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

Re: Question on PaceLinkRelMechanism / Link Rel Schemas


Graham wrote:

On 22 Mar 2004, at 4:08 pm, James Aylett wrote:

It's a trade-off. If we can get a mechanism that just uses XML
namespaces (since we're already committed to using them), then
great. But obviously not if we sacrifice something useful - and also
not, from a practical point of view, if it makes life significantly
more complex for anyone (since "cleanness" proposals have run into
practical problems in the past).


I doubt it would be that hard to implement, but I but it would be very easy to do it wrongly and equally easy for publishers to mess it up.

Maybe.

Plus it's a big mess of extra code and extra spec for very little.

That depends - at the moment there is no extension mechanism defined beyond dumping any material from other namespaces anywhere. Having some form of well-defined method of using entities and relationships beyond those of core Atom is IMHO massively important. Until we have an alternative, this proposal represents a whole lot more than "very little". (I wouldn't have exactly called it a "big mess of code...", but I take your point).

I'd veto any proposal that adds a redundant namespace mechanism on top of XML's perfectly good one.

The mechanism is essentially the same as XML namespaces, though the concrete syntax differs. This is admittedly a big disadvantage as much tool support is lost.

As far as I'm aware there is no suitable veto mechanism in the current Atom decision-making process or in IETF groups, but if there was I'd veto release of Atom 1.0 until we have a decent extension mechanism - the <link> approach is one possibility, and your earlier suggestion (appropriately fleshed out) is another.

Cheers,
Danny.