id Software: Linux Hasn't Produced Positive Results

[Scali]

But that's exactly what you did, and what you are continuing to do in your post.

No, it is Linus who put the blame on nVidia.
I merely point out the fact that this is beyond nVidia's control, yet within Linus' control.

You then make statements that are contrary to reality and use them as a base for your arguments. Nvidia 'could' open source their drivers. The only thing that prevents them from doing so is the imaginary spaghetti monster that they have told you about.

That's your opinion. Apparently nVidia and virtually all other IHVs and ISVs have a different opinion.
They could all be wrong, but that's not even relevant. Fact is that they don't consider open source an option.

I know, and respect, Nvidia's choice. I just won't pretend that they didn't have, or make one. Especially when we have living examples of companies making the opposite choice (intel).

As I said, Intel is in an entirely different situation from nVidia. nVidia has a huge capital invested in their GPUs and drivers. Intel does not.

[Scali]

.Net is being more and more tightly integrated into windows, currently it is VERY hard to use windows without .net installed. I don?t think you can draw .net as a "non-native" anymore. But it was just an example. There is VC/VB runtime libraries as another example.

I'm just pointing out that they are separate platforms. A .NET application is tied to the .NET platform, but ironically enough, that means it is no longer explicitly tied to the Windows platform (as demonstrated by Mono).
Also, you STILL don't get the point. No matter how much MS 'integrates' (ooh, there's that word again, what does it mean? Heard it with GUI before, and IE... whatever) .NET into Windows... existing non-.NET applications are not affected by anything that happens inside the .NET framework, because they simply never even start the whole .NET runtime, let alone that they call any actual .NET functions.

For .net, you need at least 6 different versions - releases and servicepacks.

Not really. There are only 3 major versions of the .NET framework, namely 1.1, 2.0 and 4.0. These have physically different specifications of the VM, and as such they are not directly compatible.
All the other versions are simply updates and extensions of the existing frameworks, and are fully backward compatible (eg .NET 3.5 is the .NET 2.0 VM with various new APIs).
However, Microsoft lets you run the different versions seamlessly side-by-side, and you can even combine 2.0 and 4.0 objects in the same application.
.NET 1.1 was never all that widespread, so you generally don't even need to install it. Vista and Win7 came with .NET 2.0 out-of-the-box, and 1.1 would have to be installed separately if you required it. But most people don't.

This is newer approach to "compatibility" dillema - instead of mixing everything together (DLL hell), they are use "isle" approach that - still eats a lot of space, minimizes yet does not solve improvement breakage.

It's not really different. .NET uses a VM, which technically is a 'CPU'. A new version of the .NET VM could be seen as a new CPU released by Intel or AMD, with new instructions etc. You generally install new OSes and new software to make use of these new instructions as well (such as x64 versions of your OS and applications, or SSE/AVX-optimized builds).
Inside the header of a .NET binary is a version number that indicates on which revision of the .NET framework it must run. So a 2.0 application won't run on the 4.0 VM in the first place, so any compatibility issues are pre-empted.

From my perspective, it is much better for developer to track changes within the libraries he uses for his applications - and to update them. This is the only single trouble.

Well, my perspective is obviously different, what can I say?
To me the tradeoffs you're suggesting are bad. You'd have to invest a lot of effort in maintaining your application, rather than actually developing it.
Just keeping the application running on people's systems becomes a dayjob in itself. That's just wasted effort.

Wine is free implementation of winapi over unix, so I think it is pretty objective and accurate - as it per definition tries to copy 1:1 winapi behaviour.

Wine is a project by and for frustrated linux users who can't quite give up their hated Windows applications.

Backwards compability is the reason for 95% all problems in windows.

