SVG12: Manipulation of animation elements

Dear Scalable Vector Graphics Working Group,

  From http://www.w3.org/TR/2004/WD-SVG12-20041027/ and any of its
dependencies it seems unclear how manipulation of animation elements
must be handled by SVG 1.2 implementations. The following SVG 1.2
document demonstrates what I believe is a bug in the Adobe SVG Viewer
6.0 Build 38363,

  <svg
    xmlns               = "http://www.w3.org/2000/svg"
    version             = "1.2"
  >
  
  <text y = '100' font-size='50'>
    <tspan>Scalable Vector Graphics</tspan>
    <animate
       begin            = '0s'
       attributeType    = 'CSS'
       attributeName    = 'opacity'
       from             = '0'
       to               = '1'
       dur              = '2s'
       fill             = 'freeze'
       id               = 'c'
    />
  </text>
  
  <script type="text/ecmascript">
  <![CDATA[
  
    var node = document.getElementById('c');
    node.setAttributeNS(null, "begin", "2s");
  
  ]]>
  </script>
  
  </svg>

For this example, the viewer will execute an animation that does not
exist when the document begins (the animation in the <animate> element)
and the animation that exists in the document tree after executing the
script. Assuming that this behavior is conforming, one would expect that
after replacing begin='0s' by begin='4s' "both" animations would still
be executed just delayed by 2 seconds. This is however not the case, the
viewer will only consider the animation that begins at 2s. If the "two"
animations overlap, the viewer will restart the animation with the new
parameters provided by the script. With begin='0s' and the <script>
element replaced by

  <script type="text/ecmascript">
  <![CDATA[
    var node1 = document.getElementById('c');
    var node2 = node1.cloneNode(false);
    node2.setAttributeNS(null, "begin", "1s");
    node1.parentNode.replaceChild(node2, node1);
  ]]>
  </script>

the effect is that "both" animations start at their corresponding begin
where the "second" animation restarts the animation but it ends with the
end of the first animation such that you end up with half-opaque text.
Strangely, if the animate element is removed via scripting no animation
is executed.

The behavior I expect is that processing of animation elements in the
document tree starts when the document begins such that elements that
are not in the document tree at that moment are not considered, etc. I
could not find any clear requirements in the SVG 1.2 Working Draft or
its dependencies that requires the behavior I expect or that supports
the behavior of the Viewer I've mentioned.

Thus, please change the SVG 1.2 Working Draft such that it clearly re-
quires the behavior I expect and include proper test cases for this re-
quirement in the SVG 1.2 test suite.

Furthermore, the Second Edition of the SMIL 2.0 Recommendation states
in http://www.w3.org/TR/2005/REC-SMIL2-20050107/animation.html that
specifications should limit manipulation of animation elements after
the document has begun and make certain other restrictions. It seems
that the SVG 1.2 Working Draft does not satisfy these requirements,
please change the SVG 1.2 Working Draft such that it satisfies these
requirements, or such that it notes why it does not satisfy them and
defines all relevant processing.

Finally, the requirements for processing animation elements that are
dynamically added after the document has begun is not clear to me. It
is for example not clear to me whether such an element begins relative
to the document begin or relative to the current presentation time. I
would expect them to begin relative to the current presentation time
but it seems easy to argue the other side and even think that such an
animation would not begin at all. Please change the SVG 1.2 Working
Draft such that this is clearly defined.

If any of this happens to be unclearly defined already, please change
the SVG 1.2 Working Draft such that this is more obvious, for example
by an informative section that refers to the normative requirements.

regards.
-- 
Björn Höhrmann · mailto:bjoern@hoehrmann.de · http://bjoern.hoehrmann.de
Weinh. Str. 22 · Telefon: +49(0)621/4309674 · http://www.bjoernsworld.de
68309 Mannheim · PGP Pub. KeyID: 0xA4357E78 · http://www.websitedev.de/

Received on Saturday, 8 January 2005 18:35:41 UTC