Star Citizen Plans To Go Vulkan-Only

[WonkoTheSaneUK]

Also, they don't actually target DX12 or Vulkan, they use an engine that can support those. First they "wisely" chose CryEngine and when Crytek started going downhill they move to Lumberyard. Lumberyard may support Vulkan, but the engine itself is not open (dev will get the source code, but that's it) and also doubles as a way to herd people onto AWS.

In short, Vulkan is preferable to DX12, but the only thing that would pique my interest, is a finished Star Citizen/Squadron 42.

Actually, Lumberyard is a fork of CryEngine, based on the same build that CIG forked from. To switch over took 2 of CIG's engineers a single day to complete.

[pete910]

I love Vulkan but what about finishing the damn game first?

http://dereksmart.com/2016/06/star-citizen-fidelity-of-failure/

*snip

WTH, your takin that guy seriously !

Have you seen his effort at a game ?

[pete910]

Actually, Lumberyard is a fork of CryEngine, based on the same build that CIG forked from. To switch over took 2 of CIG's engineers a single day to complete.

In other words, they took the modifications that Amazon had done and added it to the existing code branch

[Shevchen]

I know the story. But after so many years in development, how many features are actually finished? Is space combat done? Trading? Weapon systems? Something? Because last I heard about it, someone tried to get into the open beta week-end, only to rage-quit after a few hours playing a bug-ridden mess.

I don't mind the waiting (I haven't chipped in anyway). But to demonstrate progress you have to hit some milestones every now and then. As always, I prefer to be sceptical and let RSI surprise me, rather than the other way around.

Can't relate to that. I play it regulary with my crappy Radeon HD 7970 and it runs good. Arena Commander/Pirate Swarm is a fun game-mode against the AI and Squadron battles are also fun for the short-term vacation. In the PU, the FPS is rather limited due to the server bottleneck, but thats why they bring in a new "netcode".

I'm watching the updates they bring out and the pace is good. I just think many people underestimate the scope and size of Star Citizen and try to understand it as GTA-online or similar stuff. Its not and should not be treated this way. But I can understand, that its hard to imagine something, that has never been done before.

From what I compile from all the reports, SQ42 will be released this year and all hands are on deck to make it happen. The PU 1.0 may hit release some time next year and for all the nuggets we get in between, I'm already satisfied with what already is playable. If SQ42 won't be released this year, I would be disappointed too, but I'm chilling around here too.

Some people want to see it released as soon as possible and there are others, who want to see it "someday" as long as its good. Sadly, the project is so massive, that you can't make both sides happy. My personal opinion on this is that I don't care how long it takes, as I'm not in any rush... why should I? Is the world ending in 3 years and I want to play SC at least one until then or why the pressure? Or is it, because we get one Assassins Creed each year and are not capable to wait for a game anymore? Or is it, that we are used to "announcement of a game till release" within half a year, while the game itself too 7 years in Dev already? (Fallout 4)

They could make it 10 years and I wouldn't even be mad about it. But thats just me.

[bug77]

Can't relate to that. I play it regulary with my crappy Radeon HD 7970 and it runs good. Arena Commander/Pirate Swarm is a fun game-mode against the AI and Squadron battles are also fun for the short-term vacation. In the PU, the FPS is rather limited due to the server bottleneck, but thats why they bring in a new "netcode".

I'm watching the updates they bring out and the pace is good. I just think many people underestimate the scope and size of Star Citizen and try to understand it as GTA-online or similar stuff. Its not and should not be treated this way. But I can understand, that its hard to imagine something, that has never been done before.

From what I compile from all the reports, SQ42 will be released this year and all hands are on deck to make it happen. The PU 1.0 may hit release some time next year and for all the nuggets we get in between, I'm already satisfied with what already is playable. If SQ42 won't be released this year, I would be disappointed too, but I'm chilling around here too.

Some people want to see it released as soon as possible and there are others, who want to see it "someday" as long as its good. Sadly, the project is so massive, that you can't make both sides happy. My personal opinion on this is that I don't care how long it takes, as I'm not in any rush... why should I? Is the world ending in 3 years and I want to play SC at least one until then or why the pressure? Or is it, because we get one Assassins Creed each year and are not capable to wait for a game anymore? Or is it, that we are used to "announcement of a game till release" within half a year, while the game itself too 7 years in Dev already? (Fallout 4)

They could make it 10 years and I wouldn't even be mad about it. But thats just me.

Right. Another wall of text without mentioning one finished aspect of the game.
You may be ok with that (and I really have no problem with it), but as a software engineer, I can tell I've never worked on a project where after several years of development we didn't have at least some subsystems done. I'm not saying it's not possible that this game has to do just that, I'm saying my past experience raises some red flags.

[ArchFanatic]

Here is addition for folks with lack of any critical thinking who think that this game will be finished in original dream shape nearly impossible with current or modified Cry tech...

https://docs.google.com/spreadsheets/d/1e9rQr2_iR0XPrvYrCNoFk4nNq_5VrrJFXf2G02MmVnU/edit#gid=1998968794

[Shevchen]

Right. Another wall of text without mentioning one finished aspect of the game.
You may be ok with that (and I really have no problem with it), but as a software engineer, I can tell I've never worked on a project where after several years of development we didn't have at least some subsystems done. I'm not saying it's not possible that this game has to do just that, I'm saying my past experience raises some red flags.

The first gameplay-mechanics as far as professions outside of simple combat come online with 3.0 to 3.x where 3.0 is expected not to land within the next 2-3 months and can as well take time till fall. The problem with the "finished aspect" of the game is, that all the functions are inter-veined. AI affects economy, economy affects mission building, missions are dependent on AI again and so forth. But okay, I don't want to butter anything up for you, a healthy skepticism is good and has had some constructive effects on several occasions already. I just want to suggest, that *this* software development is different from the "standard" one, which also invokes the need to not judge it like a "standard" project.

3.0 will give you planetary landing and a basic economy system plus basic AI. While still far away from being "finished" (It may never be, as its an ongoing thing that will be refined even years from now), I'd say 3.0 will be the first usable beta-like version of the PU.

[pal666]

Uh, no - the version CIG has is heavily modified

and now it is clear that it was a mistake

[pal666]

At this point, the StarEngine has only remnants of code left from the original CryEngine and I highly doubt, that the generalized Vulkan-wrapper Crytech wrote is even remotely capable of rendering the code SC has by now.

at this point they can learn some basics of software engineering like don't fork irresponsibly

[tajjada]

so the correct solution is to upgrade their version

... except that their version is forked and modified to the point where merging their work into a newer upstream (or upstream's work into their fork) would likely be very difficult. If it was easy, am pretty sure they would have already done it.

It's not like they are using unmodified upstream cryengine and can just download the newer version and call it a day.