Uhh, no it's not. Poorly written code is the main problem. Like people writing to Program Files and restricted parts of the registry (although Windows 7 implemented virtualization for that). Although Windows did not enforce it prior to Vista and the new restricted user defaults/UAC, Microsoft has always published best-practices documentation, and always warned against doing such things. Therefore, properly written software from the early XP era, or even before that, will still run fine under Vista/Win7 with UAC, without triggering popups, needing some virtualized environment, or any other potential security issues.
It's those developers who didn't RTFM that caused problems. Same happened with Win9x to NT-based Windows. Because Win9x only has a partial implementation of the Win32API, not all parameters of each function were used or validated. So stuff that 'worked' on Win9x by accident, would break on NT because of wrong arguments being passed. As any developer should know: "Working code is not bugfree code"
Just because it worked on one version of Windows doesn't mean it's the OS that breaks the code. It was already broken, as you can see in the documentation.

I repeat, yet again for third time - it is not your job to care what version linux distribution uses.
If you want to package your software for distribution yourself, just package it. If they use outdated library versions - notify them, that they should use lastest stable.
As easy as that.

Oh I see, you are under the assumption that everyone releases sourcecode, and then prays that distro makers will package it, so your end users can download it for free?
Well, not quite. If you don't want to release sourcecode, you have no alternative but to build and package it yourself. Which is common practice on most popular platforms today, eg Windows, OS X, iOS, Android.

It does not put effort in "compatibility", because all software is meant to be compatible by default.

Nope, it isn't. Linux and many open source projects have a long history of breaking API changes, requiring code to be recompiled, or even rewritten in order to work after an update.
I should know, I've been running FreeBSD since the late 90s, and generally install my software through ports. Getting access to newer versions of applications means having to update the ports tree. As a result, your dependencies will be updated as well. Quite often, builds would break halfway through, and you'd have to manually update other projects, because they weren't caught by the dependency mechanism in the ports system itself.
In some cases you even had to manually edit source code in order to keep things working, because the ports tree is not always perfectly in sync throughout.
And I've even had cases where there was no alternative but to just upgrade the entire system in order to get access to certain software.

And this is just ONE 'distribution'.

Change by definition is breakage.

Ofcourse not. Sometimes change is even 'fixage', like when bugs are fixed. Other times changes just mean optimizations, or adding new features. None of those changes break existing software.

In order to critisize, you should at least have major experience in BOTH of the subjects. You lack the experience in one of the subjects and try to stretch one approach on another, painfully (but luckily not too much) reacting to anyone who subjects your own opinion.

Eh, what 'both subjects' would those be? I hope one of them is not 'Windows', because if all you see is Windows and linux, that's rather sad. I don't exactly qualify as a 'Windows developer'.

Get Gentoo, use it as primary OS for at least 2 years straight, then return to the topic. More or less, this would be a good experience for you.

Lol whatever. You mean that linux distro that was a fad a few years ago because 'OMG I can builds from teh srcc0d3z!'? Like FreeBSD has been doing years before anyway? Which I've been doing myself?
Yea, I'm sure Gentoo will be very valuable to me...

My own admission is to write once and package anywhere.

Yea, not an option if you have to do all the packaging, as mentioned above.

No offence, but if you need an ophthalmologist, I know one.

First look at yourself, before trying to be smug.
I will repeat:
You: "You also can't use Android 4.0 functionality on 2.0. Is this somehow different?!"
Me: "Yes it is [different]. Becase you *can* use Android 2.0 packages on Android 4.0."

So I explained how it is different. Then you respond by saying: "No you can't [use Android 2.0 packages on Android 4.0]."

Linux is exactly same

No it's not, you said so yourself: "[Build and] Package everywhere".

Humans make mistakes. Mistakes come out because humans are a) desynced, b)imperfect by design. b) is vital, because otherwise, humans will start nuclear war due to boredom.

Some users also like to be smug b'stards, even though they are 'correcting' someone who's right and knows what he's talking about, while they aren't and don't.

One can?t satisfy everyone. One will overfill and explode. I like this OpenGL approach, but lets see if this works out and be better than a previous one.

The fact that OpenGL is completely ignored by virtually all developers on the only OS that gives them any choice at all (Windows) seems to indicate that most developers don't like the OpenGL approach (yes, the few that do are very vocal about it... problem is, they're not the ones writing professional software like 3DSMax, Autocad, or multi-million dollar games).

