Steam On Linux Usage Climbs Higher Thanks To The Steam Deck

[Emmanuel Deloget]

You are contradicting yourself:
- Is my idea easy to understand because it is not very novel, or is it so novel that it is not easy to understand? Your sentence makes no sense.

Wut ?

The meaning of my question is quite clear. But the reasonning behind your analysis is... Well, kind of expected I guess.

Why don't you identify the three issues that I have been talking about. Why are you trying to discredit me, but you avoid answering my main question?

What in the world makes you believe that I want to spend hours on trying to identify these three issues ? For God's sake, your argument that if we don't find by ourselves these Three Holy Issues we won't be able to understand your argumentation is not only plain dumb, it's also insulting. Do you really believe that Yann LeCunn played that game before he unveiled his seminal paper on stochastic training for deep learning? No. He published it and other in the fields were able to understand it. The same goes for all major science papers out there. They are published first - discussion arise after. There is a very good reason for that and it seems this reason goes very far above your head, so let me state it for you : it is easier to discuss about something that both side of the discussion have shared.

For the record, that's how you shall proceed. If you have a genuinily good idea (and I sincerely doubt that, as people who have good ideas in this area tend to not be that cryptic) then state it, explain it. You don't have to play stupid clue games: we're not your lab rats.

And please note that I don't have to discredit you -- you are far better at this game that I will ever be.

I'm trying to make my thought easier to understand by highlighting the key words. There is nothing wrong with that.

Of course there is. It makes words blow out of the page and as a consequence, it makes the whole text harder to read. Instead of reading an organized text I have to fight against tiresome explosions of nonsensical words. You are using bold in the wrong way and you're not even realizing it.

And to make a thought easier to understand, it would actually be usefull to state that thought. Are you able to do it, or do you prefer to continue to play that really stupid and infuriating game?

My guess is that you cannot. My guess is that you have two or three ideas but you don't know their values (hint: they have none). So instead of stating them in the blue you started this little game of yours, trying to see if one of the mouse will find an imaginative way out of the labyrinth of your tortuous head. Then, if that happens (and if you recognize this moment, and I'm not sure you are equipped for that), you'll be able to say "yeah, exactly that, you got it -- but remember I knew that before you, so I'm the one who shall get the credit". And believe it or not, I find this behavior pathetic.

Now, you can either confort me (and, reading other comments, a lot of other people) in this vision by continuing to insult our intelligence -- or you can disprove me by saying what you have to say. It should not be difficult, as you already know what you want to say, right?

[sinepgib]

To drop the idiot persona for one second, it really is as simple as that. As someone who *actually* knows what they're talking about it really is just this. Why would someone (or a huge AAA studio) invest thousands of dollars at minimum when it will get them a fraction of the money they spent back? Note that for AAA studios it would likely veer into the hundreds of thousands or even millions depending on how many people they dedicated, not to mention the QA resources required to support such a build. There is actually no way to justify it if they're publicly traded because they have an obligation to the shareholder to not waste money and typically Linux builds actually require __more__ work than Windows builds because you have to get it working with the 20,000 possible combinations of software and you never know when some random library is going to have a bug that fucks up your basically-hieroglyphics physics code or, heaven forbid, your networking code. Even if it was 1:1 in terms of effort it wouldn't be worth it when >95% of your userbase is using Windows and the rest are using Proton anyway which is a much better target.

There is a business case tho. Maybe not that strong, but if you can remove the privileged stores (Windows and Mac are both players in that area and happen to be able to position themselves above by owning the OS) then you can have your own with a degree of success and not have to pay as much in royalties. AFAICT that's why Valve is pushing it after all, so whatever the Xbox and MS Stores are called they do not have an asymmetric advantage over Steam. But as correctly stated, for that to even be viable you need to be able to pack once and ship everyone like you do for Mac and Windows, which is currently near impossible (and when possible, it comes with a big disk and RAM usage hit) in the Linux ecosystem. If hardware support wasn't in a sorry state even the BSDs would have a better chance at succeeding for that use case than Linux.

[sinepgib]

Same way you get a stable ABI with a library? Look at the entire Windows API userland.

