA infringement. You might want to take a second look at their EULA for their Apple developer ID and you'll see that for those $100/year they sing a very different song. Money talks!
Except I am not an user, I'm a developer, hence no EULA infringement.
IANAL, but it doesn't work like that. Like any original creative work, macOS falls under copyright protection. Apple only gives you a licenses you to copy/run macOS on Apple-branded hardware.
You might want to take a second look at their EULA for their Apple developer ID
I don't see anything in the Apple Developer Agreement that would cancel the terms of the macOS EULA. But I'd love to be proven wrong, please give me a pointer.
you'll see that for those $100/year they sing a very different song. Money talks!
$100/year is less than pocket change for Apple. Even if they have a million paying developers (unlikely), it's ~1/3658th of their yearly revenue. If they'd make their iPhones 50 cents more expensive, they would make more money.
The $100 per year is only there as one of the many ways to fight scammers. It makes it more costly to open many accounts, they have a credit card (with a name) on file, etc. Sure, it is not air tight, but it is just one of many measures.
No, it's still prohibited. From Apple's terms for devs:
"You agree not to install, use or run the Apple SDKs on any non-Apple-branded computer, and not to install, use or run iOS, watchOS, tvOS, iPadOS, macOS and Provisioning Profiles on or in connection with devices other than Apple-branded products, or to enable others to do so."
Developing natively for iOS without testing on any iOS devices sounds like a shortcut to making mediocre iOS apps.
Unfortunately I didn’t see any comments other than one in passing about the trackpad’s feel and software integration. At the moment I consider macOS to still have the edge over other implementations (software, the glass feel and haptics all in harmony), but I do know that in recent years the precision touchpads have come a long way.
The trackpad is top-notch for something not a MacBook
I legit taught myself basic Linux terminal since when I was 17 I couldn't afford a windows license.
Half the fun, at least for me was always liberating an old Windows PC.
The pricing on this isn't really competitive with buying a mid Windows Laptop + seeing if Ubuntu runs. Such fun to give up on Ubuntu and use Fedora since Fedora was supported better.
This is such an interesting laptop. Hopefully an AMD based version with a more premium screen will be released.
Yeah the AMD APU laptops are absolutely amazing. Be ready though, their Linux support is quite possibly the worst I have ever experienced in almost 30 years of using Linux. I'm typing this right now on a first gen Ryzen APU laptop (Raven Ridge) and it took more than three years for enough kernel bugs to be fixed that plain old Ubuntu LTS wouldn't constantly crash or hard lockup randomly. There were straight up silicon bugs and major sleep state issues that had to be sorted out and worked around with this early generation of chips. It's finally a great machine but it really soured me on the quality of AMD GPU code in the kernel--apparently it is a huge black box even to folks like Linus and can be quite the bug farm.
Be ready though, their Linux support is quite possibly the worst I have ever experienced in almost 30 years of using Linux.
So it's not just me :)
I bought a Ryzen APU since everybody kept telling me that AMD GPUs where the best when it came to Linux support, and it has been a massive pain. It finally runs almost stable using the latest Pop_OS, but it has been a not fun journey to get here.
i've been running a ryzen 5 2500u laptop for the last 2 or so years. no issues at all getting things running with the various distros i've used
Oh the “joys” of Linux. Precisely why I use a Mac. However everyone and then I get tempted.
So the problem with AMD and Framework is that the Framework uses Thunderbolt4 as the basis for it's expansion ports. Something that AMD doesn't natively come with.Not to say that it couldn't happen but I'm guessing 12th gen intel before AMD
To add - AMD will have to support it on their mobile processors, so it will be a while. As far as I understand, AMD's current desktop CPUs do not support it, and the support will probably have to start on the desktop.
Are you sure? The adapters are all covered by USB-C and its DP alt mode. Even the storage expansion cards are attached via USB and not PCIe. No Thunderbolt required.
the zen3+ 6000 CPUs also won't have thunderbolt support:(
Looks like they will have USB4 though, maybe just no Thunderbolt 4 certification yet? As far as I understand, frame.work is capable of TB3, but just not certified (yet?)
> The screen has a 3:2 ratio, high-resolution, and super bright colors.
No. Only in backwards anachronistic PC land is 2256x1504 "high resolution" in 2022.
The 14" Macbook Pro is 3024x1964. My XPS actually has 3840x2160.
Please kill this meme. 1500 pixels tall is not a high res screen.
Last time I checked there was no EU distribution. Anyone's aware if there are any plans ?
They are now accepting preorders for UK Germany and France .
Looking at the first photo, with all the dongles in their own little boxes, and the fact the author bought one "just in case", highlights for me that USB-C is a failure. It has introduced massive technical debt into the industry as a whole. Instead of companies all moving gently towards USB-C ubiquity, we have major players* choosing to solve the failings of USB-C on behalf of their users. This is a patch. It'll be 30 years or more before we're over this hump.