[Scali]

Fact 1: [Carmack saying: "Not nearly as many people are interested in paying for games on the platform."]

Fact 2: [Josh Walrath allows himself, in interpret Carmack as: "Linux development is another story altogether. Even though Valve is now actively pursuing the Linux market, iD has been there before, and just has not seen positive results. Remember how many past titles from iD actually ran on Linux, and for how long these were supported? John says that Linux development simply does not pay the bills. It creates goodwill among the Linux crowd, but that is about it."]

Fact 3: [Phoronix & other websites cite Josh Walrath opinion as that of Carmack]

Fact 4: [Linux users get angry, because everything cited by Fact 2 is false].

What you have done is:

Fact 2 -> Fact 4, which is correct, but origin is false.
But then: Fact 4 -> Fact 2, which is invalid implication.

Fact 4 & Fact 1 will, however, very probably catalyse themself now. This was started by Carmack and is his choice, nothing to be done here.

That's not entirely how it went. Your 'fact 4' is not the ONLY time linux users ever got angry, or linux developers refused to co-operate etc. This flawed assumption of yours invalidates your entire case.

[finalzone]

I'd respond, if I had any idea what the heck you're trying to say.
The different ecosystems were the whole point (that's the difference between a kernel and a platform. You don't develop software for a kernel, you develop software for a platform. To a large extent, the kernel is irrelevant. Which obviously the linux community will never admit. But I am not as naive as to expect that anyway)... More specifically the different culture and vision of Android vs linux distributions.

Say Android/Linux and GNU/Linux to be correct. In addition, Mozilla prepare Firefox OS with use HTML5 engine platform based on Linux kernel meaning it is another Linux distribution. When Microsoft decide to replace their NT kernel by Linux, would you call it Windows or Win/Linux? So yes, in that context, kernel is relevant in term of semantic. Do you know Debian is not only a Linux distribution, it is also BSD based on kfreebsd kernel?
Careful what you said in front of audiences who challenge your own definition of Linux.

[Scali]

Say Android/Linux and GNU/Linux to be correct.

Ahh, no.
Android is an official trademark of Google. It is just that: 'Android'. Not 'Android/Linux' or anything. Android, period.
As for GNU/Linux, that's what RMS wants you to call it. There doesn't seem to be a lot of consensus about this.
For example, I don't see any reference to GNU here: http://www.ubuntu.com/project/about-ubuntu
Now that's the largest linux distribution available today, I think I'll just go with that.
In fact, note also that they refer to their OS distribution only as 'Ubuntu', not as 'Ubuntu Linux' or anything... and certainly not 'GNU/Linux'.

When Microsoft decide to replace their NT kernel by Linux, would you call it Windows or Win/Linux?

That depends on what Microsoft wants to call it. I am not sure why you think the kernel name should be in the name of the OS as well, because that is generally not the case. As far as I know, the Windows kernel doesn't even have a specific name. It's just called 'Windows kernel' because it is named after the OS, not the other way around.
Likewise, OS X uses the XNU kernel from the Darwin project, but neither XNU nor Darwin are in the name 'OS X'.

Careful what you said in front of audiences who challenge your own definition of Linux.

Just because they challenge things doesn't mean they're right. You certainly aren't!

[ninez]

Wine is a project by and for frustrated linux users who can't quite give up their hated Windows applications.

No it's not. that is nothing more than your own silly stupid opinion. I use Wine mainly for running VSTs with Jackd. - this allows me to run various freeware and commercial virtual instruments and effects plugins on my linux desktop. It wasn't a question of 'giving up' or 'hating' this software, since I didn't actually 'migrate' from Windows, but instead use both Linux and Mac - the Windows versions of these plugins are included in the ISOs that i bought to use with my Mac. (obviously, freeware is downloaded). So explain to me exactly how this makes me a 'frustrated linux user who can't give up their hated windows software'?! ~ here's a hint: it doesn't at all. It actually just means that i am taking advantage of software that allows me to run applications designed for another OS - which IS the point of wine.

