[topicmapmail] Wiki/WebLog application built on top of topicmaps...

[Guy Murphy]

[snip]
> With this approach, you could distinguish management metadata from
> user-created metadata with ease - just scope the association.  As for the
> case of [Bob]--(friend)--[Mary], is that not intrinsically a symmetrical
> assocation?
[snip]

As you present it yes, it's symetrical. But as in the case of unrequited
love, Bob may love Mary from afar, but Mary may not even know (or care) who
Bob is.... we may not ever know what role Mary plays in the association, we
may be ignorant of Mary. Mary may remain forever an enigma.

Bob may wish to assert frienship toward Mary, but we might wish to wait for
another time for Mary to assert her participation (if any) in the
association... this doesn't dimish the potential usefulness of Bob's initial
assertion.

So today we might find that Bob asserts... [Bob] --(friend)--> [Mary]
And tomorrow we might find that Mary asserts
[Mary] --(seeking-restraining-order)--> [Bob]
Only to find out the day after that [Bob]--(diagnosed)-->[sociopath]

We may care to model [sociopath]--(is-a)-->[Bob] or we may decided that we
only want to track interesting instances, or which Bob isn't one... our
choice.... remember that I'm talking about building a navigation system, not
a graph model for an inference engine.

A friendly regard doesn't necessarily run both ways. And we don't
necessarily kow the role that all participants play in an association at teh
same time, and that doesn't invalidate the potential utility of the
association.

These are silly, but useful examples... and while yes they can all be
accomodated with a TNM, it would seem that one has to go to considerable
effort to simply exhibit a non-regard on the part of a topic that's the
subject of an assertion rather than simply not having the reciprocal
assertion.

We could model this ignorance/lack of concern or whatever bidirectionally as
[Bob]--(unknown : friend)--[Mary], but it would seem to me that what we are
effectively trying to do is model a one way assertion, and that the best way
to model a directional arc is indeed with a directional arc.

I'll consider it overnight, and see if I'm merely sat in my preconception of
the issue. The discussion is useful for me as it's certainly flagging lots
of ways in which I can express interest in one direction with a
bi-directional assertion.

Cheers,
    Guy.