- From: Sjoerd Mullender <sjoerd@oratrix.nl>
- Date: Mon, 22 Mar 1999 10:47:19 +0100
- To: Leandro Marques Rodrigues <leandro@telemidia.puc-rio.br>
- Cc: www-smil <www-smil@w3.org>, Philipp Hoschka <ph@w3.org>
On Thu, Mar 18 1999 Leandro Marques Rodrigues wrote: > Hi, > > I have some questions about the specification of the begin attribute as > an event value. Can I specify the begin attribute of a child of a seq > element like this: > > begin="id(E1)(begin)", > > where E1 is an element "in scope", according to the specification? > Woudn't it seem strange, since I'm specifying that two children of a seq > element would start at the same time? Imagine if the element E1 is > declared after the element that references it, for example: > > <seq> > <ref begin="id(E1)(begin)" ... /> > <ref id="E1" .../> > ... > </seq> > > The beggining of the first element depends on the beggining of E1, but > the beggining of E1 depends on the end of the first element. So, how > would the element seq behave? > > For me, the explicit begin of a child of a seq element should be > specified only related to the effective end of its ancestor (or the > effective begin of the seq element, if it is its first child), and the > explicit begin of a child of a par element should be specifed only > related to the effective begin of the par element itself or of any other > child. I think you have discovered a slight shortcoming in the SMIL specification. In section 4.2.4.1, rule 2, there is a line that says: "It is an error if the explicit begin is earlier than the implicit begin of the element." But it should probably also say something to the effect that it is an error to refer to a later child in a seq. -- Sjoerd Mullender <Sjoerd.Mullender@cwi.nl> <URL:http://www.cwi.nl/~sjoerd/>
Received on Thursday, 25 March 1999 05:13:40 UTC