This is dishonest, if they also claim at the same time "they are working on it". Doing nothing is not call working.

I'm not talking about a shim, but about full functionality needed for games, not needed for GUI client. I.e. GOG can perfectly split tasks like updating from tasks like multiplayer and etc.

I really don't think they are going to just provide any form of libraries without the full functioning client. It'll be considered half assed by enough people that they'll get more criticizm from that than not doing anything at all IMHO. Plus they'll be advertising such games for Linux and people will be buying them thinking they're getting the full Galaxy experience to get let down after the fact that the client isn't available. The natural response to that from people will be to have them advertise on the game page "Galaxy Client functionality not available at this time, coming soon...". At a certain point it ends up being a huge ugly mess of half-assed compromises on top of compromises and excuses rather than a professionally produced product though.

They appear to be doing all or nothing, and it is currently in a state of nothing, with no evidence of all nor something, but hopefully by talking about it we might spark some fuel in them to actually do something. Personally I expect either no response, or a stock trademark "It's being worked on but will take some time, we can't give release dates at this time, hopefully you can understand." but they've more or less already said that so I doubt they'll repeat it too soon, especially into the line of fire so to speak. :)

If a store page states it supports Linux, then it must do what it takes to actually support the platform not just add a Linux category and an installer script which doesn't check for and install any missing required libraries (unless it only works on the "supported" Ubuntu and Linux Mint distributions which I don't use for various reasons).

It is 2017, there are tons of multiplatform libraries for almost everything, and especially for GUI. Choose a write once, run everywhere programming language and a lot of the porting work is non-existant. Write code using proper standards and there will be less of a need to write custom code for the different platforms.

Basically, what GOG is lacking is knowledge, programming skill and good architecture designers.

I'm not really sure what you are calling "full Galaxy experience". Tasks like updating games should be completely orthogonal to stuff like multiplayer support. The later should be baked into games, which implicitly will use client libraries, and the former can be handled by GUI client. What's the reason to conflate the two, and then stall releasing anything for Linux, because they can't figure out how to make GUI cross platform enough?

I feel the same. It sounds like they can't solve issues that shouldn't even have been issues to begin with, if they would have planned things properly.

they do use QT for their GUI, they use Chromium Embedded Framework for their embedded browsing, they do use Poco C++ libraries for network/other stuff, they do use xdelta to do their incremental patching, and sqlite, zlib, ....
Those are all open-source, multiplatform libraries. So I dare say they paid attention to the points you mention.
But there is only so much code that you can share across platforms.
Especially for a application whose main task it is to install something, you simply can't avoid platform specific code.
Handling/resolving dependencies, handling user rights, desktop shortcuts, menu entries, launching a background service task and communicate with it, ...
those things simply work different. And writing the application in C# or java instead won't solve the problem either.
Just compare their linux installer with their windows installer and tell me how much in common you see there ?

The "write once, run everywhere" is a nice paradigm in theory, but is incredible naive to assume it's going to work for every application.

And that's not even mentioning the infrastructure/tools that are needed to allow the devs to directly upload/download/test their linux builds.

Interesting math, even though unsustaned by sources of information. But it can be applied to Mac as well, just replace "1%" with "3%" and the final result with "0.09%". But somehow Mac is viable OS, and Linux is not. :(
IMO, the descision to support or not to support an OS is mostly based on managers' perceprion of OS in question. Mac OS is viewed ar OS for rich, successful people with lots of money and taste for good things, so MacOS is a must. While Linux is just a toy for geeks and those who can not afford Windows license (needless to say all Linux users are potential pirates because they share that wacky "Free OS for Free People" believes), so it is not feasible to support Linux. (Never mind Mac's generally crappy hardware, "Intel in place of video cards", buggy, outdated OpenGL and proprietary Apple-specific Metal™).

In regards to handling dependencies, user rights, desktop shortcuts, etc. they are only "supporting" Ubuntu and Linux Mint (which is Ubuntu based) so they only have to worry about making it work on that particular flavour of Linux Distribution. Of course, that would still be bad for those that are not using Ubuntu or Linux Mint, but it would still be better than the situation we currently have.

