Paul Everitt wrote:
[ snip URI layout ]
> One solution is an architecture (a la dmoz) that decouples the organization
> from the location. This is done with metadata (ours is based on Dublin
> Core). You then have virtual folder hierarchies that are actually metadata
> queries. On the downside, this requires some buyin from authors to
> participate in the system. Each of these folders can then optionally be RSS
> channels for syndication.
I'm forming the opinion that:
a) We're not going to acheive a URI layout that will be flexible enough
b) Content authors are never going to know where their new document
should be, particularly if they've written about several areas in one
document
c) Taking any random document, could we agree on what it should go all
the time?
I am therefore moving towards a "this document serves the following
areas: [list of areas" method, where the content gets uploaded and the
author selects from a [heirarchical?] list of areas on mozilla where the
document applies. We'd need certain information about the document, and
some method of achieving a filename that's rememberable i.e. not
/docs/gecko/27947562, but we could probably do it.
JG
