Debian Dropping A Number Of Old Linux Drivers Is Angering Vintage Hardware Users

[schmidtbag]

People with an interest in older hardware could start their own vintage distribution where they could enjoy their museum pieces. The rest of the world moves on.

Considering all of the different flavors of distros out there, I think this is a great idea - a modern maintained distro built to run on vintage hardware (and not just x86). I think support for all platforms that don't [inherently] have PCIe support is a good cut-off point (so, around 2003). It could have a strong focus on:
* Have an intuitive way to micro-optimize tasks (since on old single-core CPUs, every thread counts)
* Patch modern LTS kernels that may have dropped certain architectures, and then purge any changes that have nothing to do with the distro (like advanced instruction sets that old CPUs will never have). Every byte counts on old HDDs, so removing features from a kernel that just simply won't end up on the PC is worthwhile.
* Maintain support of packages like these old GPU drivers that everyone else is dropping
* Drop packages that will likely never see themselves being used (like drivers only found on PCIe devices, anything that depends on OpenGL 3.x, anything that depends on modern instruction sets, etc)
* Have multiple versions of packages readily accessible that might be broken by dependencies or vice-versa (like X.Org with certain GPU drivers).

Such a distro would have a passionate community who would be paying close attention to what needs to be maintained, and there won't be that much to maintain. They could probably do a new release for every LTS kernel release, while also having really ancient kernels available to install (like from 2.4 or earlier). There's hardly a purpose in using bleeding edge kernels and patching them every single time.

[vsteel]

I think it is interesting how people with vintage hardware feel the need to run the latest software. It is working now, just keep going with the release you are on. When the newer release is out which doesn't have the drivers you need, keep rolling with the latest version you can.

[Ardje]

Wait, elo graphics? The touchscreens? First of all: what does debian have to do with that? And second of all, do they know how many elo touch screens are still in use? It's like a defacto clone thing.

[skeevy420]

I fully understand the feelings of these users. My current workhorse is a Phenom II 365 system - albeit with 16 GB - with a NVIDIA 9600GT card and it is still running strong. No, I don't play games, but do development, office work and video editing. It is still fast enough for all those daily work. There is no pressing need for ever faster hardware because it will not help me to complete my tasks faster and thus gives me more relaxation time.
Whispers of dropping support for these class of CPU's are around for some time, but luckily this has not been done yet.

Mine is a 10 year old Dell T5500 and it is good enough to play games paired with an RX 580. I, too, feel no need to upgrade and my only real outrage with these situations is when they try to push AVX minimum standards because that'll force me to upgrade.

I have three ATI cards from 1998 in a drawer, two brand new and have never been used, and I could care less if they're dropped. I have brand new ATI Rage cards from the 90s and I don't care if their support is dropped. Why? The cheapest, bottom of the barrel, AMD card that supports AMDGPU will run circles around them.

The only real use for those cards is if I happened to have a bunch of other 90s era hardware and wanted to make a legacy system for reasons.

[Ardje]

I think Debian devs are not judging by the old age of the hardware but simply by the code activity. I'm pretty sure that as long as the code is officially maintained and active, they are not planning into removing it.

I must say that both ubuntu and debian are doing pretty well maintaining "old" hardware. Ubuntu has patched ceph to be usable on 32 bit systems, and so has debian, while upstream, they make bug after bug due to wrong size assumptions. 32 bit systems are still very common and very cheap. And I am very thankful, because my 64 bit intel can't cope with the cpu load, but my 32 bit arm has no problem with it.

[ms178]

Considering all of the different flavors of distros out there, I think this is a great idea - a modern maintained distro built to run on vintage hardware (and not just x86). I think support for all platforms that don't [inherently] have PCIe support is a good cut-off point (so, around 2003). It could have a strong focus on:
* Have an intuitive way to micro-optimize tasks (since on old single-core CPUs, every thread counts)
* Patch modern LTS kernels that may have dropped certain architectures, and then purge any changes that have nothing to do with the distro (like advanced instruction sets that old CPUs will never have). Every byte counts on old HDDs, so removing features from a kernel that just simply won't end up on the PC is worthwhile.
* Maintain support of packages like these old GPU drivers that everyone else is dropping
* Drop packages that will likely never see themselves being used (like drivers only found on PCIe devices, anything that depends on OpenGL 3.x, anything that depends on modern instruction sets, etc)
* Have multiple versions of packages readily accessible that might be broken by dependencies or vice-versa (like X.Org with certain GPU drivers).