Writing the application in C# or Java allows them to not worry about the differences in the libraries and compilers that are installed on the user's system (C++ 2015 or 2013 redistributables? Glibc or Musl?). Having just a Java or .NET Framework version to worry about is way easier.

As for the infrastructure/tools? The OS and the tools are all free and open-source. They basically have 0 software costs in regards to that compared to the Windows and MacOS platforms which have high software licensing and maintenance costs.

missing the point.
Handling those things on Linux(regardless which flavor) is different than handling those things on windows. And using a multiplatform GUI library won't magically make these things work on Linux.
In the end, if you want to support a new platform, you need to invest additional manpower/development time.
There is no magic multiplatform framework "make it once, deploy everywhere" with no additional effort.
That's just fantasy and bad advertising.

it is not about the dependencies of the application, but about the dependencies of the games they want to install.
Handling dependencies of games written for linux requires different platform specific code than handling dependencies for games written for windows.
Writing your application in C#, Java, python or whatever won't change a damn thing about that.

and ? software cost was never even a talking point here ...

That's for sure, but GOG aren't supposed to be some amateurs. How much time should it take them to hire needed programmers and implement all this? Unless they decided not to do it [soon], I don't see what's the blocker here.

Regarding dependencies, I'd say it shouldn't be client's problem to manage them. But it would be nice for it to probe them, and report to the user if something is missing or clearly incompatible. That's going to be Linux specific, no doubt about it.

By full Galaxy experience, I mean the exact same experience that one has available in Windows being available in Linux across the board. That means the full blown client, and all other API and back end services.

They can figure out anything ultimately. It's only a matter of manpower and time. They could create anything with just the right amount of these resources.

Admittedly, we can all bikeshed on this. :) We don't know what they've actually done however or what the details are as to why something isn't available for Linux yet. On the bikeshedding angle, if I were to set out to develop a project of this nature my approach would be to develop everything cross platform from day one on equal ground, utilizing as many cross platform pre-existing open source/free libraries out there as possible in order to not have to develop such things myself. They appear to have done that with quite a few elements of Galaxy (just look at the libs it is dependent on, they're all cross platform I believe). I would extend this to make the GUI cross platform as well, using one of the toolkits out there such as GTK+, Qt or similar, and I'd have the build system build on all platforms with every commit from a developer, with a build failing on one platform auto-killing the builds on all platforms. This is considered strict by many developers out there (notably Windows centric ones <grin>), but for any loss of productivity in getting things going on platform X, it increases productivity in getting and keeping things running on platform Y and Z, etc. On a personal note, my prime development would actually be done and tested on the Linux platform first before submitting a build, correcting things for all platforms, then testing on all. If there's one platform that would work for sure it'd be Linux. :)

Sadly (for me), that's not how the majority of the software development works however, and most shops that are Windows centric would likely have me put in a straight jacket at the thought. :)

I've no idea what GOG's internal pipeline is like in that regard, but based on what we see outside it appears they develop on Windows first then port to Mac, and any Linux work appears to be 100% opaque. I take them on their word that they'll have a Linux Galaxy release some day, but they're only a small team and it appears that they can't both work on a major new Galaxy release and simultaneously get the Linux bits out the door as it would seem the people who would do the Linux stuff are needed to do the core client stuff. Either they've got to have the entire team working completely cross platform always to get it out the door, or they have to have multiple teams dedicated to each platform, or they need to shift gears and have the one team work on one platform at a time in some form of rotation, the latter yielding disparate releases.

It's likely that they did the best with the limited resources they've had to date and had to make some tough choices that caused this platform fragmentation. It happens. Moving forward though, I hope that the company is making a lot more revenue now and are able to hire enough developers to work on Galaxy to have the platform come together as a unified codebase in the future that has simultaneous releases on all 3 of the target OS platforms. If not now, then hopefully next year, and if not then, then the year after. If not then, then definitely once CP2077 comes out and they're tripping over crates of money in the office. :)

I.e. they aren't dedicated to Linux support of Galaxy yet. Otherwise, if current team isn't enough, they should hire more people. There is no other way about it. There can be a marginal case when they simply can't find anyone to hire. That happens, but I'd say they had quite a long time to find some people. I see many of their engineering positions require relocation to Poland offices. That may be a blocker for them.

