Re: LANG: first sketch

On March 7, Frank van Harmelen writes:
> Smith, Michael K wrote:
> 
> 
> > Great document.
> 
> 
> Thanks!
> 
> 
> > Two comments.
> > 
> > 1. Negation and disjunction considered hard?
> > 
> > I would place these in the class "easy to understand by our target
> > group".  Is your categorization due to complexity for tool
> > builders/reasoners?
> 
> 
> Certainly reasoning etc becomes harder when disjunctions are involved.

Frank,

I wouldn't make such a bold statement. There are many ways in which
the same effect can be achieved through the "back door", e.g., using
disjointness. The language would have to be MUCH more tightly
constrained for reasoning to be easier.

Ian


> 
> For the "target group", of course we can argue about this. My experience is that:
> - people don't use union of classes all that much
> - the only use of negation is for disjointness statements,
>    and these are included as separate items
> Also: people can't/don't deal with nestings of these.
> But I agree, it's certainly an area where I can well imagine we go the other 
> way than our "first stab".
> 
>  
> > 2. Syntax (nag, nag, nag)
> > 
> > Determining the semantic components of OWL should be our priority,
> > no question.
> > 
> > The only thing I take exception to here is "we expect that
> > a single syntax won't do".  I don't know quite what that means.  In
> > one sense, I agree whole-heartedly, let a thousand flowers bloom.
> > 
> > That said, we are defining a language.  There must be a rigorous
> > statement of what the sentences of that language are.  These are the
> > strings for which our semantics will provide a meaning.
> > 
> > One syntax description will be primary.  Nothing prevents anyone from
> > providing what they think are better human or machine engineered
> > syntax on top of this.  In particular, the WG can specify a
> > translation from the definitional syntax to an alternative we deem
> > critical.
> 
> 
> I agree. One syntax should be primary. The document should have expressed that 
> better. I think your last paragraph says it exactly right (provided the 
> translation is both ways).
> 
> Frank.
>     ----
>

Received on Thursday, 7 March 2002 17:33:30 UTC