Then have the balls to just pick one. I know the Linux community loves their assortment of distros, but if fragmentation is going to prevent advancement then we need to just suck up the fact that there needs to be some stable entry point for outsiders to start depending on. New development editions are great, but nobody can make long term commitments to a moving platform. Especially to one that is so unapologetically mobile. That doesn't happen though if people interested in the platform are only willing to view it as an all distros or none proposition.

Steam drew a line in the sand, and I think it was a very rational, and reasonable choice on their part on a number of levels, and is an example others can start falling in line with. One of the great things about Linux is it is fairly simple to have a second install. If the world decides Ubuntu is going to be the commercial landing point for supported software in the Linux world then great. Lets do it. I am more than willing to have an Ubuntu install on my system if it means content will bother showing up at all. I will probably never need that support, but I can be ready on the off chance I do. In fact I already am.

I'd start with Debian if I were them, since Ubuntu is based on it anyway. I.e. start with supporting Debian stable, testing and unstable (Ubuntu is based on the last one). This will create a great base for many other distros, even not Debian based ones. Surely it'll benefit many Debian derivatives such as Mint and Ubunty directly. If supporting all 3 is too tedious (unstable can be challenging), they can support stable and testing (since Debian testing is the most used for the desktop amongst these 3, and Linux Mint which potentially even exceeds Ubuntu in popularity is switching to it as a base).

Firstly I personally don't consider Steam's choice to be the best possible. Secondly, there is no point to "standardize" it in such way. We aren't talking about standards or making standards, we are talking about preferences of a particular vendor (in this case Steam). If there is a better choice, GOG can go that way, no matter what Steam does. Also, developers making games anyway try to target many distros to gain wider user base, so they won't be limited to Ubuntu or Debian only, just because Steam or GOG prefer them. Thirdly as I said above, going with Debian will cover Ubuntu too, so that wouldn't be far apart from what Steam does, but on the other hand would bring more benefits, since Debian has more downstream derivatives than Ubuntu itself and it's a better meta distro for this scenario. And lastly Steam is known to work on Debian anyway, so here you have it.

Something doesn't need to be "the best choice" in order to be what ends up being on top, somethings just end up on top, and if there is nothing horrifically wrong that thing I don't personally want to inject conflict into a messy situation.

IF fragmentation is (and some argue it isn't) a problem then conflicting vendor preferences is a problem complicating a problem, and disharmony on top of disharmony. Consumers aren't going to be willing to chase vendor preference around the block anymore than GOG is willing to chase our distro preference around it. It's all a muddled mess a lot of people just are terrified of getting into. Even if the perception is worse than the reality, perception becomes reality and nothing happens. We all just go back to our assorted corners of a world with a very large number of corners.

If the Linux community has any interest in letting in the commercial world, then I think it needs to be willing to let there be some kind of stable side to the world to develop for. Since it is so open we are constantly getting full access to every change and variation, and that has its down sides. I'm not saying people should have to give up their pet distros or lose access to what is current. I am saying there should be some place in the Linux world where people can go without the constant sound of jackhammering, and the need for dodging orange cones. I will go as far to say that if we don't come up with such a place then it will be a huge impediment to getting people like GOG on board, and commercial growth will continue to be the frazzled mess it has been thus far.

Unfortunately I don't think the Linux community has any say in what that might be. It is up to the vendors to decide what they are willing to do, and the only thing we can do to help is not fight them just because we would have made different choices.

Of course I don't know that any of this is true, and I kind of hope I am wrong because I don't see people as being willing to compromise in such a way. I'd like to think that we will get there(wherever that is) without having to change the environment, but I have a difficult time picturing it.

I think what you are looking for is hardly attainable at the distro level. I.e. consider this scenario: Steam being comfortable with Ubuntu adds more distros to their official list (let's say OpenSUSE and Mint for starters). Do you imply that this choice should suddenly be considered standard and every other games distributor should start supporting these additional distros? I don't think so. As I said, Steam preferences don't dictate you what to do. It's good if there is enough convergence of course, but there is no need to blindly copycat every step.

What's more appropriate though is some common denominator from middleware perspectives. I.e. let's say you come up with base technologies and drivers that are supported (X or Wayland, Nvidia closed driver? Intel open driver with certain high end integrated chips? etc.). That makes more sense. Trying to placate Steam choices as standard just doesn't cut it.

Absolutely not. I'm talking about a first step, a common entry point. From there it's whatever, but if there was some common ground baseline that everyone could live with then Linux just might get their foot into a few more doors.

The only reason Steam is relevant here at all is because of their size and being the first large entity to say we aren't supporting a mess of 20 distros, we are going to do this for one and see what happens. The single supported distro angle works, but if GOG tries it and picks a totally different distro, well then people start having to ask "am I the GOG guy or the Steam guy?" and I don't think that's helpful (though I will take it).

