Re: SVG animations without SMIL

On Thu, Jun 4, 2015 at 2:40 PM Brian Birtles <bbirtles@mozilla.com> wrote:

> On 2015/06/04 12:04, Shane Stephens wrote:
> >
> >
> > On Thu, Jun 4, 2015 at 11:18 AM Brian Birtles <bbirtles@mozilla.com
> > <mailto:bbirtles@mozilla.com>> wrote:
> >
> >     On 2015/06/04 3:20, Philip Rogers wrote:
> >      > For consistency with the other geometry attributes, I don't see
> many
> >      > costs to promoting 'd' as a simple string presentation attribute.
> The
> >      > proposal would be to have it work just like animating 'd' works
> >     today:
> >      > string -> string where the interpolation is done on the parsed
> >      > representation.
> >
> >     I'd prefer we don't do this. Apart from the obvious difficulties with
> >     SVG path syntax, the requirements for interpolating between two paths
> >     are so restrictive (same number and type of segments) as to make
> >     hand-authoring path animations almost impossible.
> >
> >
> > FWIW we could pretty trivially relax this requirement when interpolating
> > in CSS. There's no fundamental reason for forcing matching segments -
> > all segment types can be arbitrarily split and from memory there are
> > adequate representations of each pairwise conversion.
>
> When you actually go to author content though, you very often need
> anchor points or some way of describing *how* you want to go from A to
> B. For example you often want to say, "This corner point here, it
> corresponds to that corner point in that path (despite the fact that the
> ratio of segments before and after is completely different)."
>
> Flash has shape hints for this and you can see the difference between
> the default tween and a hinted tween between the outline for M and N
> here.[1]
>

Tav's superpaths proposal works well for this and cleanly extends the path
syntax.

Cheers,
    -Shane


>
> [1] http://www.adobe.com/devnet/flash/learning_guide/animation/part12.html
>
> http://www.adobe.com/devnet/archive/flash/articles/concept_shape_hints.html
> also has another demo compared an unhinted morph from 5 to 3 with a
> hinted morph from L to Z
>