Rough Roadmap?

[Bertram]

Hi there,

I know this is a very common and simple question with not forcefully a simple answer.
Still, as I saw no clear roadmap in the bugtracker, and wasn't able to find anything more than mere clues here and there about what is coming in the next version, I wondered:

What are the features and blocker fixes wanted or left before the next release?
Note that I didn't say when, I'm simply asking what, even roughly.

I also wanted to know whether new textures was something planned at all in the game (while I know the renderer might cause problems there.) and/or whether this effort will be done gradually.

Best regards,

[NoQ]

Model and texture data files are something easiest to change. It doesn't have any roadmap, it can be done in any phase of development, cause updating them usually does not produce any bugs.

Minor releases just carry some of the easiest and/or most important bug fixes. They just get accumulated in the respective git branch, and a new minor version is released from time to time.

Biggest feature changes happen in git master, and usually there's a rough idea of what big features will be there. For example, in 3.2 we'd probably be having stats data converted to ini format (almost done), and maybe also rewrite campaign in js (only started a little). There were also some renderer updates being done, but i'm not quite understanding them.

[Bertram]

Hi NoQ,

Thanks a lot for the reply!

in 3.2 we'd probably be having stats data converted to ini format (almost done)

I saw two big patches from Duha in github just today. The new format is quite clean
and I hope the rebalancing will get through, while I know those kind of things are kinda sensitive
as it always frustrates someone somewhere for some reasons.
It's already good to see it's being done.

and maybe also rewrite campaign in js (only started a little).

I bet this is the best thing btw, congrats for switching to some modern scripting API.
I must say that porting the current campaign to the new format will be impressive work, while I know it's not the easiest path as it has its boring side of bringing nothing actually new. I hope this one and other campaigns using the new API will be done, though, or even multiplayer's maps with events?

There were also some renderer updates being done, but I'm not quite understanding them. [...]
Model and texture data files are something easiest to change. It doesn't have any roadmap, it can be done in any phase of development, cause updating them usually does not produce any bugs.

Nice to hear. This means one could rather easily upgrade the units look for instance?
I've seen many graphics improvements threads and I wondered then: Is there a will to replace the graphics, I mean in master?

Best regards,

[NoQ]

I must say that porting the current campaign to the new format will be impressive work, while I know it's not the easiest path as it has its boring side of bringing nothing actually new

There are a lot of long-standing bugs in campaign scripts (mostly inherited from the original game, i guess), hard to reproduce and debug. With the new scripting engine it should be much easier to write things clearly and avoid such problems; fixing them is quite an inspiring profit.

I've seen many graphics improvements threads and I wondered then: Is there a will to replace the graphics, I mean in master?

Certainly. The thing about Art Revolution guys (eg. Jorzi) is that they expect some features from the engine to support their models, so they're not putting their stuff into the main git yet. Also their models might be too high-poly for massive battles; not sure, i'm the wrong guy to ask. Those new textures by cnceo, i guess/hope, will be merged eventually. Berg throws in stuff regularly, but often more useful for custom mods than for base game, yet quite a lot of his artwork gets into the base game. Generally, any new stuff that looks similar to old stuff just better, gets accepted quickly.
____________________

Yeah, and one more thing we have in mind but hardly started is a better map format. The old tile-based format does not allow using the new terrain renderer to its fullest; we also need a support for maps with custom scripts and probably custom graphics (especially custom terrain and tilesets), that would be safer than current map-mod mechanism; challenges, tileset.ini and map.ini files are steps in this direction.

[Per]

Also their models might be too high-poly for massive battles

Not any more

Is engine limitations still holding back AR? In which case, it might be nice to have a prioritized list of things that needs to/should be fixed.

[Alpha93]

Is WZM format going to be supported starting with 3.2?

[Per]

Is WZM format going to be supported starting with 3.2?

Well, it could be. I had the code for it almost ready. But what features of WZM do you want, that has not been already added to PIE (in master)? In the long run, I wonder if it would not make more sense to start looking into something like Collada instead.

[Bertram]

Hi all

There are a lot of long-standing bugs in campaign scripts (mostly inherited from the original game, i guess), hard to reproduce and debug. With the new scripting engine it should be much easier to write things clearly and avoid such problems; fixing them is quite an inspiring profit.

Indeed, this will make the scripting API ironed out, if you can make the official campaign with it.
And I just saw Aubergine's effort on documentation about it.

Is engine limitations still holding back AR?

+1 on this one. I don't know about all the side-efforts made on this, but if we can see things from Jorzi, cnceo and Berg put in, I bet this will get visually interesting right away.

