Re: [css3-images] Proposed Gradients changes

Given that the only objection was from Brad, and I've resolved all of
his objections save a minor syntax-related one, I've gone ahead and
made the changes to the draft.

While there, I rewrote and slightly reorganized the prose to make it
easier to read as well.

I also decided against adding an interpolate() function for now.  With
the solution I pursued for linear gradients it wasn't necessary to use
interpolate() for them, and introducing a function like that for only
radial gradients seemed like overkill.  If anyone still thinks it's a
good idea, we should add it to Transitions or V&U instead.  Instead I
just defined the used value of gradients as being the explicit
'canonical' form, and interpolation to take place over used values.

So, quick rundown of actual changes:

1) Linear gradients are now defined in terms of angles, not endpoints.
 You can either provide an angle, or keywords specifying a side or
corner.  The latter is mapped to an angle at used-value time.  The
definition of how to determine the endpoints from the angle has been
slightly adjusted for conceptual purity; this should have no actual
effect.  Overall this is slightly weaker than the previous definition
of linear-gradients, as you can no longer specify an arbitrary point
as the starting-point.  I believe the loss is minor, and worth the
simplification in the model.

2) Radial gradients now have a way to explicitly set the size/shape of
the ellipse by providing two lengths, taken to be the lengths of the
horizontal and vertical axes.  The keyword-based size/shape values
still exist, but are converted into the lengths at used-value time.

3) Serialization of gradients has been amended to take into account
the syntax changes.

4) Interpolation of gradients has been defined.  I define the
constraints that the start and end gradients must satisfy to be
interpolable, and then precisely how to interpolate them.

5) A lot of minor and functionally inconsequential text editting and
formatting for easier reading.

~TJ

Received on Saturday, 20 November 2010 00:25:04 UTC