Re: Recursive graphs?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Jeremy Carroll wrote:
| This is basically the same apart from disallowing strucures that nest
| within themselves?

For me there is one relevant difference to the TriX model as I
understand it: an application/rdf+xml describes a graph; an
application/x-trix+xml describes a *set* of graphs. For example, I'm not
sure what should happen if I open a TriX document in an editor that's
made for editing a single graph. Hmm.

On the other hand, when you, as you suggested in one mail, define the
first graph as "asserted" and the other graphs as "whatever the first
graph says about them," you could take this syntax as meaning that the
other graphs are "nested" in the first one.

| On the surface syntax, which is what was driving the TriX papers, the
| nesting is too ugly (IMO).

I don't know; think of N3: It makes sense to me to think that a formula
is nested in a graph, more so in fact than to think that the formula is
delivered in a "graph set" together with a graph that uses the formula.
Especially when you're talking nested formulas, this becomes very strange.

On the other hand, I really dislike the idea of extending the abstract
syntax by allowing graph nesting! I would feel much more comfortable
with creating a de dicto reification vocabuary, have triple stores have
special support for it so that it's efficient, and providing special
syntax for it layered on top of TriX using a stylesheet. ;) ;)

I don't believe that a de dicto quoting equivalent to your proposal is
impossible; the crucial factor making this possible is that you always
quote the *whole* graph, which means that you can represent bnodes in
the quoted graph simply by a bnode in the quoting graph.

How do you feel about this approach? Probably you have considered it
while writing the paper and can immediately tell me why you don't like
it? ;-)

- - Benja

| But this sort of constraint might be the best
| way to address some of the difficulties.

Maybe; that's what I'm trying to work towards, so that I can make use of
this ;-)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org

iD8DBQFAMNBOUvR5J6wSKPMRAlVHAKDBXcZ63tnlt2gvtx2UY43q2pQBqQCeOjLI
WyI7UFGf2pISfNjB7r+dqe4=
=UTQf
-----END PGP SIGNATURE-----

Received on Monday, 16 February 2004 09:15:36 UTC