*=Apple and Framework. I'm calling Framework "major" because they have the ear of the industry. They're a computer-person's computer company.
What would USB-C being a success look like? Everyone throwing away every single device that has an old USB port immediately on day 1?
I am not OP, it a couple things:
* Power and cable clarity. For most folks there is no real way to pick up a random power supply, and a random cable, and have any idea how fast your device will actually charge. Or if it will at all. Generally plugging a laptop into a charger that doesn’t supply sufficient power ends up in battery drain while plugged in.
* A realistic hub. Every try to find a USBC hub? The best I’ve found without paying multiple hundreds of dollars includes a usbc port for power, and a usbc port for a device. I’ve yet to find a reasonable device that allows me to plug 4 usbc devices into one usbc port.
It would look like lots of new devices NOT shipping in 2022 with micro USB (and even mini USB!) and HDMI and barrel power connectors. But unfortunately they still do.
Curious- what's shipping with micro usb? I had to find a micro usb cord to charge an older device a few weeks ago and probably have 1 or 2 cords compared with 15 or so USB C cords (and I've never bought a USB C cord outright- those are all from products that shipped with C).
As for mini usb, the only thing I own that uses it is my TI-84+ calculator.
I really hope there will be some competition to the Apple M1 chip soon. All other laptops seems bulky, noisy and powerconsuming compared to the MacBook Pro M1 I've been using for a year now.
I have one of these laptops running an arch based distro. Some of the hardware was buggy especially wifi/bluetooth and the intel GPU had some issues with KDE but all of the issues are fixed. Still requires a reboot here and there but pretty much all computers I've had do at some point. Battery life isn't great. I'm not unsatisfied, but overall I wish I spent $500 more and bought an M1.
Anyone wanting to upgrade their laptop in the near future should probably wait a few months, because new models that were announced at CES will show up with the next gen Intel or AMD CPUs that will hopefully be worth the wait. This looks like one of those "generational" upgrades that we have only every 7 years or so.
I'm not sure if the Apple M1 is going to be beaten (especially in the efficiency department), but challenged for sure. Will have to wait and see.
I have one of these laptops running an arch based distro. Some of the hardware was buggy especially wifi/bluetooth and the intel GPU had some issues with KDE but all of the issues are fixed. Still requires a reboot here and there but pretty much all computers I've had do at some point. Battery life isn't great. Overall I wish I spent $500 more and bought an M1.
My has been a dream with NixOS on it. A great dev laptop really. The Intel GPU is enough to play around with graphics programming too.
What kernel do you use on NixOS? Bluetooth connection was spotty for me until I used 5.15.12.
What is even the appeal of such a device. Linux runs as smooth as it gets on Thinkpads and those are also pretty upgradeable and components can be replaced if required. Maybe not onboard Wifi or GPU but pretty much anything else. So why the fuzz about this laptop?
I just wish there was an option for having separate touchpad buttons. Not being able to clearly feel the dividing line between left and right click is a dealbreaker for me with any laptop.
Nowadays most notebooks have multitouch touchpads and thus right clicking is done not by tapping in a specific button, but by a gesture (e.g.: tap anywhere in the touchpad with two fingers).
Which is inevitably slower and clumsier than mousing with your index finger while keeping your thumb laid across the mouse buttons.
how do you even manage to that with a standard mouse?
The new Thinkpad, imo. I really like most of what they've done with it, but I'm just not into that keyboard... It doesn't look too usable, although people are saying it's good.
Long time thinkpad user here - my previous was a W520 specifically for that older style keyboard. I am not a fan of Chiclet keyboards but I put the Framework keyboard up there with Thinkpads / OG Pixelbook. Too bad they're not out at a best buy for you to walk up and try...
how is the battery life?
I run the framework with a i7-1165G7 and Ubuntu+i3wm. I have had this laptop for ~3 weeks now. Battery life for me is running around 5-8 hours with full brightness, music streaming, and somewhat active use. It isn't fantastic but it is pretty reasonable. I haven't conducted any serious measurements but I'd say the battery life is pretty solid (and the screen is pretty bright). (This is all anecdotal though, If you really want serious numbers, I'd wait for some tech reviewer)
Same question here. I’d love to ditch my MBP for this, but I rely on portability heavily and will need something that will last a work day.
Unfortunately it’s the Achilles heel of this device. I was getting 3-4 hours of battery on Linux web browsing and VSCode
This looks like a great piece of kit.
I’d usually be wary of dropping that much cash on a product from a vendor without a track record.
How well backed / funded are these guys?
I have a couple of years old Thinkpad X1 and the WiFi issues are pretty much the same on Ubuntu. Drops out after a while.
It looks like preorder is open in Germany now! Not in my country but it's a bit closer to home, maybe something can be arranged..
