23% for Linux right there, in the public poll on very popular game, posts from active members with verifications.


Yes, we can and its accurate. One quarter of user base. Its not "hopes", its hard facts.

It is bad for Linux gaming, for sure. Whether it's good or bad for GOG - no one knows. We can only compare what GOG are doing to what other distributors are doing, and conclude that others are doing more, while having less (itch.io).

as much as I love the itch.io client (especially that they are using firejail to sandbox their applications), you can't really compare it to Galaxy.
most importantly itch.io does not provide multiplayer / match making / achievements and all that modern fluff.
And without that I would argue that we would still miss these recent Linux releases.

ironically, from my POV the itch.io client is a gaming client "done right" - it concentrates first and foremost on getting the game to the player without trying to imitate steam and adding loads of extra features into it.
But sadly this is not what publishers (and most player evidently) want to see.

As it currently stands right now, GOG's way of treating Linux users is more bad than good. The most recent releases being Pillars of the Earth and Darkwood both came with their Linux version since day 1 which is a good thing.

When it comes to games with multiplayer though, you can expect the worst even before a game is released and as has already been proven: Sudden Strike 4 and the upcoming Tooth and Tail, both victims because of lack of Galaxy.

Another thing to mention, while the games on GOG are technically DRM-Free, Galaxy doesn't truly seem as optional as they want to have us believe. Linux games on Steam may have DRM, but with GOG it's even worse since the games aren't even available at all. I wonder if they even considered this, and how could they even have the face to say Linux isn't even a priority.

Honestly I'm very, VERY disappointed in GOG, they could have delayed the client ok, but at the very least they could have provided devs with the necessary API's so Linux games would be at least playable in single player (and here we make so much noise about Always-Online games on Steam, how ironic). How long has it been now, almost 2 years since Galaxy was released on Windows? More or less around the Witcher 3's release.

And after 2 years, they tell us that it's coming and (as quoted by their recent comment): "I'm not saying we're not able to, or we don't know how". Now, does GOG truly want to have me believe that? 2 years of smoke and no fire, not a single screenshot that the client even exists, all we ever got were words that hold no weight.

I think GOG should be a bit more considerate towards their Linux user base, it's small I understand that but why are we getting this treatment?

Yep, I think GOG bit off more than they could chew in this case, and got over ambitious, in result stretching their resources thin. Linux support became the victim of that planning. They should have aimed at something like itch.io first, and only when reaching that should have started working on some huge goals like multiplayer and etc. which are magnitude more complex.

That's just it, I don't think GOG is even remotely trying to "create a Linux market". Rather, they added Linux offerings to experimentally tap into whatever Linux market may or may not exist already. That they're not putting any major focus on Linux development right now is indicative that they perceive the Linux market to be too small at the moment, and that their resources are better spent elsewhere.

If GOG were trying to create a Linux Market, they'd probably be doing their own custom Linux distribution - GalaxyOS or whatever like SteamOS, and possibly Galaxy Machines as well. :) They're clearly not doing that however.

This seems to suggest that there are more people out there gaming on Linux than there are people who want in-game overlay, cloud storage, and the other various features Galaxy provides. Supporting more operating systems isn't really objectively "doing more". In this case it appears to be supporting far far less game client features, on a single additional operating system. Galaxy at this state in time is supporting much more features on one less operating system. In terms of the global gaming marketplace wants and needs, I'm of the opinion that most gamers out there would prefer Galaxy and the features it provides than people who just want a game launcher/patcher and Linux support.

There are different ways to measure what "doing more" means other than "total number of operating systems supported by product". One can slant the message based on cognitive bias however they like, but it doesn't really hold up under objective scrutiny.

I think you nailed it here. Had they aimed at cross platform support right from the beginning as an actual core product goal, and limited the initial features strictly to storefront, game library management and launching and similar simple features, they'd have had a cross platform baseline client first. The more popular modern client features could be developed incrementally after that also cross platform, and that is an approach I'd have like to have seen them taken.

I think they ended up allocating a lot of developer resources to the higher end features a bit too early under the probably correct premise that the mainstream gamer market expects such features, but underestimating how much time it would take to actually develop it all. Furthermore, their first attempt at developing the overlay didn't really pan out, and they ended up redoing it as an injected overlay, which is what all other overlays out there do. It's not clear if that was a necessary development path to get from point A to B, or if they wasted time going in one direction then later realized they needed to change gears and do it a different way. Either way it took over a year to come to fruition.

It's too late now to change things, but we probably can agree that had they kept it 100% cross platform from day 1, and focused on providing the identical baseline client across all 3 platforms, the Linux user base would probably be pretty happy overall with a basic client even if there was lag with getting certain new feature updates that involve platform specific work. That probably would have allowed developers to actually make GOG versions of their games as they wish, even if the high end client features weren't all available yet.

