[sc34wg3] What do we mean by reification?

[Bernard Vatant]

Just to second Lars Marius on this seemingly endless debate.

I remember my first TopicMaps.Org F2F in Paris in January 2001. Among
other things, I remember "reification" was (re)-introduced in the
general sense it has in XTM 1.0 prose, and in current RM. This was done
in a small working group, including if I remember well Steve Pepper,
Chris Angus, Daniel Rivers-Moore and myself. At the time we were happy
with that word. Now I think we were absolutely wrong, and should have
used representation instead. 

If we want to be understood and interoperable outside our topic maps
universe, please let's look at what is going on next door, and be
consistent with how "reification" is used, for example in RDF and formal
logic.

See http://www-formal.stanford.edu/jmc/generality/node6.html

Note that this resource dates back to '96. 

It's exactly the way it is used in SAM, for representation inside the
model, by a first-order object, of a higher-order structure (predicate,
statement, association ...). It is a technical term, needed in the model
for internal consistency purposes. 

The general relationship between a topic and its subject is one of
"representation", in its most generic sense, like in "knowledge
representation". Reification is a technical case of representation, and
RM should use only "representation".

That seems now obvious to me ... although it was not 2 years ago :)

Bernard

PS: Still uncertain? Try Googlefight!

http://www.googlefight.com/


reification : 32 900 results

representation : 1 270 000 results

_____________________________________

Bernard Vatant
Senior Consultant - Knowledge Engineering
www.mondeca.com
_____________________________________