Yes, UK, France, & Germany since 17 Dec. I noticed because I was checking every day - bit surprised then to only get the email (I'd also subscribed for updates/let me know when available) yesterday (4 Jan).
I went ahead and had mine shipped to a forwarder in the US, who then shipped it to me. Depends on how much in a hurry you are and whether you can find a forwarder you trust.
and if you are willing to pay import duties.
Yes that's a good point, especially since you're paying them twice (you're probably not getting a refund from the origin)
Some of the carrier also charge crazy high fees for doing the import service.
Like a dream laptop, the only things I could wish for is an AMD option, Dvorak keyboard option and a matte screen option.
A laptop with an ortholinear+QMK option would probably be an instabuy from me. Hell... native QMK (or ZMK) alone would get my attention.
I like the idea of the Framework laptop, but I haven't paid more than $500 for a PC, ever.
Well this isn't the market for you, and Framework have never claimed to be going for the ultra cheap laptop market. At least not yet... perhaps if they get a lot of enthusiastic support they'll see a market for the $500 upgradeable laptop.
What decent laptop can you get for $500.
I paid £500 for a laptop 10 years ago (so, more with inflation) and it was:
1) Built really shit, fell apart after a year, broken hinges, weird fan noise
2) Got super hot.
3) had about 3hrs on battery if you avoided doing anything on it.
4) was huge and heavy.
Not saying you’re wrong. But maybe your expectations vastly differ from mine.
I paid $500 for each laptop my kids needed at beginning of pandemic in March 2020 (so $1000 in total). Each is a Ryzen 7, 16 GB RAM and 1 TB SSD (one of them is actually NVMe). Both laptops are still performing greatly without hiccups and they saw 2 vacations on the other side of the country during the 2 summers. And one of them (the NVMe one) I used it to do my work during the 2 weeks vacation last summer (virtualization and virtual machines for all my projects, though I had to get an external USB SSD for the virtual machines to fit with me).
All in all if you take care of them it can be used as work horse during vacations. Only thing I can't do on them is last generation gaming, for that I need my usual work horse which is a $2500 desktop gaming rig.
Refurb last year's model off eBay. Paid a quarter of the price on a high-end ultrabook and it looks brand new.
Do you mean 75% of the price? I find it nearly impossible to believe you got a 1/4 discount off a 1 year old refurbished high-end ultra book. You’re telling me a $2k laptop dropped to $500?
You can find a $1500 laptop for $400-ish if you're willing to wait it out. Either used or refurbs from other countries, provided you can live with some minor inconveniences like the wrong keyboard. Nothing a $2 sticker set won't fix.
Interesting, I’ve never looked at importing slightly used international models
Fair, bit like buying a used car: can't be expected to pay full price.
Macbooks seemingly hold their value though, and it's a wonder of why this is.
Ok. Unpopular opinion. This was my laptop, since nobody is sharing theirs and just talking about hypothetical $500 new laptops.
This laptop is the reason I don’t buy bottom of the barrel prices with high specifications.
Do you adjust it for inflation? 1995 $500 is quite a bit more than 2021 $500
What are you using currently?
This person purchased this laptop no more than a few weeks ago. How is this a review? It's basically just an unboxing and "first impressions".
I'd like to know what somebody thinks of this laptop after having used it for a few months, at least.
I don't understand what's so configurable about this laptop. All you can choose is the expansion cards which control what ports you get, which is useless.
Before Ultrabooks this kind of upgradability was the norm, even the CPU was socketed.
These days you can't even change the battery on most laptops.
You should look at the Linus tech tips review. The components inside are all clearly marked and easy to upgrade. Even the screen has an easily removable bezel.
You can buy it without RAM, without storage, without WiFi card, and without OS and sort out all those things yourself. It's still a laptop with all that entails, but everything that was reasonable to make configurable or swappable -- they did it.
Do you actually save money if you buy all those components yourself? With all the shortages and issues it is a real nightmare sourcing inexpensive components these days. Unless you're at a huge scale and have contracts with the component manufacturers, or you've been hoarding a supply of parts from years ago... I bet it's going to be cheaper to tick a box when ordering vs. buying the same stuff from newegg, etc.
Yes, even if only because you can choose something lower-spec'd than they offer, they're certainly not ripping anyone off (like any other manufacturer does!) vs. buying the same elsewhere. I checked Amazon & Scan at my time of pre-order (UK) and decided to go for it just for the.. 'thrill' of it. (Suppose I can/should check again nearer the time.)
Doesn't any laptop already allow this?
What laptop do you know of that allows this?
Anything Dell makes allows you to choose components and OS.
To be fair I haven't purchased a laptop since 2010.
> All you can choose is the expansion cards which control what ports you get, which is useless.
Given the number of people who argue about which ports are the right ones I think it's quite a nice feature.
The usb c ports have recesses for standardised dongles, that's mostly it.