Re: A proposal for a unitary RIF phase 1

From: Michael Kifer <kifer@cs.sunysb.edu>
Subject: Re: A proposal for a unitary RIF phase 1 
Date: Sun, 28 May 2006 03:25:58 -0400

> 
> > From: Michael Kifer <kifer@cs.sunysb.edu>
> > Subject: Re: A proposal for a unitary RIF phase 1 
> > Date: Sat, 27 May 2006 19:34:17 -0400
> > 
> > > Peter,
> > > 
> > > It is no big deal to be unitary by restricting the language to Datalog.
> > > You don't even need to limit it to a function-free sublanguage. In our
> > > roadmap the language was unitary also up to this point.
> > 
> > OK.
> > 
> > > The issue is how to build such a system in an extensible way so that it
> > > could be extended to satisfy most of the RIF requirements.
> > 
> > Well, this is at least one of issues.  However, I don't see any particular
> > preference for divergent semantics here.
> 
> I am not sure if I understand you here correctly, but what I meant was that
> anything we do in Phase 1 needs to have a clear path to enabling further
> extensions. 

Well, at least it has to have a path to satisficing the requirements of
Phase 2.

> These planned extensions will most likely influence Phase 1
> because, for example, we need a way to say what the syntax and semantics of
> a rule set is intended to be. 

I think that we need this already for Phase 1.  Without semantics, what
good is RIF?

> Even though this might seem unnecessary in
> the unitary world, this same rule set will have to live in a Phase 2
> world with other semantics, so it must be prepared to declare its
> characteristics in that larger context.

Well, this assumes, for starters, that there will be other, divergent,
semantics in Phase 2.  I do not conceed this point.  However, even if this
were the case, how do divergent semantics in Phase 2 require divergent
semantics in Phase 1?

> 	--michael  

Peter F. Patel-Schneider

Received on Sunday, 28 May 2006 07:44:27 UTC