RE: Semantic Web Services: Interest Group creation discussions

> > > [...] Layered meta data, service chains and
> > > schema driven semantic approach to web services should
> > > be considered for standardization.
> >
> > Sorry to be a grouch, but I think it is WAY too early
> > for standardization in this area.
> > Let's see some systems crop up, and prove their value.
> > Then, when *they* want to interoperate, it will be time
> > to look at standards.
>
> Agreed. I do think its timely to start *early* discussions about
> requirements though.

I would go further than Jon on this.

"Let's see some systems crop up..." - systems are cropping up, and some are
proving to have value. "When *they* want to interoperate..." - many of them
already do. It may be too early for typed-in-stone standardization, but
things need to be sketched out asap to avoid balkanization. Case in point:
the weblog domain. Here developments are happening rapidly, but most are of
an ad hoc nature, put together by individuals/independent groups wishing to
overcome a single problem or implement a single bit of functionality.

In this environment the solution that is easiest in the short term has a big
advantage. Back-of-envelope specs like Userland RSS get a head start,
whereas solutions that offer advantages in the longer term by solving whole
classes of problems (like RSS 1.0) are disadvantaged. Longer term, quick but
limited solutions are likely to decline due to natural selection, but in the
interim all kinds of interop problems have been generated.

In service terms in this domain we're seeing things like the Blogger API and
trackback appear, with rapid adoption. Were there SWWS-friendly options
available early on and a little growth hormone applied by developers that
want good standards, perhaps a lot of wasted effort could be avoided a
little farther down the road.

Cheers,
Danny.

Received on Tuesday, 18 March 2003 11:59:49 UTC