On Sat, 19 Mar 2005 15:25:24 +0100, Thomas Broyer <[EMAIL PROTECTED]> wrote:
> Well actually I was thinking about creating a "unified feed API" (in > .NET, but could be ported to other languages) and trying to reduce > memory footprint by reusing Person objects, as well as providing > relations (bidi-arcs) between a Person and the Entries he/she has > authored or contributed. Yep, I don't see why not. Irrespective of memory etc, this would seem (IMHO) the best approach to Atom Persons if you're programming with objects. The Atom model is only specified through prose and the implications of syntax constraints, but I'm sure there's enough to start a "unified API". I personally believe the best approach to this is through a unified machine-readable model, specified in RDF/OWL. This is more-or-less the approach to Atom being taken by Henry Story in BlogEd (Java) and Mark Nottingham for Sparta (Python). > End of discussion, until next Atom version ;o) But only the beginning of implementation! Cheers, Danny. Early OWL model draft: http://semtext.org/atom/ Henry's refactoring/implementation: http://bblfish.net/work/atom-owl/2004-08-12/blogexample.html Mark's sparta notes: http://www.mnot.net/blog/2005/03/17/sparta -- http://dannyayers.com
