[sc34wg3] TNC backward compatibility

[Steve Pepper]

Hi Nikita,

Good questions!

At 20:28 04.03.2003 -0800, Nikita wrote:
>There are some issues related to the backward compatible TNC behavior.
>
>1)
>What happens when baseName is not typed?
>Is TNC turned on by default? I think YES.

I think NO. The more general case is when a name does NOT belong to
a controlled vocabulary. It is very natural to specify a type when
there IS a controlled vocabulary (type="iso639-alpha2"); much less
natural to do so when there is NOT a controlled vocabulary
(type="some-artificial-uncontrolled-vocabulary").

So the default should be that the TNC is not turned on by default.

My recollection is that we did in fact discuss this in Baltimore and
came to the same conclusion (which is reflected in the SAM).

>2)
>If base name's type is a controlled vocabulary,
>should the TNC be turned on along scope axis or controlled vocabulary?
>
>In other words, which of the following (if any) should be merged:
>
><baseName id="b1">
>     <instanceOf><topicRef xlink:href="#cv1"/></instanceOf>
>     <scope><topicRef xlink:href="#s1"/></scope>
>     <baseNameString>ABC</baseNameString>
></baseName>
><baseName id="b2">
>     <instanceOf><topicRef xlink:href="#cv1"/></instanceOf>
>     <scope><topicRef xlink:href="#s2"/></scope>
>     <baseNameString>ABC</baseNameString>
></baseName>
><baseName id="b3">
>     <instanceOf><topicRef xlink:href="#cv2"/></instanceOf>
>     <scope><topicRef xlink:href="#s2"/></scope>
>     <baseNameString>ABC</baseNameString>
></baseName>
><baseName id="b4">
>     <scope><topicRef xlink:href="#s2"/></scope>
>     <baseNameString>ABC</baseNameString>
></baseName>
><baseName id="b5">
>     <instanceOf><topicRef xlink:href="#cv2"/></instanceOf>
>     <baseNameString>ABC</baseNameString>
></baseName>
><baseName id="b6">
>     <baseNameString>ABC</baseNameString>
></baseName>
>
>I think that we had never discussed this.

That's true.

Assuming cv1 and cv2 are instances of "unique-characteristic" (see
http://www.isotopicmaps.org/sam/sam-model/#sect-name-types), my take
on this would be:

- b1 and b2 (possibly)
- b3 and b5 (possibly)

I think the final answer will depend on how we settle the scope
interpretation issues in SAM 3.4.3.

>3)
>Should TNC be turned on for occurrence elements if they are of controlled
>vocabulary type?

Yes. Except we should stop talking about "TNC" and find a better name for
what otherwise might be called "SIDP-based merging". Maybe "property-based
merging" is sufficient?

By the way, the phrase "controlled vocabulary type" is a bit constricting.
The PSI defined in the SAM has the following identifier:

   http://psi.topicmaps.org/sam/1.0/#unique-characteristic

so I suggest we talk about "unique characteristic types" instead.

Steve

--
Steve Pepper, Chief Executive Officer <pepper@ontopia.net>
Convenor, ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps)
Ontopia AS, Waldemar Thranes gt. 98, N-0175 Oslo, Norway.
http://www.ontopia.net/ phone: +47-23233080 GSM: +47-90827246