While hindsight is 20/20, it's all past history that can't be changed. Realistically, I think we'll have to wait until Cyberpunk 2077 comes out and GOG & CDP get another massive cash infusion and can ramp up their hiring efforts and have much larger teams that can be working in parallel on things. Right now it doesn't seem like the Galaxy team is large enough to have much parallel work being done simultaneously. With a larger team working on multiple sub-features, it would be more realistic for them to designate a developer or small team to Linux platform specific feature development and issues however.

Unless they share some major revelations with us, I estimate we'll be waiting another 2-3 years or more for a Linux Galaxy client if we ever get one, and I think that people will have reverse engineered enough of Galaxy in the mean time to have a working downloader/installer/launcher by then. There still wont be a Galaxy runtime library for Linux however, so a community driven effort based on reverse engineering wont end up increasing the unit count of Linux game releases on the platform. That'll have to wait until they have a proper client. Like many have said though, a lot of existing games that do have Linux releases elsewhere are likely to never have GOG Linux releases even once GOG does get Galaxy on Linux up to scratch.

:o(

That's not how Statistics works. What that poll tells is that 23% of the people who have an account on that forum and chose to answer the poll (29 people) use Linux. You can't extrapolate the results to the whole userbase with that methodology.

Why is that? Its still open. Your claim makes every poll incorrect.

When you perform a poll (or any kind of field work), you need to design a methodology that ensures that the data you are obtaining is an accurate depiction of the population you are studying. Or as accurate as possible.

One of the most critical points is the size sample (number of people you are interviewing). Ideally, you would interview 100% of the population and your results would be very accurate, but that's unfeasable unless you have unlimited resources (time, money and manpower), and not very practical. There are mathematical equations that, with a few assumptions, allow us to determine what is the optimal size sample that is needed to obtain results with enough certainty (researchers normally settle for 95%).

You also need to decide how the data will be collected. This process, if not done right, may introduce biases that will falsify your data. An obvious source of bias is how the participants in the poll are selected. For example, if the interviewer selects them on their own, they may not choose a heterogeneous sample of the population, while if you only interview people who volunteer, you will end up with answers only from people particularly interested in the subject. That's why a lot of polls are done via random telephone calls. Other biases may be introduced in the way the questions are formulated, which could skew the answers towards a certain option because the interviewees don't want to look bad (even if the poll is anonymous).

And even with all this work, polls can and do fail (think, for example, how most polls failed to predict that Brexit would win).

Internet polls are fun, but they don't have much scientific value (unless they are used in very small communities where almost everybody answers; for example, a group of friends or colleagues voting when and where to celebrate a dinner).

Well not to long ago peoples were trying to convince Gog to release games on Linux by telling them that the Linux user base was not like the Windows one and that the only thing they needed to do was to release the games binaries and that the community would handle the rest... :)


You have to see the other side : releasing games on Linux with minimum investments also means minimum risk for them, which in turn can make those release worthwhile (risks and ROI wise).

If the only way for them to release games on Linux was first to invest X millions on it then probably they would have considered that it was not worth it and not release anything at all.


Personally I am pretty sure that Valve investment in Linux has everything to do with Gabe acute paranoia attack and very little to do with any market research suggesting any hope of profitability of said investments.

Yeah but investing in something less "unknown" with lower risk and better ROI is often the preferred third option. (what Gog did with movies, not really a resounding success, and Gwent)

The problem is they are hurting Linux releases here, and in result hurting their own Linux user base. I.e. I don't care about Galaxy really, but I seriously care about releases being blocked by it on GOG. In this sense, Galaxy is a major anti-feature. It might be unintentional, but so far GOG didn't manage to avoid this problem.

Look, I appreciate (really) the time you took to type the theoretical part of statistics, but all that is theory and outside of practice.

95% sampling is unrealistic as it will include inactive members. The local community is pretty active on forums to notice it, there is no reason why one would try to falsify, because the votes are public and duplicate membership prohibited. I also don't get it about bias - many posted screenshots of system info or how could primitive multiple choice like "win,mac,lin,other" be inaccurate? You can't make a more unbiased poll than simply publicly ask for preference in the most active place with most simple wording.

95% certainty doesn't mean a sample size of 95% -it's actually a lot smaller. That's why we calculate the optimal size sample, to reduce the effort needed.

I don't doubt that the 29 people that answered the poll in that forum use Linux. But you can't infer that other thousands of people also do. That data is only representative of the people that visit that Internet forum, and it is skewed towards people interested in an OS poll. It is not representative of the whole base of Kerbal Space Player programs, let alone gamers worldwide.