Your definition of Wine would be like me saying 'people only use VMware because they can't give up there hated *insert OS here* applications ~ which is obviously severely MORONIC. There are many valid reasons to use VMware (or Wine) other than the small-minded reasons that you claim.

That depends on what Microsoft wants to call it. I am not sure why you think the kernel name should be in the name of the OS as well, because that is generally not the case. As far as I know, the Windows kernel doesn't even have a specific name. It's just called 'Windows kernel' because it is named after the OS, not the other way around.
Likewise, OS X uses the XNU kernel from the Darwin project, but neither XNU nor Darwin are in the name 'OS X'.

As far as you know the Windows Kernel doesn't have a specific name?!?! WTF?! - you really mean to say you have never heard of the 'NT Kernel'?!?!?! ~ which MS has been using for a VERY long time, somewhere around 1988-1990 and still use it today;

http://www.ntkernel.com/

http://en.wikipedia.org/wiki/Architecture_of_Windows_NT

http://en.wikipedia.org/wiki/Windows_NT#Development

http://www.reactos.org/en/about.html

You might want to have a read through. it's not called the 'Windows Kernel', it's the NT kernel that windows uses. You also should note that various versions of Windows are in fact known as Windows NT (citing the kernel in the name).

You also seem to not realize that XNU stands for 'X Not Unix' ...and what is the name of Apple's operating system?? -> MacOS X... it looks to me like they correspond with each other perfectly. You see, the operating system isn't called MacOS (specifically), the specific part is the "X" (meaning 10). Where as MacOS 9 was a totally different OS, built on a totally different architecture (PPC) and vastly different underlying components.... You might actually want to make sure you know what you are talking about before making such comments, they really do come off as uninformed and dumb.

One last point about developing in Linux. Scali, you do realize that you don't always have to track libs your application is using right? there does happen to be an alternative that lots of developers use - it's called bundling a specific version of xyz library (static) with your application. I have a few applications on my systems that do thi and i am aware of others that do as well, for example AutoDesk Maya bundles various libraries ~ and thus don't experience ANY breakage in Maya, having to do with system lib versions changing.

[Scali]

I use Wine mainly for running VSTs with Jackd.

Aside from the obvious anecdotal evidence/exception-to-the-rule fallacy...
How is this not running Windows software on linux (where obviously you could just run it on Windows... VSTs being the native format of Steinberg's Cubase, and apparently your VSTs are for the Windows version of that. There is no linux version btw).

Your definition of Wine would be like me saying 'people only use VMware because they can't give up there hated *insert OS here* applications ~ which is obviously severely MORONIC. There are many valid reasons to use VMware (or Wine) other than the small-minded reasons that you claim.

VMWare is also used for running Windows-on-Windows, or linux-on-linux etc. Hardware virtualization is something completely different than Wine.

As far as you know the Windows Kernel doesn't have a specific name?!?! WTF?! - you really mean to say you have never heard of the 'NT Kernel'?!?!?! ~ which MS has been using for a VERY long time, somewhere around 1988-1990 and still use it today;

Sure I know about that... but that is the NT kernel of Windows NT, again the kernel named after the OS (Windows NT goes far beyond just the kernel, the whole Win32API is part of NT, and was backported to Win16 and later Win9x variations).

You also seem to not realize that XNU stands for 'X Not Unix' ...and what is the name of Apple's operating system?? -> MacOS X... it looks to me like they correspond with each other perfectly. You see, the operating system isn't called MacOS (specifically), the specific part is the "X" (meaning 10).

Yes, so the X stands for 10, not for 'X Not Unix'. It's the same thing as Windows NT still being called Windows, but essentially being an entirely new OS. Has something to do with strong brandnames and all that.

Where as MacOS 9 was a totally different OS, built on a totally different architecture (PPC) and vastly different underlying components.... You might actually want to make sure you know what you are talking about before making such comments, they really do come off as uninformed and dumb.

