- From: Raphaël Troncy <Raphael.Troncy@cwi.nl>
- Date: Tue, 10 Mar 2009 15:54:47 +0100
- To: Conrad Parker <conrad@metadecks.org>
- CC: Media Fragment <public-media-fragment@w3.org>
Dear Conrad, > The distinction between use of a fragment or a query introduces > different retrieval protocols. This is a technical distinction and has > little to do with semantic purpose such as user context. Well, not exactly. The query creates a *new* resource that is unrelated to the original resource, i.e., you can't tell that http://www.example.org/myVideo.ogg?t=15,45 is unrelated to http://www.example.org/myVideo.ogg. You don't have this problem with the fragment. > Retrieval of segments by fragment lacks backwards-compability with > plain HTTP, whereas retrieval by query can fall back to a direct HTTP > download of the segment. The web architecture does not break when servers and clients do not understand things, they just ignore it :-) So, in the worst case, a complete resource (instead of a segment) will be sent, but this is the default behavior now. I don't see where there is no "backwards-compatibility". A query creates a new resource: it is fine if you want to do that, but for most of our use cases, we want to have a relationship between the fragment and the resource it comes from ! > This is a topic that's been discussed in the Annodex project since > about 2001. I'm hoping this WG can avoid some of the mistakes we made > early on in that, such as trying to use fragment syntax where it is > inappropriate. What do you think it is inappropriate to use the fragment syntax when actually we want to express what is commonly understood as a fragment, and *not* as a query? Cheers. Raphaël -- Raphaël Troncy CWI (Centre for Mathematics and Computer Science), Science Park 123, 1098 XG Amsterdam, The Netherlands e-mail: raphael.troncy@cwi.nl & raphael.troncy@gmail.com Tel: +31 (0)20 - 592 4093 Fax: +31 (0)20 - 592 4312 Web: http://www.cwi.nl/~troncy/
Received on Tuesday, 10 March 2009 14:55:39 UTC