It might be useful in the future as a way to support legacy games on new hardware without D3D9 support (not going to happen soon, but one day…, and this project will free driver devs from it to focus on just Vulkan), and as a way to port DX9 games for other platforms (think Android). At least that’s what comes to my mind.

Also, Linux users with Vulkan-compatible hardware will be able to use it with Wine instead of its translation layer into OpenGL (which, I believe, might be beneficial: translation layer from a high-level API into low-level one should be cleaner and faster than from a high-level API into totally different high-level one).

View this comment - View article - View full comments

Quoting: slaapliedjeThey kind of get the cast offs after the studio who made the original game decided to work on other projects. Some of their games are epic, some are not so much. Though I believe they did make the original for Pillars of Eternity, and I think Tyranny is the somewhat sequel to that?

Well, no. Tyranny is a brand new, Obsidian-developed, franchise, another setting with a very different atmosphere and characters development, another gameplay, although on the same engine and similar style of isometric RPG in the technical way. But that’s all.

I believe they are making an actual sequel to Pillars of Eternity, in PoE universe. Perhaps they also will make another RPG in Tyranny’s world.

Also, I believe PoE is heavily based on all the original Infinity Engine games, especially the setting style of Baldur’s Gate, storytelling of Planescape: Torment and combat mechanics of Icewind Dale (but shamefully didn’t play them, except for just a few hours of Planescape).

View this comment - View article - View full comments

Quoting: gurvI think we're witnessing the whole DirectX lock-in all over again. But this time it might very well be the shader language that stops developers from migrating to Vulkan (plus Microsoft offering "incentives" to go DX12)

Well, Vulkan uses intermediate language (pretty low-level assemblery, but still platform-agnostic, one) for shaders, SPIR-V, so its adoption depends on compilers from higher-level ones, like GLSL and HLSL. Vulkan came out with a glslang compiler doing the GLSL → SPIR-V compilation and it is getting HLSL support right now, so let’s hope it won’t be an issue anymore.

On the other hand, glslang does not do optimizations, so developers have to do things like loop unrolling by hand in generated SPIR-V or search for third party optimizers (which take SPIR-V representation and output optimized GLSL, so one needs to write HLSL, compile to SPIR-V, optimize to get GLSL, compile again to SPIR-V (and after that perhaps run official SPIR-V optimizer from Khronos which does just very basic optimizations at the moment). And that is pretty cumbersome, so probably not too encouraging for the switch to Vulkan.

But let’s hope it is going to change for the better in 2017, as all those tools are open-source, SPIR-V and Vulkan are pretty good technologies, and they slowly gain popularity every day. Also remember: Vulkan is the future for Android, and huge amount of gamedev is done on Android. ;-)

View this comment - View article - View full comments

I’ve found a post in Czech in their forums from one of the moderators, back from March 6th. Even though on the international threads they keep saying that they keep Linux version in mind but cannot promise anything, the post in a thread about Linux Beta says:

QuoteTak se k vám vracím s nemilou zprávou.
Linuxová verze není a jen tak asi nebude. Momentálně je to prostě technicky nemožné.

which roughly translates to:

QuoteSo I’m returning to you with an unpleasant update.
There is no Linux version and there won’t be any. At the moment it is just technically impossible.

I wonder if that’s just about beta (and what are the technical – and apparently not just manpower availability – problems) – and if beta is impossible, then how can a final game be possible on Linux.

Did they use some Windows-only middleware and are waiting for a Linux and mac versions? Did they write hard to port HLSL shaders for D3D? Are the Linux drivers and OpenGL support not feasible for the work the game requires to run decently even on high-end hardware? If so, would Vulkan help (as new version of CryEngine supports it)?

EDIT: And after a little more digging I’ve found out that they use CryEngine 3 which does not support Vulkan (nor will) – and they apparently have performance issues, so I believe not doing the Linux/macOS version is related to a difficuty of doing complex optimized OpenGL rendering (but that’s only my guess).

View this comment - View article - View full comments

Putting cross-platformness away does not seem to me as really going for quality – they might end up with platform-specific and platform-optimized code that runs OK, but actually isn’t high-quality maintainable and portable source code.

Of course it is possible that they have well structured abstractions that make code mostly platform agnostic and just do not want to spend energy now on implementing the platform specific bits on Linux and macOS. But I don’t really believe that – if they don’t try to support at least two different platforms since the beginning they cannot be sure their approach is going to make easily portable code later.

EDIT: they ARE supporting consoles, so there is hope that they are making it easily portable and not too Windows-dependent. So I might be (and hopefully am) wrong in my comment. :)

View this comment - View article - View full comments

And… It’s greenlit now. :)

View this comment - View article - View full comments

Quoting: KimyrielleThis is one of the very, VERY few commercial quality open source games (I can't even think of a second right now, but there probably is one). Most look as if a 5 year old did the art (*cough* FreeCiv...).

Battle for Wesnoth (very nice 2D IMO), SuperTuxKart, FlightGear… There are some with somehow decent graphics, but looking old, like MegaGlest. Anyway, there is quite a few open-source games that are good quality (better than some commercial titles), and there’s even more free-as-in-beer with FOSS engine and not-open art (eg. AssaultCube).

But I agree, most of the FOSS games do look like games from decades ago, or as if they were done by people without any graphics experience.

That’s understandable, as some of them are projects started some time ago, and even though are still under some development, didn’t age well (UFO:AI?, Warzone 2100, aforementioned MegaGlest) – I worry that by the time 0AD leaves alpha stage into beta stage, it’s gonna be visibly aged too.

View this comment - View article - View full comments