There are no in depth statistics anyone can pull up to give a complete scientific analysis of operating system usage out there, so we have to take what is out there and do some reasonable ballpark estimation based on that. If you look at the number of systems on Steam Stats registered as Mac versus Linux/SteamOS it is one factor to take into account. Another is the number of games available on Steam for Mac versus those available for Linux. These are loose numbers but gaming deployment on Mac versus on Linux is higher, somewhere up to double on Mac or thereabouts. In theory if Linux deployment climbed to match Mac, that would make for a level playing field on those metrics. But there's more to it than that also, as they're officially supporting 2 Linux distributions, which means potentially double the installation and quality assurance testing or more depending on how far they take it. The amount of variance that can and generally does exist on a Linux installation is generally much higher than it will be on a Mac or Windows system, and that can't be ignored either.

I'm rather sure that your comment about Linux users being potential pirates is entirely in jest, but everyone may not realize that. For the record however, if anyone goes to the Humble Bundle website and looks at the price break downs for a given bundle, where it shows bundle purchases per platform (people can choose what platform they are purchasing their game for, for statistical purposes), in every single bundle I've ever viewed the statistics for, Linux contributors consistently have paid more on "beat the average" than both Windows and Linux, which I perceived as Linux users wanting to reward the game companies for being thoughtful enough to make their games available on the Linux platform by paying extra. I don't know of any statistics other than Humble for this, but I'd wager that the wider Linux gaming community would have similar sentiments and even be willing to spend a few bucks extra to get good Linux support (or any at all) if such an option exists (as it does with Humble).

So I don't think there is any problem with regards to Linux users being willing to pay, or even to pay more. I think the problem is that even though Linux gaming is at an awesome and growing awesome state, it still has not reached that "cruise control" speed yet and unfortunately for all of us Linux enthusiasts - it is still niche. It's upper tier niche, and hopefully will cross that threshold to being more mainstream WRT gaming in the next few years, but it's just not there quite yet.

I perhaps naively thought Valve's SteamOS and Steam Machines efforts were going to catapult Linux gaming forward much further than has actually materialized to date so far. I'm glad for what they did accomplish and it did move things forward but we now seem to have hit a new plateau. I'd like to fantasize that Galaxy client would push it forward to that next level, but I have doubts about that too. :)


Yeah, I completely agree with you on all points of that. It's pure conjecture, but I speculate that it is a very hard sell on a lot of developers around the world to relocate to Poland. If true, then that's a real bummer as they're missing out on perogies out the wazoo, and a bunch of other great things Poland has to offer of course. :) I wonder how big GOG would have to become in order to consider opening offices in other countries such as the US, UK, Germany or elsewhere which could increase the pool they have to work with?

They've had job postings up literally forever, which seem to stay up perpetually. That could mean that they're hiring people and still wanting to hire more, or it could mean they want to hire but are not filling the positions, or even a combination of both perhaps. Even if they are hiring people to work on Galaxy though they're going to most likely allocate them to the projects that are the most important to their business overall, and that might not end up being Linux support for a while.

No Drm != Linux support


It's no trivial matter supporting a completely different platform, even if the program is written in java, and especially considering that it is an installer sort of application. It's not a simple matter of compiling and releasing. The requirements for installing between Windows 7/8/10 and linux are quite different.

Now the runtime libraries required to run because gog galaxy is trying to be like steam, I would say that the easiest solution is to provide an empty shell of a library that really does not support any of the features, or simply mimics and does nothing. This would mean no multiplayer type support if galaxy provides that, but at least it would get some games running single player.

No DRM = no Windows which has built in DRM ;)

I proposed it to them, and they claimed it doens't help them.

Or they didn't tried. :-P