[tmql-wg] TMQL: How to query TMRM?

[Robert Barta]

On Thu, Jan 25, 2007 at 01:49:18PM +0100, Lars Heuer wrote:
> I wonder if it is possible to query TMRM with TMQL? While TMQL uses
> aspects of TMRM it seems not to be possible to work on the 'raw'
> TMRM proxies. Is my suspiction right?

Yes, TMQL as it stands is "TMDMish".

> So we need Yet Another Topic Maps Query Language (YATMQL) which is a
> superset of TMQL to query TMRM?

Hmmm, good question. The question is then, in which way is TMQL
"TMDMish"? One obvious answer is 'in the various navigation axes',

   http://topicmaps.it.bond.edu.au/junk/tmql.html#Navigation

because it is them along which users move along the data model. This
may not sound like much dependency, but there is also the connection
with TM content which - written as CTM - will be very TMDMish. The
rest of the language is mechanics.

So my hipshot is:

  - incoming: navigation
  - outgoing: TM generation

If these would be replaced with TMRM concepts

  - incoming: TMRM navigation
  - outgoing: TMRM 'notation' such as

>        {(shoesize, 43), (beardcolor, white), (beardlength, verylong)}

then you might get closer to a TMRM/QL.

> The proxy neither has a subject identifier, nor subject locator nor
> item identifier.

Right, subject identification is covered as navigation. So either this
will be replaced with something else, or completely disappear.

[ Not quite, there is one shortcut in TMQL so that you can write

    http://en.wikipedia.org/wiki/U2

or

    wp:U2

to identify a subject without the longwinded

    "http://en.wikipedia.org/wiki/U2" ~

So this would have to go as well. ]

> It has just a label and its properties. Is it possible to fectch the
> proxy via TMQL and inspect its properties?

No, but this is where I would use my own set of navigation, probably
just stealing what is in TMRM at the moment.

\rho