RE: UML/XMI and semantic web

Eric Neumann wrote:

> Although I am coming at this with a distinct life sciences 
> bias, a group
> of us (the Bio-Ontologies Consortium) did evaluate (by feature) a
> variety of languages for describing ontologies for the life sciences.
> The results are published in
> http://www.ai.sri.com/~pkarp/pubs/00-ismb.pdf. In summary, the group
> found that UML was not as expressive as some of the other alternatives
> (XOL, OML).

For what it's worth, I think the evaluation of UML in that paper is
flawed.  Table 2 has "No" as the answer to a number of questions
about the expressive capabilities of UML.  Seven of these answers are
just plain wrong, and the others are also wrong if you allow for the
use of annotations using the object constraint language.  I'm happy to
provide evidence for this claim, but it's a little off-topic for this
list, so if anyone wants more detail, please mail me.

I'm also surprised that UML scored a "low" on the "simplicity to
humans" scale compared to the other models considered, but this
probably depends more on an individual's familiarity with
object-oriented modelling than on anything else.

It may still be the case that UML is not well suited to representing
ontologies for biological applications (I'm not qualified to comment),
but I don't think it follows from the analysis mentioned above.

- Stephen

----------------------------------------------------------------------
Stephen Cranefield                       
Department of Information Science 
University of Otago                               Phone: 64 3 479 8083
PO Box 56, Dunedin		                    Fax:   64 3 479 8311
New Zealand	               E-mail: scranefield@infoscience.otago.ac.nz
----------------------------------------------------------------------

Received on Tuesday, 13 November 2001 21:03:33 UTC