[sc34wg3] Look Ma! No Properties!

[Robert Barta]

On Fri, Jul 16, 2004 at 03:30:38PM +0200, Jan Algermissen wrote:
> Steve Pepper wrote:
> > 
> > * Jan Algermissen:
> > 
> > | Wow....you are the first person I see to come up with a purely declarative
> > | merging rule, which is (warning: implementation issue ahead :o) extremely
> > | important in order to get efficient merging. To make it possible at all.
> > | At least if you want to have 100% generic merging rules.
> > | (Note: IMHO, a merging rule is simply a constraint. In the case of tau, it
> > | is a constraint on the set of all members.
> > 
> > I'm curious to know why the examples Lars Marius and I gave
> > in http://www.jtc1sc34.org/repository/0497.htm don't qualify
> > as "purely declarative merging rules". 
> 
> I can't see one that does not involve procedural aspects (but
> you might want to point me to one that I missed).
> 
> Maybe our understanding of 'declarativeness' differs, though.
> 
> It is more precise to say that Robert's 'merging rules' are
> constraints on the topic map and that I have not seen *this*
> before.

Steve,

I see it that way that 'merging' is an operation that is (a) triggered
on two or more vortices (uhm topics) and (b) does something with them.

The tolog code

   basename($T, $BN) :-
   topic-name($T, $N),
   value($N, $BN).

   select $T1, $T2 from
   basename($T1, $BN),
    gabx($T1, $X),
    gaby($T1, $Y),
   basename($T2, $BN),
    gabx($T2, $X),
    gaby($T2, $Y),
   $T1 /= $T2?

takes care about (a), but not of (b). (b) is covered by the generic
merging rules defined in TMDM.

BUT: Does this makes sense in all circumstances? In the example of the
cities which should merge if their are _sufficiently close_, for
instance, generic merging would add up all characteristics of $T1 and
$T2 and would drop them into a new topic. Does this now mean that that
new topic has a whole lot of new coordinates so that we end up with
more than one latitude and more than one longitude?

How can an application actually control it? In the above case what we
want is to merge freely the names, that is ok. What we also want is
that the longitute/latitude coordinates are combined 'properly',
whatever this means for our application.

TMQL to the rescue:

   http://topicmaps.it.bond.edu.au/docs/23/toc

If all pans out well, TMQL can generate maps, i.e. generate topics and
associations from existing ones. Exactly what we need to cover
application cases which are not happy with the generic 'drop
everything in this basket' merging.

Needless to say, that generic merging is MUCH faster as it can be
hard-coded in TM processors.

\rho









> 
> Ok?
> 
> Jan
> 
> 
> 
> 
> Is it simply because
> > you didn't see that document, or have I missed something?
> > 
> > Steve
> 
> -- 
> Jan Algermissen                           http://www.topicmapping.com
> Consultant & Programmer	                  http://www.gooseworks.org
> _______________________________________________
> sc34wg3 mailing list
> sc34wg3@isotopicmaps.org
> http://www.isotopicmaps.org/mailman/listinfo/sc34wg3
>