[tmql-wg] Result set requirements

Lars Marius Garshol larsga@ontopia.net
Tue, 16 Mar 2004 02:09:54 +0100

* Lars Marius Garshol
| tolog doesn't have explicit typing in the language, [...]

* Robert Barta
| I think it has explicit typing. All predicates have predefined
| profiles.

In that sense it does, but those defining inference rules do not need
to specify the types of the arguments. The engine infers those on its

| But I am not talking (writing actually) about types in the sense of
| the implementation (what kind of item in terms of TMDM), but typing
| in an application-specific sense.
| In XQuery there are not only those types of WXS (the predefined
| ones, number, integer, ....), but also this [ not real syntax ]
|    element Book := ( element Title, element Author )
| This 'typing' gives queries a boost but requires a strong mechanism
| to define the type. In the TM area this would be the task of an
| ontology language.

I've seen the XQuery folks do this, though I have to say I am not very
keen to follow in their tracks. This quickly gets very complicated,
and to make use of it requires an unusually advanced implementation,
plus close integration between QL and CL. There's a lot of pain down
this route; I'm not convinced there's a lot of gain.

To top it all, I'm not sure you need to integrate things as closely as
the XQuery/WXS folks have actually done.

Are you advocating this approach, or just describing what the XQuery
gang did?

| I do not see this coming with TMCL.

What do you mean?
* Lars Marius Garshol
| yet it is sufficiently amenable to analysis that the current
| tolog implementation in the OKS can infer the possible types of every
| variable used in a query.
* Robert Barta
| Without application-specific typing this is straightforward, yes.

Given certain constraints, anyway.

Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
GSM: +47 98 21 55 50                  <URL: http://www.garshol.priv.no >