[sc34wg3] TMDM Conditional associations?

Lars Marius Garshol larsga at garshol.priv.no
Fri Jan 30 05:48:24 EST 2009

* Inge Henriksen
> Would it be an idea to implement the definition of conditional  
> associations in future versions of TMDM?

What you're describing here is generally known as inferencing or  
reasoning. There is actually a little bit of this in Topic Maps  
already, with superclass-subclass, as Graham described in his email to  

There has been talk of making an inferencing or rule standard in ISO,  
but this is a very complex subject, though, and it's not clear that  
it's ready for standardization just yet. More practical work and  
experimentation on this would be very welcome.

In any case, it's not necessary to change TMDM to support this. In  
your example you describe the inferred association as a normal TMDM  
association, and that's what it would be, except that it would never  
be input by the user directly, and might behave a little differently  
from associations which are.

--Lars M.

More information about the sc34wg3 mailing list