Re: 3D -- WebGL and SVG?

Hi David, Dean,

I agree with Dean that a markup language is by itself difficult to design, and even more so, if not impossible, for rich 3D content. If it's about real world use cases for 2.5d content, I'm happy to review any proposal. The same goes for mixed formats (2D with 3D avatars). We are building these games, and are happy to review and move forward any spec that helps. WebGL is very versatile, but definitely not something web developers would use.

The order in which I would attack things is the following:

 1.  Design a higher level JS API to work with immediate mode WebGL
 2.  Design generic layer to handle collisions (especially on mouse events) on multiple tech formats – I.e. WebGL and HTML + CSS transforms mashed up
 3.  Design a very high level retained mode 3D API (something like O3D, with a scene graph etc.)

Dean is right that often the right way is to let library developers jump on it first, and distill the important parts for an actual browser-level implementation (I've been a lib dev myself). There's a fundamental flaw in this strategy however when it comes to WebGL – even actually great library web developers don't understand 3D shader languages. The tricky bit is to design an API that actually works for web developers moving into 3D and game development. As much as people sometimes bash jQuery, it enabled the same transition: web designers suddenly became web developers.

Paul

Von: Dean Jackson <dino@apple.com<mailto:dino@apple.com>>
Datum: Sun, 7 Aug 2011 12:37:21 -0700
An: David Dailey <ddailey@zoominternet.net<mailto:ddailey@zoominternet.net>>
Cc: 'www-svg' <www-svg@w3.org<mailto:www-svg@w3.org>>, "public-xg-declarative3d@w3.org<mailto:public-xg-declarative3d@w3.org>" <public-xg-declarative3d@w3.org<mailto:public-xg-declarative3d@w3.org>>
Betreff: Re: 3D -- WebGL and SVG?

Hi David,

On 07/08/2011, at 6:23 AM, David Dailey wrote:

So, maybe, as you say, the domain of SVG is limited to making sure it can be used with 3D technology, but it is possible that the similarities, and markup *could* share a lot more in common. SVG has done a lot of things very very well. It is conceivable that slight extensions to the conceptual space outlined by SVG + SMIL + replicate could motivate just the sort of declarative language that the incubator group seems to be seeking. After all, most of the mathematics of higher dimensional space is written in notation that appears in 2D <humor />.

I think this is where I am most reluctant - maybe I misunderstand your suggestion. I don't think there are slight extensions to SVG that would provide a useful 3d format. Even if there are, I think that effort would distract from work necessary in SVG.

There definitely is a community for declarative 3d markup. As I said in my first message, the problem is developing something that is useful outside a niche, and making sure the right people are involved.

On a tangential note, irrelevant to the basic discussion, if by the one truly successful markup language you mean HTML, I think its success was less due to its careful crafting than to the remarkable void that existed pulling *any* suitable declarative language into that hungry space at the time.

Accepting that seems to imply that there is no space hungry for 3d markup at the moment. Collada and X3D are available but not (yet?) popular on the Web.

By exposing WebGL, which is extremely low level but also somewhat limitless, we now have the luxury of waiting to see what libraries people build upon it. Maybe then it will be more obvious what areas 3d markup could be useful.

Dean