Re: B.1, B.2

From: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>
Subject: RE: B.1, B.2
Date: Fri, 21 Mar 2003 14:34:26 +0100

> 
> Minutes:
> > > > JimH: Agrees with DanC. Suggests taking out B1/2 and close issue with
> > > > proviso that could be re-opened if volunteer appears before last call.
> > > >
> 
> Jeremy:
> > > I suggest the following change is all that is necessary.
> > >
> > > in S&AS 4.1
> > > [[
> > > Bnode identifiers here must be taken as local to each
> > transformation, i.e.,
> > > different identifiers should be used for each invocation of a
> > transformation
> > > rule.
> > > ]]
> > >
> > > ==>
> > >
> > > [[
> > > Bnode identifiers here are local to each transformation.
> > > When the construct being transformed matches the *restriction* or
> > > *description*
> > > productions from the abstract syntax then
> > > the bnode may be shared between multiple identical transformations of
> > > identical
> > > *restriction*s or *description*s. Otherwise the bnode used in each
> > > transformation
> > > should be unique for each invocation of a transformation rule.
> > > ]]
> > >
> > > ==
> 
> Peter:
> > This does not include changes to the equivalence proof.
> 
> Agreed.
> The proof is more difficult.
> 
> Two points:
> - the proof is informative, and so by the earlier discussion, could be left
> with a "to-do" until after last call. This seems to have been your
> suggestion about the graph as triples - since I do not believe we have any
> text that correctly describes the current graphs as triples.

The graph as triples section is not referenced by anything else in the
document, which makes it very different from the proofs.

> - I think the easier (but perhaps not most attractive) fix, would be to
> leave Lemma 1 and 2 unchanged except adding that the bnodes are not shared
> between multiple identical transformations, and then relaxing that condition
> in a new lemma (2.1?). This would then justify the merging of bnodes which
> denote the same classes. A more attractive option, requiring more detailed
> work, would be to inspect each use of the disjoint bnodes assumption in the
> proof, and make necessary changes.

I was planning on doing something along this latter line.  Given the other
demands on my time I can't promise to sufficiently attend to this until
mid-April, as I stated in a recent telecom. 

> Jeremy

peter

Received on Tuesday, 25 March 2003 16:35:20 UTC