I wonder if it would not make more sense to start looking into something like Collada instead.

I guess WZM is the old, PIE the current, and Collada one of the possible next model formats (.dae)?
Why is collada better? Not criticizing, I'm just wondering why you want to change. There must be something attracting in this format compared to the current, right?
(For the record, I used this to get thing: http://en.wikipedia.org/wiki/List_of_fi ... D_graphics)

Ah, I also read somewhere that the map format was restictive in the freedom of tilesets used and compared to the renderer capabilities. Is that so? And is that somehow preventing textures to land in master?

Anyway, thanks a lot for all the answers, it's not easy to follow all that is happening.

Best regards,

[Alpha93]

PIE is the old format, WZM is being worked upon and I have no idea of what COLLADA is. Just started 3D modeling so I'm kinda new to this stuff.
Do you mind telling me what additional stuff WZM would be supporting compared to PIE, again? Didn't follow the discussion in Mangust's thread at the time.

[Jorzi]

Last time I checked, the main things holding the wzm format back were the lack of animation support & other special features of certain .pie files, like:
-scav dudes
-oil derrick
-cyborgs
-
The main advantage of the wzm format is the inclusion of precalculated normals, tangents & bitangents, allowing tangent space normal maps to be correctly displayed.

(Another feasture that would take animation to the next level would be the addition of bones and per-vertex bone weights. (oil derrick uses a simple form of bone animation where each part is its own bone)
A "bone" is essentially an affine transform matrix, and having weights would allow things like deforming realistic cyborg animation. Before going to bone deformation, however, I think we could benefite quite a bit from having a moddable general animation system.)

Edit: about performance:
Since we are now no longer using immediate-mode rendering (if I understood correctly) polycounts aren't really that much of a limiting factor anymore, except when it comes to shadow casting.
Since shadow casting is cpu based and uses shadow volumes, the calculation time is directly proportional to the vertex count. With shadow casting off, however, you can essentially throw millions of polygons at the screen with even a fairly weak gpu.

[Bertram]

Ok,

PIE is the old (and current format, I guess), WZM format is the next one. If I get it, it's a custom format?

I had the code for it almost ready.

Cool to hear.

the main things holding the wzm format back were the lack of animation support & other special features

What are those special features? As you only listed unit names, I didn't get it, in fact .
Are you talking about "bones" support?

(oil derrick uses a simple form of bone animation where each part is its own bone)
[...] A "bone" is essentially an affine transform matrix, and having weights would allow things like deforming realistic cyborg animation.

But what features of WZM do you want, that has not been already added to PIE (in master)?

So, the WZM format would lack animation and bones support, as a good start?

Speaking of Collada or an open format for 3D assets. Is there one format being open and having the wanted features?

Before going to bone deformation, however, I think we could benefite quite a bit from having a moddable general animation system.)

How are the animations defined, btw?

Thanks for all the precise answers. It's great to have it all in one thread.

Best regards,

[cnceo]

Speaking of Collada or an open format for 3D assets. Is there one format being open and having the wanted features?

Collada on the one had supports animation and on the other is supported by quite some 3D modelling programms out of the box and it is open. So I guess it would be a huge benefit for the artists. However, developers are pretty busy with bug fixing, netcode, gameplay etc. Also I think there are a lot of things that can be done about graphics, without waiting for a new format. It is true that there is no real roadmap for the graphics, which is partialy due to the lack of a semi-complete asset of new structures/textures, which maintain the original style (See the art revolution mod thread for one of the approaches). So in principle you could take a look at those mods and start with anything that is not done yet. Once you have done your model just release the raw material and everyone can export it to any format that you possibly want in the future.

[Bertram]

Hi there

I just realized 0ad is using Collada for their models, btw.

So in principle you could take a look at those mods and start with anything that is not done yet.

Would an incremental effort be accepted in master?

For instance, thanks to Jorzi's report:

Among the alpha campaign weapons only the lancer and truck are still unfinished (MaNGusT has WIP models)

This means the other models can be imported, I guess? Would they fit in PIE format, or should at least the WZM format be added first?

Best regards,

[d00fd00f]

Any news on what the required/preferred new features in the rendering code should be?

As far as the rendering code goes.. what is "missing" to bring in AR content? or has all of this been fixed already?

[Per]

The big things remaining in the rendering code is converting the remaining holdouts using immediate mode (I have work in progress at https://github.com/perim/warzone2100/tree/gfxqueue), changing the terrain renderer (a tough one due to the way the data is currently structured), and rewriting the shadow code to use more modern, faster techniques (requires changes to the terrain renderer, which is probably the hardest part).