Such a distro would have a passionate community who would be paying close attention to what needs to be maintained, and there won't be that much to maintain. They could probably do a new release for every LTS kernel release, while also having really ancient kernels available to install (like from 2.4 or earlier). There's hardly a purpose in using bleeding edge kernels and patching them every single time.

Thanks for your detailed proposal. That is exactly on the line of my thinking. Such a distribution can be optimized for their needs and other people on newer hardware can move on.

[fransdb]

These are 12 years old hardware, not 20 years old (as mentioned in the article). I am unsure you will have the same argumentation in 8 years.

Granted, that is no the same age, however, I also talked about the whispers for some time now. That might make some of us nervous.
Just this year I retired an old 32-bit system (Slot-1) I have for ca. 30 years. Just because I could not add more memory to it and - as a result - due to swapping was getting to slow to serve as a mail/web-server. With that system I finally reached the limit of what I could squeeze out of it. However, I was lucky that the hardware was supported all those years since I switched from Windows to Linux some 20+ years ago.

As Danny3 stated, as long as it is usable, bug free and no updates are required because of compiler, kernel or other API changes, why would anyone thinker with the existing codebase? In your view it seems that as long as the date of the repository is changing and/or a file is changed just for the sake of making a change, it might still be included. Creating work where no work is required.

As for any good reader, I just talked about my feeling. For some a strange concept.

[L_A_G]

I fully understand the feelings of these users. My current workhorse is a Phenom II 365 system - albeit with 16 GB - with a NVIDIA 9600GT card and it is still running strong. No, I don't play games, but do development, office work and video editing. It is still fast enough for all those daily work. There is no pressing need for ever faster hardware because it will not help me to complete my tasks faster and thus gives me more relaxation time. Whispers of dropping support for these class of CPU's are around for some time, but luckily this has not been done yet.

I highly doubt those whispers of dropping support for an architecture as recent as K10 are anything serious assuming you didn't just make that up so you could have an argument. Hardware of roughly that age is, as you said, still very much usable for current day use cases and you can still go out and buy brand new hardware with lower performance than that. This simply isn't true for the hardware that's now being dropped, which is currently only really usable with other equally obsolete hardware. We're talking hardware so old that most of it has already ended up at eWaste recycling plants by now and the rest is mostly lying in people's attics or basements.

The reality is that the more hardware you have to support the harder it gets to make any significant improvements, particularly architectural ones, without breaking something. Fixing stuff that's still maintained is way less of a hassle than unmaintained code as not only do you have people on hand that are familiar with the code in question, even if you have people familiar with the unmaintained code on hand those people won't have it in recent memory is its been years and years since they touched that code.

[skeevy420]

I highly doubt those whispers of dropping support for an architecture as recent as K10 are anything serious assuming you didn't just make that up so you could have an argument. Hardware of roughly that age is, as you said, still very much usable for current day use cases and you can still go out and buy brand new hardware with lower performance than that. This simply isn't true for the hardware that's now being dropped, which is currently only really usable with other equally obsolete hardware. We're talking hardware so old that most of it has already ended up at eWaste recycling plants by now.

Red Hat's been tossing around the idea of dropping support for pre-AVX x86_64 which includes K10s and my system. Just view the articles from here from the past 30 days and you'll come across it.

[fransdb]

People with an interest in older hardware could start their own vintage distribution where they could enjoy their museum pieces. The rest of the world moves on.

That is, your assuming - so it seems - that these people don't use more modern hardware, at least if I look at your last words about "..the world moves one.". Some people have build a past life, with memories and feelings. Holding to some of them is not wrong and show others that there was a life before that age. Young people will find themself there too, after some years. So be aware of qualifying (explicit or implicit) expressions.