Hold that thought there... OS X initially also ran on PPC, you know.. Apple didn't make the transition to x86 until a few years later.

[crazycheese]

I'm just pointing out that they are separate platforms. A .NET application is tied to the .NET platform, but ironically enough, that means it is no longer explicitly tied to the Windows platform (as demonstrated by Mono).
Also, you STILL don't get the point. No matter how much MS 'integrates' (ooh, there's that word again, what does it mean? Heard it with GUI before, and IE... whatever) .NET into Windows... existing non-.NET applications are not affected by anything that happens inside the .NET framework, because they simply never even start the whole .NET runtime, let alone that they call any actual .NET functions.

Yes, non-managed code is still allowed. They haven?t touched it, yet.

Well, my perspective is obviously different, what can I say?
To me the tradeoffs you're suggesting are bad. You'd have to invest a lot of effort in maintaining your application, rather than actually developing it.
Just keeping the application running on people's systems becomes a dayjob in itself. That's just wasted effort.

Thats decision from people "above", and they find it ok - a fair trade-off for a clean ecosystem.

Wine is a project by and for frustrated linux users who can't quite give up their hated Windows applications.

Quite a unique approach you have here - so, you are using your OS because you can?t install Linux or what?
Nope, WINE is a free implementation of winapi on unix - no more, no less.
Exact cases why people use it are different - and they are multiple, but some are to avoid instability and insecurity of windows without porting and maintaining whole source, due to original author having quite little time resources.
In best case, you don?t need wine.

Uhh, no it's not. Poorly written code is the main problem. Like people writing to Program Files and restricted parts of the registry (although Windows 7 implemented virtualization for that). Although Windows did not enforce it prior to Vista and the new restricted user defaults/UAC, Microsoft has always published best-practices documentation, and always warned against doing such things. Therefore, properly written software from the early XP era, or even before that, will still run fine under Vista/Win7 with UAC, without triggering popups, needing some virtualized environment, or any other potential security issues.
It's those developers who didn't RTFM that caused problems. Same happened with Win9x to NT-based Windows. Because Win9x only has a partial implementation of the Win32API, not all parameters of each function were used or validated. So stuff that 'worked' on Win9x by accident, would break on NT because of wrong arguments being passed. As any developer should know: "Working code is not bugfree code"
Just because it worked on one version of Windows doesn't mean it's the OS that breaks the code. It was already broken, as you can see in the documentation.

So, even applications such as Adobe Director are poorly written? The problem is, you can?t keep anything as before if you change stuff - you can hold up older interfaces, till appoaches completely differ and then they are broken no matter how good they are written. Software is a process. So, you idea is to put the weight of maintaining ABI compatibility on library and kernel hackers shoulders. Their idea is however, no matter how long you maintain, it still will break at some time. Linux is a moving target, Windows is a lagging target. I?m quite sure most of the programmers are capable to catch up and there is no need for (effort of) compability layer. As I said, being Linux user for 5 years, I have never seen segfaults on a stable version of a supported application.

Oh I see, you are under the assumption that everyone releases sourcecode, and then prays that distro makers will package it, so your end users can download it for free?
Well, not quite. If you don't want to release sourcecode, you have no alternative but to build and package it yourself. Which is common practice on most popular platforms today, eg Windows, OS X, iOS, Android.

Frankly speaking, I avoid closed source completely where possible, so I can?t help you here.

Nope, it isn't. Linux and many open source projects have a long history of breaking API changes, requiring code to be recompiled, or even rewritten in order to work after an update.
I should know, I've been running FreeBSD since the late 90s, and generally install my software through ports. Getting access to newer versions of applications means having to update the ports tree. As a result, your dependencies will be updated as well. Quite often, builds would break halfway through, and you'd have to manually update other projects, because they weren't caught by the dependency mechanism in the ports system itself.
In some cases you even had to manually edit source code in order to keep things working, because the ports tree is not always perfectly in sync throughout.
And I've even had cases where there was no alternative but to just upgrade the entire system in order to get access to certain software.
And this is just ONE 'distribution'.

