[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: The <id> element


Asbjørn Ulsberg wrote:

Before we can finish Atom and say «The <id> element should contain
a string of the 'tag' URI scheme», I think we need a final
specification of the scheme, and Atom might help to push the
development of the scheme forward, or maybe we should develop the
scheme forward ourselves.

I don't think it makes sense to limit URIs by scheme, in Atom or (almost) anywhere else. A URI is a URI, and there are going to be new schemes invented, and in the meantime very few of them have any practical advantages over http: and urn: URIs. It's just a string, after all. In my own work a string that can (potentially) be dereferenced to obtain information is always a better choice than one that can't, so I tend to use http: URIs for everything, but others have different trade-offs.

--
Cheers, Tim Bray (http://www.tbray.org/ongoing/)