Re: Ontologies of the Web

Sandro Hawke wrote,
> Miles Sabin wrote:
> > All are legitimate. All have their uses. All of them can be used as
> > guiding principles for building systems which do useful work. And
> > if someone adopts one rather than another that won't stop any code
> > from working. So I think it's critical that the Web Architecture
> > not underwrite any one of these views at the expense of the others.
>
> I'm torn.  On the one hand, yes, this is what ontologies are all
> about.  We need to model things, sometimes the same things, in
> different ways for different applications.  I recently started a page
> of Ontologies of the Web [1], which I just updated a bit, ... but
> it's hard work.  Ontologies (like computer programs) are best
> developed and maintained in the light of real use cases.
>
> On the other hand -- architectural guidance is important.  There are
> surely many naive AND BAD models of the web, like the common
> first-attempt ones which don't account for changing content, and many
> more subtley bad ones.

OK, I guess I exaggerated to make a point. I'll qualify.

Some are useful in some circumstances, some in others. Some are complete 
junk. The proof of the pudding is in the eating, and the test is 
whether by using a model you're able to build robust and useful 
systems, and one of the components of a skilled engineers skill is an 
ability to pick an appropriate model most of the time (cp. OO vs. 
functional vs. procedural programming models).

> Perhaps best would be several completed models, with both gentle
> tutorials and formal ontologies, endorsed with commentary by the TAG
> as being appropriate for certain kinds of work.

This sounds like an excellent proposal.

Cheers,


Miles