All of this is really just political, and perceptual though. If we can just get someone like GOG into any door at all, then chances are we can get their offerings to work across a good deal of the spectrum. Which really is what we all want. Unfortunately they are afraid some angry Linux boogieman, that apparently can't figure out video games don't run on e-ink displays, is going to rip them a new one. That and multiple distros.

What worries me is singling out one distro and saying: let's support just this - it can backfire. In case of Ubuntu specifically, since lately Canonical made several steps that put them more apart from other distros than before. For example with introduction of their own graphics server - Mir. The rest of the Linux world is transitioning fom X.org to Wayland (and it will still take some time until Wayland drivers will be widely available and major DEs like KDE and Gnome would adopt full Wayland support).

Canonical on the other hand happily marches from X.org to Mir. Now, do you want to be tied to one such distro, while the rest is clearly moving in the different direction? And specifically in the context of graphical stack which affects gaming the most!

This requires serious consideration and as I said above better approach is not to focus on one exclusive distro which can happen even not to represent the majority of Linux users, but to focus on some common middleware components.

Fallout was released here on late 2008 (or was it early 2009? Not sure). GOG doesn't have the source code for it. GOG provided support. How? Did they go looking for Black Isle to ask for assistance? Did they go to Interplay to ask them to fix the patches? Did they send Fargo an e-mail asking him if he knew who they could contact? Or did they do support themselves?
Alternatively, Litil Divil was released here last year. Developer was Gremlin Interactive, which is long gone. Who supports the game, Gremlin Interactive or GOG?

So tell me, how can they fix bugs without sources for various games which don't have any developers behind them? Does GOG reverse engineer these games? I'm genuinely interested, since I don't really understand how else you can "support" such things. Alternatively, may be GOG contributes to such projects as DosBox and Scummvm to improve playability of some old DOS games in those emulators. But that's not relevant to Windows games at least. Some games however (more recent ones) do have developers and existing sources. How does GOG support them? Also fixing the bugs themselves through reverse engineering, or through the developers?

Depends on the bug I'd say. It could be a ddl replacement to allow proper DirectDraw usage (JA2 on Win8 for example), it could be a wrapper for better emulation, it could be hex editing of the file to allow a different check of the APIs, or it could be a special build of DOSBox to allow better compatibility.
And please do note that there are 2 kinds of support. One is game bugs, which GOG can't (usually) fix, like the AI Deadlock on JA2, while the other is execution bugs, which GOG can (usually) fix, like the JA2 not working on Win8 with the vanilla dll files, or the mouse cursor leaving trails and screen not refreshing if run on an SLI/Crossfire setup.
I am using JA2 as an example since I have had quite a bit of experience troubleshooting it, both the vanilla and the 1.13 version, but it should be more or less the same for most games.

Not so much 'reverse engineering' as applying workarounds. In the case of Windows games, all they do is make tweaks or set compatibility settings that would be required to run the game under modern Windows OSs.

I don't know of any game where they've actually fixed bugs. Hell, they even left Gothic 3 in its original buggy state.

Edit: Ninja'd.

OK, so GOG does a significant amount of tweaking on old games, and possibly contributes to DosBox and Scummvm. But still, what's the point for GOG doing that for more recent games, if their developers anyway plan to update them with bug fixes. I don't think they would spend resources on that, if they can direct bug related info to developers and get patches back. Note that scenarios with reverse engineering that you described mostly apply to old games. Recent ones are in different situation.

Since this discussion was in context of Linux, I think you'd mostly expect recent titles, or at least recent ports of some games. I.e. that's what you see on Humble Bundle for example. These have active developers behind them, so I doubt GOG would find themselves in position of reverse engineering to fix those games bugs, same as with their recent Windows games I suppose. Therefore the argument of super duper heavy burden of support for these Linux titles which is in some way very different from developers support offered by Humble Bundle, doesn't sound that convincing to me.

Let me put it this way.
I bought my Toshiba Laptop from a local large chain retailer. If it has any troubles while under warranty, I will take it to the retailer. What is better, the retailer saying "Sure, our technicians will get right on it" or "Sure, we'll send it to Toshiba and they'll get right on it"?
GOG has the very good habit of doing all troubleshooting and support themselves, thus being able for a bit better response time and result. Sure, for newer games I do believe they also send the bug reports to the developer so they can make a fix, but even that takes some time. See for example Torchlight and Mod support, which I'm still not sure why it was broken on GOG's release.

OK, so you agree that for recent games GOG gets reports on bugs from GOG customers, and directs them to the developers. That's exactly what Humble Bundle does as well. I don't see any difference in context of Linux or Windows in this regard. Or you are saying that GOG also asks for sources from these developers to help debug these issues?

In case of Humble Bundle by the way, they actually do manage sources in some way in order to make ports (at least of some games). I.e. they offer porting opportuninty for games which don't have a Linux version yet.