Of course it doesn't exist on Linux. Thanks for stating the obvious. That's their decision. They want to freely change the interfaces within the kernel without maintaining old interfaces. It's a choice.

Just decide on a fucking ABI and interface and never change it, only add to it. That's how stable ABIs are born. It's the same way with microkernels, you must decide on an interface, and never break it.

Linus does that, with userspace, but not with kernel. It's his stupid choice, not a technical limitation.

I already gave 3 analogies with userland libraries and IPC. Not sure what kind of evidence do you want.

Windows has stable ABI with both microkernel and in-process interfaces (i.e. "monolithic" stuff).

To be fair, you don't really need evidence if you knew what you were talking about. This is basic programming concepts. Feels like I'm talking to oiaohm. Are you sure you're a developer? Do you know basic assembly at least? (required for binary stability)

But there's an additional problem with userspace-to-userspace interfaces as well. E.g. GTK and Qt never bother trying to keep backwards compatibility, glibc does but it's a pain to achieve (you need to pass special linker flags IIRC to tell it to use an older version of the symbols). Linus doesn't have power over userspace, and has indeed recommended they get their crap together and strive for stability. I think a microkernel architecture may help due to the restrictions it imposes, but not necessarily solve it on its own, as it is a very human problem.

[blackiwid]

Shame on you, blackiwid, and on others who misbehaved.

You insult people with various levels of knowledge here that they must be so dumb even their career is often about that field, that they can't understand a simple pro / con list of a idea. Like toddlers, you play a game with others nobody did consent close to a form of abuse, probably some sort of trolling to enjoy to get people mad with the only goal to get attention and or anger.
You "misbehaved" so much that you lied to us all when you said you unveil your genius reasoning after 1 day this date expired.

if the 10% case that you have a real point and not the 90% of you just trolling would be true, stop it and give us back the free will to decide if we keep interested in your idea or not. You try to take away that free will by trying to condition us or manipulate us to find the information in some way that we have no choice to accept it, at least that is what you imply.

If you insult people don't wonder that people insult you back, if this insults are explicit or implicit does not really matter.

[AlanTuring69]

Blackiwid just got onto my "Who is Who on Phoronix forums" list.

I also made a special post about him so you can easily see what he is doing.

And you have just gotten added to a list of my own that prevents seeing your posts or threads or even seeing the posts of people quoting you. Note that while almost everything about the script is very facetious it does work.

[qarium]

Sorry, dude, if you want to live in your vacuum of "Linux games compatibility is perfect and there are no issues ever" - good for you, only it makes you a cult follower/believer and I don't deal with such people. Please never reply to my further messages. I highly recommend going to https://www.phoronix.com/forums/settings/account and adding me to your personal ignore list. I just don't want to see notifications from people who consciously avoid getting smarter/more knowledgable and for whom their beliefs are above everything. That makes you worse than a bacteria though as even they try to evolve (meaning becoming "smarter") to survive better.

you have a lot of people on your personal ignore list who just want to become smarter and read here only because they want to get more knowledge.

and yes you thread them like shit.

"It's a Proton publisher then"

nonsense a Proton publisher is a linux publisher.

[qarium]

