hd_ wrote:The first questions should be: Who is willing to work on what?
There are many talented and enthusiastic contributors to the Urho project, I think it might be possible to gather a small group (2 or 3 people) to at least attempt one feature request together.
cadaver wrote:Aiming for a modern rendering architecture while keeping D3D9 level compatibility sounds contradictory, or at least something that would increase the difficulty. However I'd agree that given current compatibility and platforms that Urho can run on this is necessary (not D3D9 in itself, but mobile GLES2.0), because otherwise too much currently working platforms would be ruled out.
cadaver wrote:Vulkan and D3D12 are not at all certain to bring performance improvement. Instead you have to do low-level management of resources that the driver used to do for you, and it'd be rather easy to shoot yourself in the foot, on multiple levels from outright crashes to just inefficiency.
cadaver wrote:Having one shader language only is a well-defined goal and would benefit all use cases of Urho. Though one would still need to be careful of the implementation. For example, if this was done by runtime-translating HLSL into GLSL, then mobiles could suffer in longer startup time. An offline process would be a better approach, but remember that Urho does not have a "project asset build/preparation step" at least at the moment, but just expects to load ready-to-go resources (like models, dds textures and API's native shader code). For practical purposes I'd personally be fine with a manual approach like "when you have changed shaders, run this tool or batch file to translate into API-native shaders"
franck22000 wrote:DX11 features for examples like geometry shaders, compute shaders
yushli wrote:I don't think overhauling Urho3D 3D part is a good idea. Don't fix what isn't broken at the first place.
Enhex wrote: Multi-threaded rendering
- Isn't it the same as Vulkan/DX12?
Enhex wrote: Auto generated script bindings
- Not sure if it's worth it. Valve had a paper about how they made a very complicated special tool that can generate bindings (doesn't work for all cases) and they spent a year making it instead of few days doing the bindings, and now you need specialized programmers to be able to even maintain it.
So you got partial solution that you'll have very hard time to be able to find replacement developers for and spent orders of magnitude more time making than just writing the bindings.
In general against, unless it can be kept almost as simple is writing the bindings.
Enhex wrote: Material Prototypes
- Don't know what that means.
Enhex wrote: Support for Code Plug-ins
- Not sure what you mean by plug-ins, but you can already use Urho as a C++ library and add whatever you want. And there's scripting.
Enhex wrote: Hand written documentation
- Already exists
Enhex wrote: Updated UI textures
- What updating means? Better default ones?
dragonCASTjosh wrote:My plan for this was to compile the shader to its target format upon saving the shader file.
Users browsing this forum: No registered users and 1 guest