Quoting: Liam DaweThey did say very clearly in an official video, that if something doesn't work they consider it a bug in Proton.

What they said was "something that we said earlier on is that we really want the entire library to work... and if it doesn't work we see that as a bug and we want to fix it."

Of course there aren't games that they don't want to work, and of course they want to fix Proton bugs.

Developer docs even tell developers to log bugs directly to Valve too FYI https://partner.steamgames.com/doc/steamdeck/proton [External Link] under "Reporting Issues".

They want developers to test their games in Proton, and to report Proton bugs to Valve so they can be fixed. They'll take bug reports from the community, like they always have, and fix Proton bugs. They are not taking over support, and if a game hasn't been tested by the developer and breaks in Proton then Valve will (eventually) patch Proton (if they can), or redo the validation tests and take away its Validated status (and potentially issue refunds) and they absolutely don't want to have to that.

Valve aren't bothered if a game is native or runs through Proton, but it's the developer that needs to test and support it, not Valve. Valve's customer support is for the sales - "we think it works enough for you to buy it" - and their developer support is providing tools to make it easier and a communications channel.

View this comment - View article - View full comments

Quoting: Liam DaweEdit: and to be clear, what Valve said before still stands - if there's a problem with a game they believe it's on them to fix with Proton. Outside of making things better specifically for the Steam Deck like input and text sizing, it's largely up to Valve. I'm happy to see developers jump in like this though, I think it's great to see more pay a bit of attention. So really on the subject of what level of support - no matter what it remains the same with Proton - bugs go to Valve and main support goes to Valve.

They've said that they'll fix bugs in Proton by fixing bugs in Proton. They aren't taking over all game support like they did with the whitelist. They expect developers to test their games in Proton and give instructions on how to do so.

View this comment - View article - View full comments

Quoting: denyasisWell, yes. I believe a Windows build is still required to release on Steam. No Linux exclusive releases.

I understand that is no longer policy, although it did used to be. Most games are going to be made available for Windows because marketshare. The number of all games currently on Steam is 64,289, and the number of Windows games is 64,270.

Perhaps with the Deck, that has changed, hopefully.

They said during the dev conference that they don't want Deck exclusives. The big objective is to demonstrate the viability of Linux as a gaming platform to all those Windows gamers out there with their existing Steam libraries, rather than to sell lots of units.

View this comment - View article - View full comments

Quoting: pete910They have some serious work to do :huh:

Yep. With a backlog of 64,000 games and about 28 new games released per day, if they could test 63 games each day then they could catch up within five years. But the rate of new games is increasing, so they'd need to do more than that. And some games are going to need multiple tests, so they'd need to do more still.

Realistically, there will be games that never get tested. But their own decision on which games to test first is influenced by the games that people who are interested in the Deck want to play, and game developers can request a test if the Deck market is of particular interest to them.

View this comment - View article - View full comments

Quoting: pete910

Quoting: CatKiller

Quoting: pete910Only 10.5k to go to get all of steams library done :whistle:

More than that: about 64,000. [External Link]

Thats including DLC :wink:

My number was from 2021 too

That's not including DLC. With DLC it's about 98,000. [External Link]

Your 11,000 would be the number that were released just in 2021.

View this comment - View article - View full comments

Quoting: pete910Only 10.5k to go to get all of steams library done :whistle:

More than that: about 64,000. [External Link]

View this comment - View article - View full comments

Quoting: AciDI know that Steam is a DRM...

Not really. You do need to download the games through the client, but after that it's entirely down to what the publisher chooses. If a game doesn't make use of Steam features (which need the Steam client) and hasn't included any other DRM, you can just run the games however you like. There are lots of games on Steam like that.

View this comment - View article - View full comments

Quoting: elmapulwhat about server side anti cheat? its possible to do the validation server only?

Ish. For some games server side anti-cheat is perfect, with no downsides. But if you can push more of the work to the client then the player gets better performance, better latency, and the user pays the electricity costs rather than you. So you can try to really nail down the absolute minimum information that you pass to the client, and check every interaction against a definitive server game state... or you just bung everything at the client and try to stop the client being tampered with.

Server-side anti-cheat can't stop someone from just pointing a camera at the screen, but client-side anti-cheat can't, either.

View this comment - View article - View full comments

Quoting: EhvisAren't you kind of arguing against yourself now? Why would a company that believes in the value of client side anti cheat want their system hidden from the kernel? That goes against their philosophy.

No. You've got your Windows application doing things that you want to keep secret. Cheats move to kernel space, so they can keep looking at (and changing) your application. So you move part of your application to kernel space. So the cheats run your application (and Windows) in a VM, so the hypervisor can keep looking at (and changing) your application. So you move your application to its own VM with your own hypervisor. And so on.

On Linux, you don't really need to do that, because there are people in the Linux space (such as, say, Amazon) that are quite keen on keeping one user's application isolated from another user's application, and from their infrastructure.

Tim's claim is that you can't trust Linux users because they might have compiled their own cheating kernel (like Windows users are able to create their own cheating kernel space drivers) so you want to hide your secrets from a potentially untrustworthy kernel.

View this comment - View article - View full comments

Quoting: EhvisIn the end, the only solution that would satisfy the creators of client side anti-cheat would be to have some sort of a signed Linux system that they can get root access to.

Not really. I mean, that's what they want to do, because that's what they do on Windows, and having your application running all the time gets you valuable marketing and metrics. But userspace applications under Linux can have secure enclaves that are hidden from other applications, from the kernel, from hypervisors, whatever. They don't need kernel-level client-side anti-cheat; they just want it.

View this comment - View article - View full comments