Is there any single example of an implemented, readily available, full microkernel out there that would validate your theory? Or is it Your Insanely Clever Idea that everybody should follow because it's Insanely Clever I Promise You?
You're probably very intelligent, but can you stop please ? Either you tell what you, and only you in the world have figured out and let us judge your ideas on their merit or you stop using this condescending tone that really does not make you shine.
And I can tell you that, in the end, you're still wrong (and since you're so clever, you'll have to figure out why).
Oh, and by the way, in a world where we want our voice chat system to use less than 3% of our CPU so that we can use it while doing Other Important Stuff, following a world where we wanted our MP3 player to use less than 3% of our CPU so that our desktop was still responsive, to a world where we are trying to undo the micro-impact of handling IRQs in order to limite system latencies because we want to have a smooth desktop experience, I can alreasy tell you that lossing 5% of your CPU time in context switches in just absolutely insane, and no, you won't be able to completely optimize that out. That it would make the whole plateform insufferable. That's the consumer desktop use case.
And while I'm at it, the forces out there that do not favor Linux is not the fact that it's not a microkernel. It's the very fact that all freaking softwares that people really use to do important things are running on Windows. Sure, there are open source alternatives but unless I'm completely out of this world, these alternatives are not yet as good as the original (Gimp vs. Photoshop ? Blender vs. 3DSMax? OpenOffice vs. MS Office? LMMS vs. FL Studio? Inkscape vs. Illustrator?...). I have 3 Linux machines at home and everybody uses them. And I still have to use a Windows machine because sometimes, only MS Word can correctly handle the documents that I receive. And some other times, my son want to use FL Studio, which does not have a Linux version. See the problem? The problem is compatibility. And we're then joining a sweet spot of a thread that should not have derailed: that the Steam on Linux is working great, it's (slowly) growing, and this is really what's important here. It means that Valve has an incentive to propose a compatibility layer which is the best possible and to continue to enhance it. And since it's opensource, all Linux users can use their code. Suddenly, wine went from an interresting and fun project to something really useful and tons of Windows applications can now run over Linux. That's waaaaaay more important than the kernel design.
The success of the platform is defined by the number of people that can use it, not by the complexity of its internal design.

"Sure, there are open source alternatives but unless I'm completely out of this world, these alternatives are not yet as good as the original (Gimp vs. Photoshop ? Blender vs. 3DSMax? OpenOffice vs. MS Office? LMMS vs. FL Studio? Inkscape vs. Illustrator?...). I have 3 Linux machines at home and everybody uses them. And I still have to use a Windows machine because sometimes, only MS Word can correctly handle the documents that I receive. And some other times, my son want to use FL Studio, which does not have a Linux version."

about an FL Studio alternative instead of LMMS you should try Ardour

https://de.wikipedia.org/wiki/Ardour

about OpenOffice vs. MS Office
believe it or not but microsoft is slowly losing the Microsoft Office monopole because google docs already won the war for young people and for shared online editing ability...
also just in case you don't know it the opensource community abandoned the OpenOffice and now all are on libreoffice.

and also about "these alternatives are not yet as good"

I am sure Blender already won agaist 3DSMax...

don't just read my post and do nothing instead test libreoffice and ardour and if you want test google docs.

and then tell me if these closed source monopole overlords are still king or what ?

[qarium]

I would ask you a question then: "How do you get a stable ABI without a microkernel?". Notice that all the attempts to produce such a thing in the Linux ecosystem have failed miserably. A stable ABI does not exist for the desktop Linux ecosystem. How do I know that? If a stable ABI was available, then developers would ship their applications as executables (i.e. binaries), not as source code.

can you tell me what is wrong with ship applications as source code ?

to ship source code is in fact the superior way to do it.

[Emmanuel Deloget]

"Sure, there are open source alternatives but unless I'm completely out of this world, these alternatives are not yet as good as the original (Gimp vs. Photoshop ? Blender vs. 3DSMax? OpenOffice vs. MS Office? LMMS vs. FL Studio? Inkscape vs. Illustrator?...). I have 3 Linux machines at home and everybody uses them. And I still have to use a Windows machine because sometimes, only MS Word can correctly handle the documents that I receive. And some other times, my son want to use FL Studio, which does not have a Linux version."

about an FL Studio alternative instead of LMMS you should try Ardour

https://de.wikipedia.org/wiki/Ardour

about OpenOffice vs. MS Office
believe it or not but microsoft is slowly losing the Microsoft Office monopole because google docs already won the war for young people and for shared online editing ability...
also just in case you don't know it the opensource community abandoned the OpenOffice and now all are on libreoffice.

Sorry, I'm quite old, and I've used OpenOffice for years before I (as almost anybody else with a hair of sanity) switched to LibreOffice. But sometimes, ancient memories takes the best of me and I mislabel LibreOffice and erroneously call it OpenOffice. Again, sorry for that.

and also about "these alternatives are not yet as good"

I am sure Blender already won agaist 3DSMax...

