Developers can use reverse-FQDNs (like Java's packages) for their namespaces, which prevents collisions without actually requiring nesting.
-James A. Rosen On Apr 16, 2:51 pm, Marcel Molina <[email protected]> wrote: > More namespace nesting would of course increase people's ability to > taxonomize. It's a splippery slope though and we are trying to balance > expressiveness with simplicity. Providing for arbitrarily nested namespaces > increases complexity considerably both from an implementation perspective > and a comprehension perspective. > > > > On Fri, Apr 16, 2010 at 11:45 AM, gabriele renzi <[email protected]> wrote: > > > * What is an annotation more exactly exactly? > > > First off let's be clearer about what an annotation is. An annotation is > > a > > > namespace, key, value triple. A tweet can have one or more annotations. > > > Namespaces can have one or more key/value pairs. > > > first, annotations are cool, thanks. But why triples instead of quads? > > Say, we'd like to store three groups of movie data . If I do > > movie: rating: 5 > > > then we risk conflict with someone else using the same namespace in a > > different way, e.g. > > movie: rating: ***** > > > . At the same time, if I use the namespace for my > > application to avoid conflicts, I have to encode two of the fields in > > one > > > cascaad: movie_rating : XXXX > > or > > cascaad_movie : rating: XXXX > > > Did you consider this and decided it's not worth making the effort for > > a fourth field, or just ignored the issue, or simply didn't think of > > it? > > If triples are staying, can we establish a _convention_ early on on > > how to best handle this? > > > -- > > Subscription settings: > >http://groups.google.com/group/twitter-development-talk/subscribe?hl=en > > -- > Marcel Molina > Twitter Platform Teamhttp://twitter.com/noradio