Its same as with Gentoo, although it can track and rebuild affected packages by itself. Being unable to install something and having to upgrade the system is rather poor approach I?d say. It is best you keep the entire system up-to-date, then no parts lag behind and cause circular stuff.

Ofcourse not. Sometimes change is even 'fixage', like when bugs are fixed. Other times changes just mean optimizations, or adding new features. None of those changes break existing software.

You absolutely need to break something in order to change it and as a result, one of the characteristics changes - be it newer functions (and shifted adress space), different memory footprint, different load. This has different impact from introducing newer bugs, surfacing hidden ones or actually making it better, but breaking something that depended on it now. I see it as "better is an enemy of the good", it worked for me..

First look at yourself, before trying to be smug.
I will repeat:
You: "You also can't use Android 4.0 functionality on 2.0. Is this somehow different?!"
Me: "Yes it is [different]. Becase you *can* use Android 2.0 packages on Android 4.0."

So I explained how it is different. Then you respond by saying: "No you can't [use Android 2.0 packages on Android 4.0]."

"No you can't [use Android 4.0 packages on Android 2.0]"

Cmmon, stop perverting my quotes, you can do it

Some users also like to be smug b'stards, even though they are 'correcting' someone who's right and knows what he's talking about, while they aren't and don't.

Sir, you are so self-sufficient and all knowing that I don?t know if this conversation actually makes any sense. You can go troll with Linus, I don?t really want to waste your genious.
But please don?t push on porting windows registry in linux, or that "compability" thingy, so that linux space is poisoned by a mirriad of outdated breaking software.

The fact that OpenGL is completely ignored by virtually all developers on the only OS that gives them any choice at all (Windows) seems to indicate that most developers don't like the OpenGL approach (yes, the few that do are very vocal about it... problem is, they're not the ones writing professional software like 3DSMax, Autocad, or multi-million dollar games).

Lets talk about Valve or Blender then.
Also, you are again forgetting about WINE - an effort to port closed source system is much more difficult than simply getting open spec?ed stack cast down for free.
So, it is not correct to all Windows as an only "OS that gives them any choice".

[russofris]

No, it is Linus who put the blame on nVidia.
I merely point out the fact that this is beyond nVidia's control, yet within Linus' control.
That's your opinion. Apparently nVidia and virtually all other IHVs and ISVs have a different opinion.
They could all be wrong, but that's not even relevant. Fact is that they don't consider open source an option.
As I said, Intel is in an entirely different situation from nVidia. nVidia has a huge capital invested in their GPUs and drivers. Intel does not.

I think that you're off base, and need to give the matter additional consideration. If I say that the earth revolves around the sun, and you respond with "That's your opinion", then there's really nothing more for us to argue about. When I show you an example of planets revolving around other stars and you say "Other ISV's have a different opinion", it confuses me.

What propaganda did you consume, or what trick of intellect did it take, to make you choose to operate in such an injured fashion?

Yesterday I told my 2-year old to clean up his matchbox cars. He proceeded to lay on the floor a couple inches away, extend his arm towards the cars, and say "I can't do it!". What nvidia/amd/etc is doing is no different.

F

[johnc]

I think that you're off base, and need to give the matter additional consideration. If I say that the earth revolves around the sun, and you respond with "That's your opinion", then there's really nothing more for us to argue about. When I show you an example of planets revolving around other stars and you say "Other ISV's have a different opinion", it confuses me.

What propaganda did you consume, or what trick of intellect did it take, to make you choose to operate in such an injured fashion?

Yesterday I told my 2-year old to clean up his matchbox cars. He proceeded to lay on the floor a couple inches away, extend his arm towards the cars, and say "I can't do it!". What nvidia/amd/etc is doing is no different.

F

What point are you making? That NVIDIA has the capacity to open source their drivers or that they can make Optimus work on Linux? Because neither one seems reasonably feasible at this point.