Re: SYMM WG comments on XBase 2nd last call

Thanks for your support in having this re-examined, Jonathan.  Patrick
Schmitz can communicate how important referential bases are better
than I can.  My description of the issue is that in multimedia (SMIL)
documents, much more than in hypertext (HTML) documents, there is the
use of a large number of many different media types.  These different
types vary greatly in how much bandwidth they use.  They also vary in
how they are delivered, with the primary example being the growing use
of streamed media.  Thus, it is common practice in multimedia for
different media types to be distributed from different servers that
are each specialized for distributing certain media types.  These
different media types would be scattered throughout the SMIL
hierarchy, since the SMIL hierarchy is determined by timing
structure. Thus, you could not group common media types sharing common
bases into hierarchies so that XBase as it currently stands could be
used.

Any SMIL presentation with streamed media, which I believe most have,
will also have non-streamed media, such as images and text.  Streamed
media needs to come from a streaming server.  Other media would more
often be delivered from a static file server.  These documents would
thus use at least two servers, perhaps more.

Another use case is annotation, which many people looking into SMIL
are considering using it for.  The media that is annotated and the
media comprising the annotation would most likely come from different
places.  Yet another use case comes from the use of the <switch>
element to select a media object from alternative servers, in case the
primary one is not available.  One could have switches through the
document hierarchy separating primary from backup servers.

The defining characteristic of multimedia is that is brings together
many different types of media components and integrates them in what
is often an intricate fashion.  It is this characteristic that makes
referential bases so important for SMIL.

-Lloyd

On Mon, Jul 3 2000 Jonathan Marsh wrote:

> Indeed it complicates things greatly.  The WG already rejected this type of
> functionality because of the level of complication introduced.  But I think
> re-examining this decision is warranted because of the new information
> revealed in this thread:
> 
> 1) SYMM may have use for this kind of capability.  Previously we had
> considerred this only in the context of XHTML.  So we've doubled the number
> of "customers".
> 
> 2) The xml:base mechanism may not be suitable for extension by languages
> needing it, because a different media type (text/smil vs. text/xml) may
> produce different results.  If our design precludes extension, we may want
> to rethink a bit.
> 
> 3) Adding this in an XML Base 2.0 appears quite difficult.
> 
> I agree that this feature has a high complexity price tag, and that a number
> of workarounds exist.  I trust that SYMM will communicate just how important
> this capability is for them so we can evaluate the tradeoffs.

--
Lloyd Rutledge  vox: +31 20 592 41 27       fax: +31 20 592 41 99
CWI             net: Lloyd.Rutledge@cwi.nl  Web: http://www.cwi.nl/~lloyd
Post:   PO Box 94079   |  NL-1090 GB Amsterdam  |  The Netherlands
Street: Kruislaan 413  |  NL-1098 SJ Amsterdam  |  The Netherlands

Received on Tuesday, 4 July 2000 03:23:39 UTC