Lists Home |
Date Index |
> I see type safety as being orthogonal to REST.
I was just talking about the time at which the client becomes aware of
the finer grained type that it's dealing with (i.e. beyond the generic
thing that supports GET/PUT/POST/DELETE), not any type of safety
> That's why I'm creating
> WRDL. If you want type safety, you should get it. If you don't, you can
> just ignore WRDL.
By "type" in this context, I mean rdf:type. WRDL doesn't help me there.
If I was told that a resource was an email inbox, but didn't know what
an email inbox was, I might still be able to find out that an email inbox
is a subclass of some generic container, so I'd know that I could POST to
it because of that.
I think there's value in knowing *why* something accepts PUT or POST,
rather than the simple fact that it does, which is available at runtime
anyhow by just poking at it with HTTP. Not that there isn't value in
WRDL, it's just interesting to see what can be done a level up. It
could be a very useful level for developers too, at least those used to
component based programming.
Mark Baker, Chief Science Officer, Planetfred, Inc.
Ottawa, Ontario, CANADA. firstname.lastname@example.org