Re: [SVG Transforms 1.0, Part 2: Language (2009-03-20)] - Feedback for "3. TransformList Extensions"


another option would be another name for the attribute
of course. There is already transform, gradientTransform
and patternTransform, one could introduce perspectiveTransform
(or something like this) to avoid conflicts and backward incompatibilities.
Well, still different transform functions can be confusing for some 
authors ...
The advantage of another attribute would be too, that this
could act independently from other transformations, as for
example animateMotion already is.

About right handed or left handed coordinate system and the
filter coordinate system: They need not necessarily the same
coordinates, it is just simpler to explain to authors, that z-directions
in different chapters or modules of SVG point in different directions.
Because filters are already part of a recommendation and are already
in use, it is to late to change.
Until it is to late, I think it is preferable to avoid this in the current
drafts. If it is not avoided, it needs to be discussed and noted in all 
parts carefully. And still several readers will be surprised by different
directions of the z-axis. We can assume, that several authors are
not very familar with matrices, therefore some careful prose about
what results from the mathematics for the 'simple user' would help
them anyway.
With the proposed z-rendering-order indication there is now a third 
part defining a meaning of a z-direction within SVG and maybe with 
the CSS-transforms there is a fourth. Looks like a lot of z-directions 
for an originally 2D format ;o) 
Because we can assume that the spatial sense
of several authors is somehow limited, every confusion may
result in additional frustration.
Personally I have no much problems in switching from left
to right handed systems or from switching the z-direction for
every document, if necessary. But I think, there are less flexible 
people having problems with this. Several seem to have already 
problems to imagine what are results form a matrix transform ;o)


Received on Friday, 4 December 2009 10:24:35 UTC