[Attempto] ace2owl. ace-variable as rdf-nodeID

[Kaarel Kaljurand]

Hi,

On Sat, Aug 22, 2009 at 13:55, Alex Shkotin <alex.shkotin at gmail.com> wrote:
>
> It would be very nice to use ace-variable in rdf-nodeID.

Do you mean the ACE variable or the DRS variable? E.g.
for the ACE text

   There is a man X.

which gets the DRS

   [A]
   object(A, man, countable, na, eq, 1)-1/4

The ACE variable is "X" and the DRS variable is "A". Note that
the ACE variable is not preserved in the DRS. This "A" is just a pretty-printed
Prolog variable.

The only function of ACE variables is to help to
formulate certain (complex) anaphoric references. In case you need to
communicate a name of an individual to the machine, rather use
a proper name.

Like most of the ACE tools, the ACE->OWL converter only "sees" the DRS
i.e. it has no idea what the original ACE variables were. (Although there
might be a way to recover them using token IDs.)

> It gives possibility to see easy in ontology which anonymous individual
> represents which variable.

Some time ago the anonymous individuals where "named" by the corresponding
DRS variable. But because at the time the OWL 2 (or then 1.1) syntax nor
the OWL-API supported this very well we started generating global
unique IDs. Once the new OWL 2 becomes a standard and OWL-API 3 comes
out we'll probably go back to the original solution as there is no need
to spend time on generating globally unique names for anonymous individuals.
Then it will become easier to relate the generated OWL to the DRS as more
of the identifiers are similar. However, the similarity between OWL
and ACE will remain the same.

Sorry for the late answer,
kaarel