| Horde3D http://horde3d.org/forums/ |
|
| FCollada or Collada DOM in converter http://horde3d.org/forums/viewtopic.php?f=2&t=259 |
Page 1 of 1 |
| Author: | bbilz [ 20.03.2008, 19:17 ] |
| Post subject: | FCollada or Collada DOM in converter |
Hi, I have been doing a bit of work with the ColldaConverter, mainly expanding on the physics integration; supporting constraints etc.. I was just wondering if there was ever a plan to implement a more standard method of reading in collada files such as FCollada or Collada DOM instead of the current direct XML parsing? This would make extending the converter a lot easier as you are guaranteed to have access to everything in the Collada file, and that it will have been interpreted correctly according to the standards instead of having to figure them out for yourself.. It would be incredibly useful for further development! -bbilz |
|
| Author: | marciano [ 21.03.2008, 21:37 ] |
| Post subject: | |
I had a look at the standard COLLADA parsers a longer time ago but was discouraged by their huge size. I don't think that they would bring such a great advantage. They map the XML to memory structures but you still need to handle and interpret these structures correctly. So there are also many chances to miss the standard or leave out a feature. |
|
| Author: | swiftcoder [ 22.03.2008, 13:33 ] |
| Post subject: | |
marciano wrote: I don't think that they would bring such a great advantage. They map the XML to memory structures but you still need to handle and interpret these structures correctly. So there are also many chances to miss the standard or leave out a feature.
The problem is that we currently have a very pedantic parser. If we could make it a little more lenient, then it would handle output from the FBXConverter, Blender, etc. In particular, it only can process <triangles> tags, and we should be able to force it to assume that <polygons> contains only triangles. We should also process and combination of <f> tags, not just the ones containing 3 indices (FBXConvertex dumps all the indices in one <f> tag). |
|
| Author: | bbilz [ 25.03.2008, 16:27 ] |
| Post subject: | |
Indeed, my point is that implementing a standard Collada parser now would save a lot of this kind of work later. In order to support all the various situations that you may require from a Collada file, you may eventually end up with your ColladaConverter as large as the original standard parsers anyway! I understand that as a graphics engine, you only require a small subset of the information in a Collada file anyway - I am more talking about the developers who need the other bits of the data (Physics, ColladaFX) who are having to add large chunks of parsing code to the ColladaConverter. Having a standard Collada interface would make this task a lot quicker and easier in my opinion. |
|
| Page 1 of 1 | All times are UTC + 1 hour |
| Powered by phpBB® Forum Software © phpBB Group https://www.phpbb.com/ |
|