- From: Sjoerd Mullender <sjoerd@acm.org>
- Date: Mon, 23 Sep 2002 13:21:07 +0200
- To: "Sigurd Lerstad" <sigler@bredband.no>
- cc: www-smil@w3.org
On Tue, Sep 24 2002 "Sigurd Lerstad" wrote: > > Hello, > > The SMIL timing model are talking about syncbase elements, e.g > > begin="video1.begin" > > and they're talking about events, e.g > > begin="video1.someevent" > > But the syncbase values begin/end are also events (DOM events) > > Reading up on the TimeGraph model, they set slightly different rules for > handling syncbases and events > > Is it correct that an implementation MUST differ between these two (instead > of also treating begin/end as events?) > (The statement above seems to indicate this, but I just want to make sure > (because it would be simpler if one could treat them the same :) Yes, an implementation must distinguish between events and syncbases. For one thing, a syncbase can be predicted, so if you use a negative offset to a syncbase. E.g. if you have begin="foo.begin - 5s", you can actually begin 5 seconds before foo begins. Events on the other hand cannot be predicted, so if you have begin="foo.activateEvent - 5s", the elements starts when the activateEvent happens, but 5 seconds *into* the element. > Is it correct that it's IMPOSSIBLE to specify e.g begin="video1.begin" and > tell the implementation to treat it as an event? (since the syntax for > syncbase would be the same? You can use begin="video1.beginEvent" which refers to the evens of video1 beginning and is an event in the SMIL sense and not a syncbase. This event triggers when video1 actually begins. -- Sjoerd Mullender <sjoerd@acm.org>
Received on Monday, 23 September 2002 07:21:22 UTC