Valve adds documentation for Steam Deck development, suggests Manjaro Linux for now

That NUC is frickin' adorable tho.

There are some others that come with Linux pre-installed that GOL has reported on before, if you're interested in that kind of thing.

Alas, I have 9 cats to support, I cannot spend 600 bucks on electronics. :P

This just reminds me of this;
https://www.youtube.com/watch?v=_dhjmSp30jc

Still not complete though. They don't say which kernel to test against and which branch of Proton Experimental.

We're four months from the release. I'm pretty sure, they might update the kernel and Proton will definitely progress a lot.

I have 9 cats to support

I'm using that line the next time I ask for a raise!

(Also, wtf? I had one (1) cat, with her own room and own queen size bed, and she was impossible to control. Nine!?)

Isn't "NUC" an Intel-specific marketing name (Next Unit of Computing), nothing to apply to AMD-based hardware? So it's now a proprietary eponym like Band-Aid, Kleenex, Hoover etc.! 🤔

Intel must be proud, especially since they didn't even pioneer the concept 😛

Yes, it is. It's more catchy than the "proper" term barebone PC though, which is probably why it's being used to describe non-NUC barebone PCs as well.

And the technical purpose of the banana is....????

And the technical purpose of the banana is....????

https://knowyourmeme.com/memes/banana-for-scale

so let me get that straight...
they cant produce enough steam deck devices to cover the needs of developers who want to develop for then...
but they expect to produce enough devices to the people who want to buy then?
there are 50.000 games on steam, and considering some devs made more than 1 game, less than 50.000 developers/publishers, i can understand that big companies that need tons of developers to make/port their games are an priority, but how many devices is valve producing?

so let me get that straight...
they cant produce enough steam deck devices to cover the needs of developers who want to develop for then...
but they expect to produce enough devices to the people who want to buy then?
there are 50.000 games on steam, and considering some devs made more than 1 game, less than 50.000 developers/publishers, i can understand that big companies that need tons of developers to make/port their games are an priority, but how many devices is valve producing?

They don't charge developers for dev kits; they do charge customers for retail units, and they've essentially sold two years' supply already. If they cannibalise their retail units to send to devs, they're costing themselves $400 a pop and making their customers (who also give them money for games) unhappy. They had intended to have a second batch of dev kits available by this point, but there's a component shortage on at the moment, so they're giving some information so that devs can make do.

Still not complete though. They don't say which kernel to test against and which branch of Proton Experimental.

As for recommending Manjaro, that's a big oooof. Much wiser to either install EndeavourOS *SNIP*

Full stop right there. lets not continue the old adage of "you should be using this" that the Linux community has. the main reason Valve picked this, is for 1, ease-of-use and 2, its more then likely what the DE is closest to Valve OS. at the same time, just because the coders hold the title of game Devs, doesn't mean they are well versed in the ins and outs of Linux environments.

so let me get that straight...
they cant produce enough steam deck devices to cover the needs of developers who want to develop for then...
but they expect to produce enough devices to the people who want to buy then?
there are 50.000 games on steam, and considering some devs made more than 1 game, less than 50.000 developers/publishers, i can understand that big companies that need tons of developers to make/port their games are an priority, but how many devices is valve producing?

They don't charge developers for dev kits; they do charge customers for retail units, and they've essentially sold two years' supply already. If they cannibalise their retail units to send to devs, they're costing themselves $400 a pop and making their customers (who also give them money for games) unhappy. They had intended to have a second batch of dev kits available by this point, but there's a component shortage on at the moment, so they're giving some information so that devs can make do.

400 include cost of production (and i guess shipping?) i dont know if its subsidized as consoles are to make a profit on games sold.
i dont know, i just think valve investment is too small.

• Nintendo paid 1 billion to unity to support their platforms a few years ago


• google paid capcom 10 millions to port Resident Evil 7 and 8


• valve on the other hand seem to expect developers to support their platform for free, while they have an umproven track record on hardware compared to others


• we were praising valve for hiring 1~2 developers to write/improve an driver while modern games take up to 1500 developers to be produced and sometimes windows get drivers specific for an game


• valve said they aim to make every game run on steam deck, but there are still 3000 open issues on proton github, some of then are 2 years old (the same age as proton, so they might be even older issues that existed on wine)


• valve missed the critical window to relase it that is the end of the year, they will have to wait another year to sell as much as they could

its just seem that valve dont know how much money they need to invest to make this product reach it full potential, and they are orders of magnitude below what they need.
i hope this can sell at least as much as the first xbox, so we might have an second chance next gen, but at this rate microsoft will enter the portable pc market before steamOS can make any dent, and they will compensate any performance loss of windows with an better hardware.


Last edited by elmapul on 13 November 2021 at 9:13 am UTC

Still not complete though. They don't say which kernel to test against and which branch of Proton Experimental.

They're on 5.13 for the kernel, not sure about Proton Experimental.
Also, installing the Zen kernel will definitely be beneficial, as it's pretty close to the SteamOS specific kernel Valve will eventually ship.

As for recommending Manjaro, that's a big oooof. Much wiser to either install EndeavourOS if you're new to GNU+Linux or use Arch if you're an experienced user/dev.

5.16 would be the logical target.

And as has been stated, Manjaro is fine. It's close to their setup, easy to install, and has Steam installed already and that is pretty much all they need. They don't need to use the OS beyond running Steam, so it really does not matter how user friendly any of it is.

valve said they aim to make every game run on steam deck, but there are still 3000 open issues on proton github, some of then are 2 years old (the same age as proton, so they might be even older issues that existed on wine)

That's not a very good metric. Bug reports are per game, and people are encouraged to report games that work without issues too, so that list will grow and grow.

