Simon St.Laurent wrote,
> (Mark Baker) writes:
> > We need less protocols, not more, because we need more
> > interoperability, not less.
>
> Hmm... that feels to me a lot like the trade-offs in XML between
> obsessively standardizing for interoperability and allowing people to
> customize systems to meet their own needs.
It's _exactly_ like the trade-offs in XML.
I've said it before, and I'll say it again: the costs of standardization
(of protocols, of schemas, of ontologies, of whatever) is typically
wildly underestimated. Technically it's a doddle, but in practice the
creation and widespread adoption of a standard is a political act. It
involves large numbers of actors, with a disparate mix of cooperative,
competitive and changing interests, somehow or other reaching some kind
of consensus. As a lot of us on this list know, this is a time-
consuming, expensive and often frustrating process which often produces
results which satisfy noone.
If you weigh the costs of agreement against the gains in
interoperability then at least some of the time interoperability loses.
And in that case a fork isn't just an option, it's the Right Thing to
do.
Cheers,
Miles
