4 ms·
WebGL in Internet Explorer
- AshleysBrain 15y ago"I'm still confident that Microsoft will add WebGL support to IE" - hmm! Really?
- azakai 15y agoI think Microsoft will. Yes, Microsoft has opposed it so far. But WebGL is gaining traction and will soon be enabled in all web browsers but IE. Interesting websites using it are starting to appear, and that will only accelerate. So Microsoft will lose out by being the only browser not supporting a useful technology. People will prefer other browsers. Where Microsoft has nothing to lose by implementing WebGL.
- windsurfer 15y agoWon't it be WebDirectX then? DirectX is a direct competitor to OpenGL.
- azakai 15y agoBut WebDirectX would benefit Microsoft in no way. It would be a proprietary extension to the web. It would not let IE run WebGL websites. Microsoft has recently seen the light when it comes to web standards - the latest versions of IE are very standards-compliant. I agree that Microsoft hates OpenGL, and promotes DirectX in any way it can. But Microsoft also used to hate JavaScript and HTML, and tried to get people to develop in .NET on the web (Silverlight). But it has shifted to actually supporting the standards-based web. WebGL is part of that web.
- cdmckay 15y agoEven if they made WebDirectX, people would just write WebGL wrappers.
- AshleysBrain 15y agoI think if it comes to it, Microsoft will invent WebDirectX under the guise it's for Windows 8 metro apps, and simply left it on in IE10. 3D on the web is a niche so it doesn't make their browser uncompetitive. Besides, they can also claim "you can do the same thing in WebDirectX anyway, just use that" - and they don't have to be seen supporting OpenGL (who cares if anyone writes wrappers? they don't have to pretend WebGL is a good idea). For the record, it's an utterly backward and damaging thing to do to the web, but they'll do it anyway.
- extension 15y agoa work-in-progress that implements WebGL with Canvas That's a neat hack but it is absolutely not going to work in any practical sense. There is already a many orders of magnitude performance gap, and depth buffering will at least add yet another order of magnitude. Also, canvas only supports affine texture mapping, as opposed to perspective-correct 3D mapping. This looks completely ridiculous for many common cases: http://en.wikipedia.org/wiki/File:Perspective_correct_texture_mapping.jpg http://en.wikipedia.org/wiki/File:Perspective_correct_textur... And, canvas can't cleanly stitch textured polygons together, so there will be visible seams everywhere. I don't mean to be a killjoy, but I think one would find this endeavor very discouraging in the end.