[sc34wg3] Towards TMDM 3.0

Lars Marius Garshol larsga at garshol.priv.no
Mon Feb 23 04:08:48 EST 2009

* Rani Pinchuk
> Benjamin Bock, Robert Barta, Xuân Baldauf, and Lutz Maicher  
> presented in
> TMRA'2008 the presentation "A step towards TMDM 3.0" -
> http://www.slideshare.net/tmra/a-step-towards-tmdm-30-presentation

While I have some sympathy for the proposals in this presentation I  
don't think this is the time to start work on that. We've just  
finished TMDM and XTM 2.0 and are currently trying to finish CTM and  
TMCL. This is not a good time to stop work on CTM and TMCL to go back  
to revise TMDM and XTM 2.0.

> Alexander Mikhailian, Xuan Baldauf and myself contributed a paper  
> which
>  demonstrated a certain problem in the current TMDM (see
> http://www.tmra.de/2008/talks/pdf/121-135.pdf).
> In working on our Topic Maps engine, we find that we are unable to
> implement this part of the current TMDM.

I guess you are referring to section 4.2 of this paper? I'm afraid I  
don't see any problem in this section. Item identifiers (or subject  
identifiers, or subject locators) were never meant to be unique  
*across* topic maps, only *within* each topic map.

I think if you reflect on it for a moment, you will see that in fact  
no other policy is possible. Even your proposed change (just one [item  
identifier], new property [item origins]) doesn't solve this. It's  
entirely possible to create two topics in different topic maps with  
the same item identifier. For example, in different TM engines, or on  
different machines.

In any case, there is no problem to be solved here. Different topics  
in different topic maps can have the same subject identifier, which is  
how we can have interoperability between topic maps. Item identifiers  
are essentially the same.

--Lars M.

More information about the sc34wg3 mailing list