No subject

Sat Apr 15 14:32:53 EDT 2006

have the matrix manipulations in a separate package in case they are needed
in other packages (say a 3d extension, shading, etc)

I'm already looking into the matrix transformations for VML and will report
back tomorrow at the meeting, with findings and questions. My general
approach is to decouple the math logic from the vml representation, so that
all transformations are done at the level of the dojo.gfx layer and are
directly applied to vml.


On 7/4/06, Gavin Doughtie <gavin at> wrote:
> I would certainly expect the matrices to be 3x3, and at least compatible
> with 4x4. I wouldn't be surprised if Eugene had his reasons, though.
> I think Eugene's idea was that if you wanted to do a lot of matrix
> manipulation, you could do that as a separate step and then set or apply
> the transform at the end of the pipeline rather than making the shape do
> it. The idea being that if you didn't need complicated transforms, that
> you wouldn't need to include the code.
> In that vein, some of the utility methods to create shapes (polyline,
> polygon, etc.) might even go in their own package for finer grained
> control over download size.
> At any rate, we're going to need the matrix code you talk about, it's
> just a question of where to put it. dojo.gfx.matrix perhaps?
> There's also the issue of implementing matrices with VML, which doesn't
> support them directly. It does, however, have the idea of a "formula"
> node that can be contained by a shape and procedurally change the shapes
> vertices. I think. I'm hoping Ioannis can dig into that, unless there's
> somebody else reading this who already has.
> Gavin
> Tom Trenka wrote:
> > Gavin,
> >
> > Looking good so far.  The only thing I think I'd like to
> > change/address is the way we're doing matrix transformations...under
> > the current codebase, you apply the matrix immediately.  SVG supports
> > (in almost everything but it's very important with transformations)
> > what amounts to a history of transformation matrices; basically it
> > allows you to push a set of matrices, in order, into a queue and then
> > apply them, in order, all at once.
> >
> > I would like to see if we can do this with the drawing API as well.
> >
> > Also, I notice that the current matrix transforms are based on 3 x 2
> > matrices; IIRC 3 x 3 matrices are really required for proper
> > transformations (I have to go look again, it's been a while).  If it
> > helps at all, I sent the the f(m) Drawing code specifically for the
> > matrix transformations (at the beginning of this discussion), and I'd
> > be happy to add that to this API if you want.
> >
> > trt
> >
> > On 7/4/06, Gavin Doughtie <gavin at> wrote:
> >
> >>
> >>
> >> Copy src/gfx to dojo/src, tests/gfx to dojo/tests, then run
> >> test_gfx.html.
> >>
> >> You should see something similar in both IE and Firefox, but there's
> >> still a lot of work to do in the VML/IE side (particularly using vml
> >> formulas for the matrix transforms and making the graphics, well,
> >> unified...).
> >>
> >> Just re-sending the URL since the zip file was slightly too large for
> >> the mailing list.
> >> _______________________________________________
> >> dojo-contributors mailing list
> >> dojo-contributors at
> >>
> >>
> > _______________________________________________
> > dojo-contributors mailing list
> > dojo-contributors at
> >
> _______________________________________________
> dojo-contributors mailing list
> dojo-contributors at

Content-Type: text/html; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Content-Disposition: inline

More information about the dojo-contributors mailing list