[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Public WebGL] WebGL extension registry




----- Mail original -----
> On Mon, Dec 6, 2010 at 4:44 PM, Steve Baker <steve@sjbaker.org> wrote:
> > Damn! And I'd *just* finished writing my shadow code! :-)
> >
> > I think this is a great thing - but it's a bit of a change in
> > direction
> > isn't it? I thought the plan was to have no extensions in v1.0 ?
> 
> No. The concern was over whether these extensions would be *part of*
> the 1.0 spec. Separating them into their own registry resolved that
> concern.
> 
> > Also, doesn't this open up the whole lack-of-anonymity can-of-worms
> > again?
> 
> Perhaps, but there are always cost/benefit tradeoffs. In this case, I
> think the availability of floating-point textures as a WebGL extension
> is so incredibly useful that it dwarfs any concerns over exposing one
> more bit of information that could be used to identify individuals.

FWIW, I have no particular problem with extensions, because the potential problem here is the same as we're already having with getParameter() MAX_... pnames, so this is comparable to adding just one more pname there when we already have many. I have yet to find a solution there for the long term; in the short term, as I explained, just disabling the RENDERER/VERSION/VENDOR strings mitigates the issue enough to make it not immediately exploitable afaics (because the remaining pnames/extensions are far from mutually orthogonal, they expose a lesser amount of information).

Oops, did I bite? Let's not have another long discussion on this topic :-)

> (Also, floating-point textures are likely to be supported on all
> hardware supporting WebGL, so this extension's availability is very
> unlikely to expose additional information.)

Great.

Benoit
-----------------------------------------------------------
You are currently subscribed to public_webgl@khronos.org.
To unsubscribe, send an email to majordomo@khronos.org with
the following command in the body of your email: