[sc34wg3] revised draft Reference Model document N0298

Steven R. Newcomb sc34wg3@isotopicmaps.org
11 Apr 2002 10:47:44 -0500


"Martin Bryan" <mtbryan@sgml.u-net.com> writes:

> I'm confused about Steve Newcomb's assertion that:
> 
> * Any node that serves as the R end of any CR arc:
> 
>    * MAY serve as the R end of any number of CR arcs, and
> 
>    * CANNOT serve as the A end of any AC or AP arc, and
> 
>    * CANNOT serve as the P end of any AP arc, and
> 
>    * CANNOT serve as the C end of any CR, Cx, or AC arc.
> 
>    The subject of such a node is always a role (or,
>    using the jargon established in HyTM, a "role
>    type").

> How does a HyTM occurs element that has both an
> implied occrl (as indicated by the name of the
> element) and a type attribute that indicates the
> Topic whose names/characterizes the topic type create
> a node that Cannot serve as the A node of any AC or
> AP arc? Surely there is nothing in ISO 13250 that
> stops a topic being used to name both a role and an
> assertion.

Caveat: I'm not sure what you mean when you speak of
        using a topic to name something.  When you say
        "topic", do you mean <topic> element, or "node
        in a dRM-conforming graph"?  For clarity, in
        the rest of this note, I'm adopting the
        convention that when I mean an instance of an
        SGML/XML element type, I surround its generic
        identifier with angle brackets.  When I say
        "topic" without angle brackets, I mean "node in
        a dRM-conforming graph".

Topics can *have* names, and a topic can *be* a name,
but, in the dRM, only an *assertion* can associate a
name with a topic.  Topics don't name anything.

Now I'll make some (possibly incorrect) assumptions
about the question you've just asked, and ask it again
in a way that I can answer:

Q: "How can the generic identifier of a HyTM <occrl>
    element be reflected in a dRM-conforming graph?"

A: The name specified via the generic identifier can
   become the subject of a topic (let's call it the
   "name topic").  By means of a naming assertion, the
   name topic's subject can be asserted to be a name of
   the topic whose subject is the occurrence role.

> Surely there is nothing in ISO 13250 that
> stops a topic being used to name both a role and an
> assertion.

I've repeated your last sentence, above, because I'm
still trying to understand it.  If you're saying that a
single name can certainly be the name of both a topic
whose subject is a role (i.e., an R-node) and of
another topic whose subject is an assertion (i.e., an
A-node), then I agree with you.  But that possibility
does not conflict with the dRM rules that you quote
above, and that I will now restate more compactly
below:

* A single subject cannot be both a role and an
  assertion type (no R-node can also be a P-node).

* A single subject cannot be both a role and an
  assertion (no R-node can also be an A-node).

* A single subject cannot be both an assertion and an
  assertion type (no A-node can also be a P-node).

Assertions, assertion types, and roles are three very
different kinds of things.  It's logically impossible
for a single subject to be any combination of them.
Their names, if any, are irrelevant.  Anybody can
assign any name to anything (or even, Orwellianly, to
everything).

> O298 states that "Each assertion asserts the
> existence of a strongly-typed relationship between
> some specific set of subjects of conversation" and
> "The ontologies of Applications may include an
> unbounded number of kinds of assertions."  Given I
> choose to use a topic whose name is "PublishedBy" as
> the type of an occurrence element (so that I can have
> documents published by people who are not defined as
> topics) why cannot I also have an association that
> asserts, where books are published by people
> mentioned in the topic map, that the same topic
> serves as the name of an association? 

It is perfectly OK for any number of subjects to be
asserted to have the same name.  (If we're using the
naming assertion type defined by the SAM, which imposes
the name-based merging constraint, then we'd better use
different scopes for each of the naming assertions,
lest these different subjects be incorrectly merged!
But that's a SAM issue, not a dRM issue.)

> Would this latter case have to be a role or would it
> be, as I feel it should, an assertion?

I guess I'm not sure what the antecedent of "latter
case" is.

-- 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