Re: SVGMatrix in Canvas spec

On 9/19/2012 2:37 PM, Ian Hickson wrote:
> On Wed, 19 Sep 2012, Cameron McCormack wrote:
>> Dirk Schulze:
>>> I just noticed that the Canvas2D specification references SVGMatrix.
>>> canvas.currentTransform returns a SVGMatrix object[1]. I am not sure
>>> if we should convince the Canvas editors to drive Deans Matrix4x4
>>> proposal[2] forward instead. I would like to rely on SVGMatrix as less
>>> as possible, since I still hope to get a API for CSS Transforms in the
>>> next level. What do you think?
>> When Hixie was developing that feature he asked if SVGMatrix was an
>> appropriate object to use here, and I said yes that would work.  I would
>> feel even better if we could actually manage some unification between it
>> and CSSMatrix / Dean's proposal.  If you have an idea how to do that,
>> please post it. :)
> The spec (at least the WHATWG spec) is just gonna follow what browser
> vendors do here. Right now it uses SVGMatrix because (as heycam says
> above), that's what browser vendors wanted, but I'd be happy to change
> that to whatever is more appropriate.
>
> Note that whatever is picked, the transform has to be limited to stuff
> that doesn't fiddle with the Z axis. I am no graphics expert, but it's my
> understanding a 4D matrix can end up making this possible? So we'd have to
> make sure we limit what the matrices can be set to...
>

I'd imagine we just drop the information from the extra dimension when 
used with Canvas.

Here in section 8:
http://dev.w3.org/csswg/css3-3d-transforms/#transform-functions

CSSMatrix has already been through this process, I believe, of 
supporting 2D then being extended later.

-Charles

Received on Thursday, 20 September 2012 00:02:48 UTC