Quoting: WYWIt's a very tricky thing to balance. They don't want to release a SD2 too soon since it will make current buyers feel burned, and they don't want to wait too long since interest will fall as it ages.

Having a stable target platform is important, and Valve have said as much.

A CPU generation is one year. A GPU generation is two years. A typical PC upgrade cycle is 3-4 years. A console generation is 5-7 years. Slotting a Deck generation in at 4-5 years makes sense to me, to balance the fixed target against staying relevant as a part of the wider PC ecosystem, but I don't work for Valve or AMD. For ease of reference, the Deck was announced in mid-2021 and released early-2022, with the OLED released late-2023. I'm expecting 2026 to be the time that the Deck 2 is released (or perhaps only announced) as 5 years from the announcement and 4 years from the release of the Deck 1. (Edit: although 2027 works, too, as 4 years from the release of the OLED and 5 years from the release of the LCD)

View this comment - View article - View full comments

Quoting: pleasereadthemanualIf they could release it in Australia first...

Oops, yes, that's something that I forgot to put on my list: worldwide availability.

View this comment - View article - View full comments

QuoteAnd what do you now want from a Steam Deck 2?

A lot of the failings of the first-gen Deck got fixed with the OLED - the bezels got reduced, the battery life got improved, L1/R1 got less janky, the not-great display got swapped, the too-small 64 GB SKU got dropped, and the power LED has status colours.

There are still remaining improvements that can be made, though, so I'd like

• enough performance to run 2026-current and near-future AAA games at native resolution at 30 fps or higher
  
• OLED for every SKU
  
• stacked cache on the APU for more effective bandwidth
  
• any magical new-fangled battery tech that lets you fit more battery inside.
  
• additional intake vents along the bottom edge so that neither "propped in lap" nor "balanced on a sofa arm" will completely starve the cooling of air
  
• a Kensington slot. If the Deck becomes the de facto machine to leave at trade shows or hand out to focus groups, then "the machine that my in-development game absolutely has to run well on" is the Deck
  
• further improved L1/R1 buttons, and potentially make those analogue
  
• have some means, either tactile or backlight, to find the supplemental buttons.
  
• since both "game developers don't test below 1080p" and "this game is only Playable rather than Verified because the text is too small" are both things, a size bump to 8½-9 inches and a resolution bump to 1920×1200 if they can get that to meet the performance requirements
  
• a kickstand, or a prop solution that comes in the box, since external controller support is already good.
  
• sticky strips instead of glue, and maintaining or improving the deconstruction process
  

Whether it will need a RAM bump and a storage bump is hard to say right now, but it seems like it might. Streaming-to-Deck and shopping-on-Deck need work already, but they'll want to fix up both of those before a Deck 2 is released.

View this comment - View article - View full comments

Quoting: pleasereadthemanualI kind of wonder if there's much point in adopting a Flatpak package as a proprietary vendor.

It's certainly possible with any of {flatpak, snap, appimage} to have that be your official distribution method for your software. Any of them will work on any distro. Whether there's any point for a given project is a different question. The Firefox snap and the Firefox flatpak, for example, are both owned by Mozilla, and they specifically requested the Firefox snap so that updates weren't held up by the distro's update schedule and maintenance burden. But most projects are going to be quite comfortable with fan-made binary-only packages - whether in distro repositories or container repositories - without having to do any maintenance themselves.

View this comment - View article - View full comments

Quoting: pleasereadthemanualI'm coming at this from the perspective of someone who is trying to build a Flatpak package and convince upstream to take control of it once I have it generally working.

I can't help you there. But even if Valve, say, wanted to take over the maintenance of a flatpak that doesn't mean that any other project would want to with theirs.

Quoting: pleasereadthemanualI don't know what Valve's official statement is on the Flatpak, but I have heard they don't want to take control of it for some reason. That might be due to Steam trying to spawn another sandbox inside the Flatpak sandbox, which isn't possible. I'd be curious what exactly the situation is there.

Their official statement is that it's unsupported. Same as snap, and same as any distro but Ubuntu LTS and SteamOS.

Quoting: ValveSteam only officially supports Ubuntu running Ubuntu 12.04 LTS or newer and SteamOS, but the Steam for Linux community is extremely resourceful and has managed to run Steam on a large variety of distros. Valve approves of these efforts but does not officially endorse or provide support for them.

Quoting: ValveSteam has been packaged as a Flatpak app by the Flathub community, but this Flatpak app is not officially supported by Valve...

Steam has been packaged as a Snap app by Canonical, but this Snap app is not officially supported by Valve.

As far as the container goes, the Steam container is the same as flatpaks - bubblewrap - and having a container in a container specifically didn't work for bubblewrap. They had to make changes on both ends so that it would work, as I already mentioned. There are probably Phoronix or similar articles going into what exactly they did to make it work.

Edit: found the change. It was

QuoteAllow a subsandbox to have a different /usr and/or /app.
Steam will use this to launch games with its own container runtime
as /usr (the "Steam Linux Runtime" mechanism).

from flatpak 1.11.1 (in 2021) with concurrent changes to pressure-vessel. Prior to that games using the SLR or using later versions of Proton wouldn't work in the flatpak.

View this comment - View article - View full comments

Quoting: pleasereadthemanual

QuoteFixed startup delays when running in flatpak environments.

Interesting that Valve cares that much about the Flatpak package, but don't want to officially adopt it.

Flatpak is the fourth most popular distro in the Steam Hardware Survey. Having had to do a bunch of stuff already on both the Steam side and the flatpak side to have container-in-a-container work at all, if there's more they can do to make it work better - why not?

View this comment - View article - View full comments

QuoteOne of the interesting tweaks is that Valve adjusted the CSS to "reduce feature leaks", I guess they're growing a little tired of certain people scraping it constantly to see what they're working on.

I mean, it is weird that SteamDB knows which keyboard I've picked for my Deck.

QuoteFixed an issue where after playing a game it would require extra B button presses to back out of the game's library page.

Yay!

View this comment - View article - View full comments

QuoteI think the price is the biggest issue here. I really don't like the idea of developers undervaluing their games as they can take a lot of time and effort to make, and pricing even for indies is tough but at £24.99 for a game that really is relying on nostalgia, in years when we've had constant amazing new releases that are either cheaper or a similar price. It's a pretty tough sell.

Yeah, that is way too high. The quality and breadth of their current competitors at that price and below, and of course the long tail of the Steam catalogue, means that game doesn't stand a chance at £25. £7-8 is likely a feasible window if it's mainly fuelled by nostalgia but is otherwise competently executed.

And, for me, since it's got no Linux support, hasn't managed to get Verified, and doesn't entirely work on Linux, it's only worth at most 10% of that.

View this comment - View article - View full comments

Quoting: BrokattRemember that most PC gamers are not that tech savvy.

This still seems weird to me. I can accept that it's true, but it's weird.

I was around for the start of PC gaming, and learning what you could configure, what it did, experimenting, personalising, tweaking, overclocking, and all of that... was PC gaming. When I switched to Linux 20 years ago, nothing particular changed; I carried on gaming, I carried on learning, and I carried on personalising - I just had a broader palette to work with.

That there are generations of PC gamers since that do gaming like cubicle work, with none of that knowledge or variety, always strikes me as a bit odd... and a little bit sad.

View this comment - View article - View full comments

Quoting: NathanaelKStottlemyerWait. Why? Is valve...? Why are they doing this? Is something with nvidia in the works or is it steamos?

Why not? They want Linux to be a viable platform, and most Steam users have an Nvidia GPU by a huge margin. Open source allows them to improve things directly. They poached Pierre-Loup Griffais from the Nvidia Linux driver team, so it's not like they don't have the expertise.

View this comment - View article - View full comments