[topicmapmail] Scoping: Associations or roles?

[Jan Algermissen]

"Thomas B. Passin" wrote:
> 
> Jan Algermissen wrote:
> 
> > "Thomas B. Passin" wrote:
> >
> >>I can.  Say you have an association representing the relationship of
> >>players to a sports team.  The athletes play roles in the association.
> >>Not only to they come and go at various times, but from time to time one
> >>player will assume the additional role of, say, team captain.  Thus the
> >>"team captain" role would be played by different athletes at different
> >>times.
> >>
> >>Without the ability to apply scopes to the roles, we are forced to some
> >>kind of roundabout construction, such as relating a time history to the
> >>team, even though what is really desired is to relate the players, not a
> >>history.
> >
> > an association represents a particular relationship and the subject
> > (the is-ness, the nature) of an association is completely determined
> > by the sum of all role playings. If the role playings change, the
> > relationship changes and thus you end up with a different association.
> > Scoping inside an association really makes no sense.
> >
> 
> So every time any single member of the team changes (say the goalie),
> you want to start a whole new association?  No way.  

You simply get a different association, there is nothing you can do
about it.

Consider this: How do you determine if two associations are equal
(represent the same relationship and thus are to merge)?


The team continues
> to exist, the role continnues to exist, and the role gets filled by
> someone else.  This is exactly like changing the content of a cell in a
> relational table.  You don't delete the row and replace it with an
> entirely different one, you just change the contents of the one field.
> This is no different.  You certainly do not have either a different team
> or a different role.

Hmm...sure the team is different after echanging a member (it could even
play better ;-). Why do you object this?

> 
> Incidentally, this point illustrates what I said in my last post, that
> roles are different from topic names and occurrences, because they are
> second-order characteristics.

Uh...and what is a "second-order characteristic"?


Jan
-- 
Jan Algermissen                           http://www.topicmapping.com
Consultant & Programmer	                  http://www.gooseworks.org