[sc34wg3] What do we mean by reification?

[Patrick Durusau]

Steve,

<snip>

> The only way you can make an assertion about anything in topic maps is 
> by creating a topic for it and then assigning characteristics to that 
> topic.
>
> Right?
>
> That's all Lars Marius is saying. Surely that's Topic Maps 101? (Or 
> has your head been completely turned by the RM ;-)

Nice of you to ask! Don't think so but self-reports on sanity or clarity 
are suspect at best. ;-)

It was mainly a rhetorical question to make sure we were starting from a 
common point.

>
> If you want to say something more about, say, the relationship 
> represented by an association, you have to create a topic for it.

Just as you created a topic for the subject outside the topic map, yes?

>
> There is a technique - fairly well documented since XTM 1.0 - of using 
> the source locator of a topic map object (e.g. the ID of an 
> <association> element) as the value of a <subjectIndictorRef> in a 
> <subjectIdentity> element. This is, in effect, saying that the subject 
> of the topic in question is the relationship represented by the 
> <assocation> element. SAM wants to call *this* process "reification" 
> and distinguish it from the general, run-of-the-mill act of creating a 
> topic.
>
> I think that makes sense. There is a very real need for a special term 
> that distinguishes these two cases and the SAM's usage *does* seem to 
> accord with general practice in AI.

I am not questioning whether the technique makes sense. What I am 
questioning is why the value of the <subjectIndicatorRef> being the ID 
of an <association> element differs from some other value of the 
<subjectIndicatorRef> such that it needs a different name? In other 
words, what do we gain by having a different name from the "general, 
run-of-the-mill act of creating a topic."? I think Lars' point that it 
be recognized by the software still obtains whatever the name.

>
> I believe we should adopt the SAM's usage and modify the way the term 
> is used in RM. (Does the RM actually even need it?)

Do you mean reification?

>
> Anyone that insists on going with the (more general) RM usage should 
> be prepared to suggest terms to use instead of 
> reification/reified/reifier/reify where SAM needs them.

I have (yet) to hear a compelling reason why the SAM needs to 
distinguish the two cases with different names. I think that case has to 
be made first before we start talking about who should be suggesting 
names for the SAM to use in this case.

Note that I agree with the facts you have stated about the operation of 
what the SAM calls "reification" so that part is not in dispute. What I 
am interested in learning is what difference there is in that technique 
and regular topic creation that merits a different name. As far as I can 
see they both have subjects about which the topic map creator wants to 
make a statement.

I am leaving for all day meetings at the Oriental Institute so I won't 
see any replies until late today (US Central time).

Patrick

-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
pdurusau@emory.edu
Co-Editor, ISO Reference Model for Topic Maps