Most of the closed bugs are duplicate reports.

And the technical purpose of the banana is....????

Reminding you to get your daily potassium intake.

valve said they aim to make every game run on steam deck, but there are still 3000 open issues on proton github, some of then are 2 years old (the same age as proton, so they might be even older issues that existed on wine)

That's not a very good metric. Bug reports are per game, and people are encouraged to report games that work without issues too, so that list will grow and grow.

Most of the closed bugs are duplicate reports.

that is a good metric because i see the numbers a lot of times to see if they changed.
the number of bugs being closed isnt changing too fast, unlike the number of new bugs being reported.
sure, people can open a bug report to report that something is working fine, but even if we filter out those (white list request) there are still too many bugs.

and if there are many duplicate reports, then the tag duplicate would be full of entries, unless no one is looking at this.

Where did Valve say they've sold two years supply of Steam Deck already?
Because if that's correct, and those devices are not yet manufactured but already "sold", that's impressive!

The first six months' supply went in less than an hour. They briefly had more granularity for the later units (taking it out to a year's supply), but fairly quickly put everything past that into the "after Q2 2022" bucket. Some people might not want to put the full amount down when their turn in the queue comes up, and I'm sure Valve will try to increase their production rate, but as it stands now every unit they can make for the foreseeable future has already had dibs called on it.


Last edited by CatKiller on 13 November 2021 at 8:23 pm UTC

that is a good metric because i see the numbers a lot of times to see if they changed.
the number of bugs being closed isnt changing too fast, unlike the number of new bugs being reported.
sure, people can open a bug report to report that something is working fine, but even if we filter out those (white list request) there are still too many bugs.

and if there are many duplicate reports, then the tag duplicate would be full of entries, unless no one is looking at this.

No, it's not even a good metric for this, since Valve uses one bug report for each game. One game can have zero issues, or a hundred.

Still not complete though. They don't say which kernel to test against and which branch of Proton Experimental.

As for recommending Manjaro, that's a big oooof. Much wiser to either install EndeavourOS *SNIP*

Full stop right there. lets not continue the old adage of "you should be using this" that the Linux community has. the main reason Valve picked this, is for 1, ease-of-use and 2, its more then likely what the DE is closest to Valve OS. at the same time, just because the coders hold the title of game Devs, doesn't mean they are well versed in the ins and outs of Linux environments.

The reasons I suggested EndeavourOS is because of how Manjaro have handled security issues in the past and EndeavourOS is closer to pure Arch.

I'm not suggesting any one distro is better than another, simply basing my recommendation on personal experience.

I think it's good Valve have the recommendation in place, even though I don't agree with it.

Well, hopefully the suggestion of using Manjaro is just a temporary one until SteamOS gets a new release. I have tried Manjaro a few times and think just pure Arch, while a pain to do an initial install, has a much more likely long lasting set up before something breaks.

Still not complete though. They don't say which kernel to test against and which branch of Proton Experimental.

As for recommending Manjaro, that's a big oooof. Much wiser to either install EndeavourOS *SNIP*

Full stop right there. lets not continue the old adage of "you should be using this" that the Linux community has. the main reason Valve picked this, is for 1, ease-of-use and 2, its more then likely what the DE is closest to Valve OS. at the same time, just because the coders hold the title of game Devs, doesn't mean they are well versed in the ins and outs of Linux environments.

The reasons I suggested EndeavourOS is because of how Manjaro have handled security issues in the past and EndeavourOS is closer to pure Arch.

I'm not suggesting any one distro is better than another, simply basing my recommendation on personal experience.

I think it's good Valve have the recommendation in place, even though I don't agree with it.

Well, hopefully the suggestion of using Manjaro is just a temporary one until SteamOS gets a new release. I have tried Manjaro a few times and think just pure Arch, while a pain to do an initial install, has a much more likely long lasting set up before something breaks.

Ha, quoting myself...

Maybe it is the effort, but I usually find distributions that are a just a bit more of a hassle to set up are longer lasting for me (as in I am less likely to distro hop if there was some effort in getting it set up.) Every distribution has some sort of imperfection. Which is why there are so many of them.
Had Ubuntu stuck with their original goal of being Debian with a 6 month release and always the most current Gnome (which used to be a huge problem for Debian, even if you were running Sid), than it really would have been my favorite distribution. But NiH attitudes basically made me go back to Debian. It makes a lot more sense to base things on it rather than Arch, as rolling releases always have their own growing pains, by their very nature. But for sure, the packaging of things is WAY easier on Arch.

that is a good metric because i see the numbers a lot of times to see if they changed.
the number of bugs being closed isnt changing too fast, unlike the number of new bugs being reported.
sure, people can open a bug report to report that something is working fine, but even if we filter out those (white list request) there are still too many bugs.

and if there are many duplicate reports, then the tag duplicate would be full of entries, unless no one is looking at this.

No, it's not even a good metric for this, since Valve uses one bug report for each game. One game can have zero issues, or a hundred.

fair enough, i'm making an script to meassure this.

version 0.9 of the script is done:
https://github.com/Elmapul/proton-good-metric
i recomend reading the commit notes to better understand what it is:
https://github.com/Elmapul/proton-good-metric/commits/main

sorry for the sarcastic name =p


Last edited by elmapul on 15 November 2021 at 8:48 am UTC