[sc34wg3] RM: Mandatory role types?

[Steven R. Newcomb]

Steve Pepper <pepper@ontopia.net> writes:

> At 00:23 19.11.2002 +0100, Lars Marius Garshol wrote:
> >Steve Newcomb wrote:
> >| Here's the rationale (at least as I see it):
> >|
> >| (1) In the graph, all role types are mandatory because
> >|     without them, there's literally no way to tell
> >|     which role player is playing which role.
> >
> >Agreed, and this one on its own clinches it for me. Without roletypes
> >you have a meaningless assertion.

> Not meaningless, just incomplete. Why shouldn't it be
> possible to express incomplete assertions in topic
> maps? For example:

> (1) "A and B are related, but I don't know how" (no
>     association type or role types; role players
>     known)

The RM4TM idiom does not prevent you from saying this,
but it does require you to handle this situation by
defining an "unknown relationship" assertion type, with
role types "unknown role type 1" and "unknown role type
2".

> (2) "A is married, but I don't know who his wife is"
>     (association types and role types known; one role
>     player unknown)

The RM4TM idiom does not prevent you from saying this,
either, but it does require you to handle this
situation by defining a distinct "unknown person"
subject for each unknown role player.

> (3) "A and B are married, but I don't know who wears
>     the trousers" (association type and role players
>     known; role types unknown)

This is more like your example (1), above.  You can
certainly make these assertions; all you need is the
necessary assertion types.  (And, in XTM and HyTM, you
are free to define your own assertion types!)  What you
cannot do, at least under the draft RM4TM, is to extend
the semantics of an assertion type by making the
relationship between role players and role types
indeterminate.  In other words, if you don't know which
roles the role players are playing, you can't make an
assertion under RM4TM.

However, role types with unambiguously indeterminate
definitions are OK.  For example, you *can* define an
assertion type that has symmetrical role types,
"spouse.1" and "spouse.2", and that does not confer any
"wears the trousers" value on any property of either
spouse.

> Another example of (3), from the Italian Opera topic
> map:

>      "Puccini was the teacher of Ponchielli... Is was
>      it Ponchielli that taught Puccini? Damned if I
>      can remember!"

Right.  Same issue; same solution.

> Why should the RM, the SAM or the syntax prevent us
> from expressing these kinds of incomplete knowledge?

It *doesn't* prohibit you from saying anything you want
to say, but it *does* make you say what you mean.  That
seems a reasonable price to pay for information
interchange.

-- Steve

Steven R. Newcomb, Consultant
srn@coolheads.com

Coolheads Consulting
http://www.coolheads.com

voice: +1 972 359 8160
fax:   +1 972 359 0270

1527 Northaven Drive
Allen, Texas 75002-1648 USA