Re: minutes, 3 November 2011 FX Taskforce F2F meeting at TPAC 2011

Hi David,

On 3/11/11 7:28 PM, David Dailey wrote:
> First just a comment: my goodness what an extraordinary number of
> things you've talked about -- I can imagine your heads must hurt.
> Thanks to the note-takers!

Not my head, but my hands after a day of minute taking. :)

...
> Is the work on SVG Filters now separate from the work on
> compositing?

They were always separate, I believe.  Compositing, since being split
out from the old SVG 1.2 Full WD, has never been published separately.
Filter Effects has been progressing as a joint SVG/CSS deliverable.  The
SVG Filters specifications is going away in favour of the Filter Effects
specification.

> Is the implementation of 3D transforms only going to be done at the
> level of the bitmap and not to the actual geometry of the vectors
> affected? and there are accessibility reasons for applying SVG
> filters (that preserve geometry and not just bitmaps) to HTML
> content. (e.g. colorblindness)

The way 3D transforms would be done in hardware would be to render the
SVG content to a raster, upload that as a texture, and then have the GPU
apply the 3D transformation to it (AIUI).  It's not clear to me what the
exact accessibility concerns with that would be.

> Is there stuff that Porter-Duff can do that feComposite and
> feColorMatrix together can't?

Seems like it might be possible to implement the Porter-Duff blend modes
with those.  The blend modes probably not.

> How about if you add Vector effects (to allow the Euler diagrams)?

Not sure that would help.