Re: Request feedback on fragment identifiers

Hi Al, Martin, Larry,

many thanks for your consideration of our work,
some short answers below,
Best regards

Myriam

Al Gilman wrote:

> Myriam,
>
> Please review the earlier discussion of "temporal URI fragments"
> which IIRC concluded that a name-value-pair-[list | bag] structure
> such as is provided by the ?search-part syntax is a better fit to the
> application than fragment syntax. At least where the application is
> time slices of media streams [temporally distributed media objects].
> I am not sure if the MPEG objects you are dealing with are internally
> polychronic but I would tend to think that most of the applications
> for subsetting or indexing into MPEG objects would be covered by
> time-sliced tails and segments. 

Actually, the syntax we have defined goes a little further than 
addressing a time point or a time range. The syntax also supports 
spatial and spatiotemportal locations and it uses XPath-like 
constructions for referencing a logical unit such as a chapter of a DVD 
or a Track of a CD audio. So yes, we need "more complex selection 
patterns".
I'm preparing a short summary of the syntax, that I'll send to Martin, 
Larry and you shorlty.

> http://www.w3.org/Search/Mail/Public/search?keywords=URI+fragment+temporal&hdr-1-name=subject&hdr-1-query=&index-grp=Public__FULL&index-type=t&type-index=uri 


Thanks a lot for the link.

> While that reference will ultimately lead you to Silvia Pfeiffer
> <Silvia.Pfeiffer@csiro.au> let me suggest up front that you might
> want to discuss your project with her. 

Actually, we initially harmonized our syntax toghether with Silvia, 
however, what we are proposing is *only* a syntax for addressing a 
fragment of an audiovisual content, there is no particular system 
behaviour on top of it, so for retrieving the fragment the URI spec 
rules apply (the fragment shall be handled by the client...).
Best regards,
Myriam

Received on Sunday, 11 July 2004 20:38:47 UTC