Re: request for guidance: centralized extensibility for HTML5 using X3D graphics

On 23 Feb 2010, at 23:03, Chris Marrin wrote:

> On Feb 23, 2010, at 12:29 PM, Krzysztof Maczyński wrote:
>> Chris,
>>> GeoNodes, metadata, and even all the primitives (Box, Sphere, etc.) could be easily done in a JavaScript library and kept out of the core node set. There would also be no need for TimeSensors or other timing or animation related nodes. You could also eliminate the complexity of Prototypes, which can be done in a JavaScript library.
>> Your point of view doesn't generalize beyond scripting-centricity. Standard arguments about declarative approaches being superior under many circumstances, especially in markup, apply here.
> I'm not sure what you mean by that statement. Advocating the use of scripts to extend a given set of functionality (like 3D) is not at odds with a declarative approach. X3DOM is proof of that. The question is what needs to be hardcoded into the browser. I think WebGL is a great first pass at what is needed.

Right. WebGL was a important step and I hope it stays forever. 
Without WebGL it would be (almost) impossible to implement something like X3DOM today.

> Any further steps have to be done with a minimalist approach or they will never gain any traction. The main point of my statements was that you don't need to give up functionality because of that minimalist approach.

Well, there are at least some issues on the current UA/JS/WebGL layer.
There is e.g. no 3D-Sound API which is a must for immersive, game like application.
(and needed for the X3D-Sound node)

Knowing the type of image before it get's uploaded to WebGL would also help.
For some application (e.g. X3D lighting model) it is important to know if
it is an gray-scale, RGB or RGB+Alpha texture. 
This is not just X3D-related but important for sorting in any (forward) rendering system.

I personally think there should be a method to ask an HTML-Image for 
the number and types of channels (and not WebGL).

But both are right now minor issues. We will see how this evolves. 

However, a native X3D implementation, as part of the UA, could easily solve this issues. 


> -----
> ~Chris

Dr. Johannes Behr
Leiter Projektbereich VR

Fraunhofer-Institut für Graphische Datenverarbeitung IGD
Fraunhoferstr. 5  |  64283 Darmstadt  |  Germany
Tel +49 6151 155-510  |  Fax +49 6151 155-196  |

Received on Wednesday, 24 February 2010 17:55:18 UTC