Valve looking to drop support for Ubuntu 19.10 and up due to Canonical's 32bit decision (updated)

Brainstorming session at Canonical Offices: The Cloud, IoT, Kubernets, AI. Other ideas? Yes: The Cloud, IoT, Kubernets, AI. The Cloud, IoT, Kubernets, AI. The Cloud, IoT, Kubernets, AI. The Cloud, IoT, Kubernets, AI...........

Quoting: TheSHEEEPI simply look beyond a few personal inconveniences at the bigger picture.
And the bigger picture is that progress requires sacrifice. Can't play some games/use some apps anymore? So be it, if that's the price to pay to finally get rid of old stuff like 32bit for good.

Because . . . the libraries take up a few megs on your terabyte+ hard drive? Yeah, surely it's worth getting rid of some functionality to get that!!!
But you know, all you have to do is delete the relevant libs. Nobody's stopping you.

Quoting: Schattenspiegel

Quoting: TheSyldatPop!_OS is the best solution really . Hell they are already better than Ubuntu on some various things .

Not sure -they probably are, but they are also a bit small, based on Ubuntu and unfortunately ship only the gnome desktop - so maybe something more...diverse ;-) concerning DEs and a little more upstream. I would love to see Mint but they have a similar situation. openSUSE may actually not be a bad idea.

at least they have an source of income

Quoting: BeamboomAnd like I said in that other discussion: One can't expect an old binary to run on new computers for all eternity.

Why not?

Quoting: EzyRhinoI'm beginning to think that the decision Canonical made is exactly what they want. This is all speculation on my part, but I think they want to focus 100% on the enterprise (AWS, Azure, etc...) and get out of the end-user desktop market. Sad if true.

Could be. I've noticed a tendency with distros produced by companies: They enter the scene by creating a lot of buzz as the saviors of the Linux desktop. They leverage this into a more or less prosperous business selling their distro for servers and other back-end work. They gradually ease away from the desktop, which they have found impossible to make money from; it seems to be merely overhead. Once they are no longer seen as major desktop players, their share of back-end stuff gradually erodes because there is no buzz about them. Then some new company comes along, creating a lot of buzz as the saviors of the Linux desktop . . .

The next one could be either Valve or someone partnering with Valve, the big difference being that Valve seem pretty convinced a robust Linux end-user desktop is one way or another going to make them money.

Quoting: Purple Library Guy

Quoting: TheSHEEEPI simply look beyond a few personal inconveniences at the bigger picture.
And the bigger picture is that progress requires sacrifice. Can't play some games/use some apps anymore? So be it, if that's the price to pay to finally get rid of old stuff like 32bit for good.

Because . . . the libraries take up a few megs on your terabyte+ hard drive? Yeah, surely it's worth getting rid of some functionality to get that!!!
But you know, all you have to do is delete the relevant libs. Nobody's stopping you.

That wouldn't serve any purpose, as it would only be for me. As I said, I'm looking at the bigger picture here.
And the bigger picture is that right now, 32bit is like an annoying mosquito at the butt of many developers and users... just look at this very situation.
If 32bit were gone for good, well, so would be situations like these.

But obviously, this was not the way to go about it, Canonical should have gathered some support and actual solutions from bigger linux players beforehand.

Quoting: Guest

Quoting: TheSyldat

Quoting: GuestIs it really because of the decision to drop 32bit OS support? Nobody, literally nobody, especially not Valve's target audience, is still using a 32 bit OS unless they just don't know any better. I think it's more because they already have steamOS and ubuntu is falling out of fashion anyways.

Again they are not just dropping the 32 bit flavor , they are freezing the 32 bit multilib in terms of update (essentially deprecating it ) sooo as a result a lot of your games on steam won't work out of the box on newer versions of Ubuntu but more importantly even after trying to get the 32 bit libs it might still not work all that well because they're not updating them anymore either .

Oh damn, I mean I get why canonical would want this, 32-bit should basically only exist for legacy applications, modern applications should not be relying on it, and theoretically on a modern OS you should never need to use 32-bit libs, but there's always some legacy application creeping around somewhere that relies on it, and there's always some stupid developer that didn't bother compiling for 64 bit here and there.

I mean this kinda needs to happen, but maybe it's a bit ahead of it's time, this decision.

Not only legacy application, but also applications with legacy code which is not 64bit compatible - there are old project with huge codebase. Or for some instance for application for which is 32bit better. Here is post from enemy camp developer about their visual studio (2019 is still 32bit aplication, i asume point are still valid, or their codebase is realy incompatible)

Quoting: Purple Library Guy

Quoting: BeamboomAnd like I said in that other discussion: One can't expect an old binary to run on new computers for all eternity.

Why not?

Do I really need to explain that? Why do do think they want to do this to begin with? Why do we phase out technology quite regularly - despite the hard struggle every bloody time we do it?

Why did Windows become such a bloody, security issue riddled mess? Several reasons, but the need to be backward compatible is one massive reason. Messy as f*ck. A patchwork out of this world.

In a world of limited resources, spending a lot of those resources on backward compatibility can be argued is a energy wasted that could be spent a lot better. Especially now that those old blobs of binaries can be run in virtualized environments.


Last edited by Beamboom on 24 June 2019 at 10:56 am UTC

Quoting: BeamboomIn a world of limited resources, spending a lot of those resources on backward compatibility can be argued is a energy wasted that could be spent a lot better.

I guess it depends on the value of the old stuff (high for many of us) and the prices for the different solutions to keep the value. The price for Ubuntu to keeping these libs seems not high for me, especially considering Debian still maintains them. The price of hundreds of thousands of people setting up VM (or the like) solutions seems a bit higher. Yes, we do need some other solution some day. I just don't think it's this day yet.

What the heck does "frozen" mean? In order to have it work correctly, if you're building/providing multiarch, you're going to have to release 32-bit code in lock-step with the 64-bit.

The problem with "support" for games right now is that there was a lot (and I do mean that...I have my own dev number, etc...) that were provided in the early stages with Steam support only being 32-bit, so you couldn't even PROVIDE a 64-bit version. The question was posed back in the beginnings to the dev forums and Sam Lantinga replied that they'd not had the resources to roll that out at that time.