Re: SVG animations without SMIL

Erik,

I agree with Dirk.

Interpolating between path definitions in SVG at the moment is
straightforward by putting a straight jacket on what is allowed.
The two path definitions must be quite similar, curves to curves, arcs to
arcs and lines to lines.
There is freedom as to whether the definitions are relative or absolute and
the type of bezier curve used.

At the level that SVG exists that may be the correct approach. Some minor
changes would help but that is all.

For example, in figure animation, animating an arm when the person touches
his shoulder isn't easy.

In 1908, Emile Cohl had problems with inbetweening that. He gave up
straightforward interpolation and tried articulated cutout models and then
gave up on those as they didn't look realistic.

In 1966 Burtnyk and Wein at the National Film Board of Canada had the same
problem in their computer animation system and solved it in their figure
animation by adding single axis rotation as part of the interpolation. As
well as interpolating path segments between path A and path B you define an
anchor point about which parts of the path rotate.

In 1969, Computer Image of Denver built a system called CAESAR (Computer
Animated Episodes for Single Axis Rotation) which had almost the same
algorithm built into their video animation systems.

Later systems started with skeletons and put flesh on them and so on.

And none of this helps the person who has a puddle that is expanding when
it hits the floor. The constraints are quite different and depend on what
the puddle is made of and the floor surface.

I think SVG should stick to animating simple paths and not go too far in
trying to second guess what the application might want in terms of
interpolation.

It is probably best left to possible several higher level systems that
handle constraints and sit on top of SVG.

>From the original submissions made to W3C, the decision was made to base
the first graphic rendering system, SVG,  at a low level and look at
higher-level constraint-based solutions later.

Perhaps now is the time to do that?

Bob

Received on Thursday, 4 June 2015 10:35:13 UTC