[sc34wg3] Illustrating SIDPs

Patrick Durusau sc34wg3@isotopicmaps.org
Tue, 04 May 2004 14:30:03 -0400


Greetings!

In a conversation with Graham at XML Europe, he indicated that the
TMRM had no identity attribute for topics. I was too jet lagged at
that point to follow up properly so I am hopeful we can continue that
discussion on the list.

While I will be illustrating the notion of SIPDs as processing of a
topic map, note that the TMRM DOES NOT mandate the processing or rules
that I am using as an illustration.

It DOES provide the terminology and checklist needed by application
designers and topic map authors in order to allow them to disclose the
rules for identifying the subjects of topics -- the rules that apply
in their applications and/or topic maps.

The rules for a topic map has defined a "personName" property as
follows:

     * Property name:              "personName"

     * Value type:                 string

     * SIDP or OP?:                SIDP

     * Test for subject identity:  If the strings are identical, they
                                   represent the same subject.

So now we know that whenever two topics have a "personName" property
whose value is the same string, they are regarded as having the same 
subject.

The Topic Map Author (who happens, in this example, to be writing the
rules for the topic map) has created a topic T1 in which:

     T1:
       personName="Patrick"

(We know, therefore, that T1 will merge with every other topic that
has a personName property whose value is "Patrick".)

Now, it turns out, sure enough, that the Topic Map Author later
creates another topic, T3, with a different subject in mind than the
subject that T1 is intended to reify, but with the same name:
"Patrick".  After merging, the two Patricks have the same reifier; the
topic map makes them appear to be the same person.  In our little
vignette, here, the Topic Map Author detects the problem by
discovering that the merged Patrick has two spouses.

Now, our Topic Map Author may not be the sharpest tack in the box, but
he's not a complete idiot, either.  He reasonably says to himself,

"Oh, I can distinguish between people who have the same name by means
of their spouses.  If they have different spouses, even if they have
the same name, they must be different persons."

Changing hats, he becomes the rules maker, and redesigns the
rules.  He replaces the "personName" property with a "personID"
property, in which the value is a combination of name and spouse:

     * Property name:              "personID"

     * Value type:                 complex:
                                      "name"   : string
                                      "spouse" : topic

     * SIDP or OP?:                SIDP

     * Test for subject identity:  If the "name" strings are identical, 
and the "spouse" topic is the same, the subject is the same.

Changing hats again, he becomes the Topic Map Author, and
specifies:

     T1:
       personID:
         name:   "Patrick"
         spouse: T2

     T2:
       personID:
         name:   "Carol"
         spouse: T1


     T3:
       personID:
         name:   "Patrick"
         spouse: T4

     T4:
       personID:
         name:   "Brunnhilde"
         spouse: T3

...and T1 and T3 no longer merge, because Carol (T2) is not Brunnhilde
(T4).

Important points to draw from this illustration:

There is only one SIDP per topic (per TMA). (note that personName was
replaced by personalID)

SIDPs (and, for that matter, OPs) can be arbitrarily complex.

The tests for determining whether the values of two instances of the
same SIDP class will be regarded as identifying the same subject can
be arbitrarily complex.

Working on other posts to hopefully explore where we have missed each 
other in prior discussions. Not suggesting we will always agree, but 
hope we will all have a shot at understanding the basis for our 
disagreements in a technical sense.

Hope everyone is having a great day!

Patrick

-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
Patrick.Durusau@sbl-site.org
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model

Topic Maps: Human, not artificial, intelligence at work!