There's a whole industry out there that would like to know about that

If that was true,

1/ it would be used by all major 3D shops around there ; and yet this is not the case
2/ 3D plugins would be developped for Blender, and not for 3DSMax/Maya ; and yet this is not the case
3/ you would see mentions of non-free softwares as contenders to break the monopoly of Blender in the area ; and yet this is not the case.

Blender is quite incredible and can successfully be used in a professional setting - that's already quite a good achievement. But it still has steps to overcome before it can be deemed as "as good as" the current best players in this specific area.

don't just read my post and do nothing instead test libreoffice and ardour and if you want test google docs.

Last time I checked, Google Docs was not an opensource product. Not to mention that while you can use it daily as a personnal-oriented tool, it lacks many of the professional-oriented features of MS Office and is not really comparable to it.

About Ardour, it's a good program. I've used it for years (from the time when it was crashing when you tried to do something too sophisticated with a wav file) and it has grown to become quite impressive. But unfortunately, it does not adress the same problematics as FL Studio, and as such will never be as good as FL Studio. These are really two different tools with *some* common points. But the very fact that you cite Ardour leads me to some kind of (tenuous) link with the original thread: Ardour, as LMMS, can load VST plugins. Unfortunately, most commercial-grade plugins are closed-source and are only available on Windows (and sometimes on Mac). As a consequence, you have to bring up wine in the discussion whenever you want to load a VST on Linux (IIRC LMMS runs a VST server on wine and then connect to this server to use the plugins).

and then tell me if these closed source monopole overlords are still king or what ?

I think you are misrepresenting my words. What you may have read as an attack on Open Source is not, so you don't have to assume that I consider any of the unnamed publishers above as some kind of kings. But you also have to admit that, whatever wishfull thinking can do, it will not make the existing crop of free softwares better than their alternative non-free counterpart. There are reasons for that, the first one being money - huge trucks of money that are thrown to the development of non-free softwares. As a consequence, free software don't have much choice but to try to follow, and this is not an easy task. And how could it be different? How many people are working on non-free softwares at Adobe vs. how many people are working on The Gimp, Inkscape....?

Don't get me wrong: all the programs I cited are already quite good. Once you have accepted their respective limitations, they all can be used in a professionnal setting (I use LibreOffice daily, and I quite often summon up The Gimp and Inkscape). I can even add more: PostgreSQL against Oracle SQL for example, or Scribus vs. InDesign. I use PostgreSQL every single work day and I think I have a good grasp on its stengths and limitations. I have used Scribus to build several documents that have been shared even with the outside world. But as good as they are, our beloved open source softwares are still distant challengers when you compare them to their respective inspiration. That's unfortenately not an attack: that's a fact.

[Weasel]

But there's an additional problem with userspace-to-userspace interfaces as well. E.g. GTK and Qt never bother trying to keep backwards compatibility, glibc does but it's a pain to achieve (you need to pass special linker flags IIRC to tell it to use an older version of the symbols). Linus doesn't have power over userspace, and has indeed recommended they get their crap together and strive for stability. I think a microkernel architecture may help due to the restrictions it imposes, but not necessarily solve it on its own, as it is a very human problem.

What do those libs have to do with the kernel? It's their choice whether they break ABI or not. I don't agree with them at all, but it has nothing to do with the kernel's decisions.

Breaking ABI is a design decision. It has nothing to do with kernel, userspace, monolithic, or microkernel.

Let me put it in simple terms. Linux decided, by design, to not keep a stable ABI for its in-kernel interfaces. But it also decided to keep the userspace interface stable. Design decision.

Windows decided to keep stable ABI for its userspace interface, but also a relatively stable ABI for its in-kernel interfaces. Windows also happens to be microkernel by design.

This doesn't mean microkernel forced them to be ABI stable. In fact, there are ABI breakages sometimes even in the Windows world for in-kernel interfaces. Most drivers written for Windows XP, Windows 7, and Windows 10, are not compatible. They broke the ABI by design. -.-

Note I said "most" because there's a lot of kernel interfaces on Windows (same with Linux). Just because they broke some of them doesn't mean they broke all of them. So some may still work if they used the interfaces which didn't get broken.