Re: [SMIL] A question

On Thu, Apr 29 1999 herve_foucher@ds-fr.com wrote:

> 
> 
> Philipp Hoschka wrote:
> > Timeline 1 is correct, because of the paragraph in SMIL 1.0
>         ^^^^^^
> > Lloyd cites:
> >
> >   An element with a "repeat" attribute with a value other than
> >   "indefinite" has an implicit end equal to the implicit end of a seq
> >   element with the stated number of copies of the element without
> >   "repeat" attribute as children.
> >
> 
> Lloyd also wrote:
> > Wording in one part of the SMIL specification suggests the second
> > interpretation.                                           ^^^^^^^^
> 
> So... you do not agree with Lloyd?
> 
> My point of view, reading what Lloyd cites is:
> 
> 
> <seq repeat="3" begin="3s">
>  <img src="foo1" dur="4s" region="reg1" />
>  <img src="foo2" dur="4s" region="reg2" />
> </seq>
> 
> is therefore equivalent to:
> 
> <seq begin="3s">
>  <img src="foo1" dur="4s" region="reg1" />
>  <img src="foo2" dur="4s" region="reg2" />
> </seq>
> <seq begin="3s">
>  <img src="foo1" dur="4s" region="reg1" />
>  <img src="foo2" dur="4s" region="reg2" />
> </seq>
> <seq begin="3s">
>  <img src="foo1" dur="4s" region="reg1" />
>  <img src="foo2" dur="4s" region="reg2" />
> </seq>
> 
> So, TIMELINE 1 IS CORRECT. Does everybody agrees? Lloyd?
> 
>                  1    1    2    2    3
>         ....5....0....5....0....5....0
> seq        [------]   [------]   [------]
> foo1       ****       ****       ****
> foo2           ****       ****       ****
> 

I don't agree.  As the author of the cited part of the spec I can tell 
you what my intention was, even if the wording is (to say the least)
ambiguous and less than clear.

The intention is that the code

<seq repeat="3" begin="3s">
 <img src="foo1" dur="4s" region="reg1" />
 <img src="foo2" dur="4s" region="reg2" />
</seq>

is equivalent to

<seq repeat="3" begin="3s">
 <seq>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </seq>
 <seq>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </seq>
 <seq>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </seq>
</seq>

If we change this a little to make it even clearer, if in the example
you had par instead of seq, you would get

<seq repeat="3" begin="3s">
 <par>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </par>
 <par>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </par>
 <par>
  <img src="foo1" dur="4s" region="reg1" />
  <img src="foo2" dur="4s" region="reg2" />
 </par>
</seq>

Also, dur and end attributes are put in the outer seq.  If you don't
do that, the timing model doesn't make much sense.

-- Sjoerd Mullender <Sjoerd.Mullender@cwi.nl>
   <URL:http://www.cwi.nl/~sjoerd/>

Received on Thursday, 29 April 1999 10:28:59 UTC