Re: [xtm-wg] ANNOUNCE: Update of XTM Repository (new XTM prototype DTD)

["Steven R. Newcomb" <srn@coolheads.com>]

[Murray:]
> [Preface: I may be slightly wrong in some of these answers, but essentially
> I'm hoping to describe what I believe we have in xtm07.dtd. The sample
> instance xtm07.xml might help a bit, although I don't know if Steve, Michel
> and Sam have looked that hard at the sample. Perhaps one of them can chime
> in on my weaknesses...]

I'm just going to quibble a little.

[Kal Ahmed:]
> > ...I am assuming that only
> > baseNameString's are subject to the topic naming constraint and that this
> > would be described more clearly in the accompanying prose text.

Right.

> > I think, however that it might be useful to define two new PSIs for
> > 'displayable' and 'sortable' to use as <parameter>'s to partition the
> > variantNames in some consistent manner.

You read our minds, or we read yours.

> > 3) Identity
> > Where did it go ? Perhaps I'm missing something in my reading of the spec,
> > but it seems that I have no way to declare the subject of a topic. This is a
> > major concern
> 
> The <instanceOf> child element in <topic> says what the subject of the
> topic is: 
> 
>    <topic>
>      <instanceOf><topicRef xlink:href="#us-railroad"/></instanceOf>
>      <baseName>
>        <baseNameString>Union Pacific</baseNameString>
>      </baseName>
>      ...
>    </topic>
> 
>  could be translated to:
> 
>    "this topic's subject is an instance of 'U.S. Railroad'."

No.  <instanceof> establishes a class-instance relationship between
the topic node corresponding to the containing <topic> element
(playing the role of "instance"), and the topic node corresponding to
the <topic> element that plays the "class" role in the same
relationship.  That has nothing to do with the subject
identity of the <topic>.  


Whatever happened to the -identity- attribute of <topic> in 13250?
------------------------------------------------------------------

First of all, we decided that, as a practical matter, it's hard to see
any reason why a topic should not regard its subject descriptor as an
occurrence.  What, after all, could possibly be more relevant to the
subject of the topic, than some sort of hard-to-misinterpret statement
of what the subject is?  So, isn't the difference really a defined
public subject that is included in the scope of such occurrences?
Well, yes, of course it is, but that should only govern the
occurrence-related semantics -- basically, to help the user find the
occurrence(s) that are regarded as the subject descriptors of the
subject of the topic.  Anyway, in the graph, the presence of a special
XTM-defined public subject -- conveying, among other things, the
notion of "subject-descriptor-ness" -- in the scope of an occurrence
indicates that the occurrence is a subject descriptor.

But there is a very real sense in which a subject descriptor applies
to a topic node in the broadest, most unlimited scope possible.  Each
subject descriptor is supposed to be absolutely self-sufficient: fully
adequate to the task of identifying the subject unambiguously.  True,
it may only be capable of identifying the subject to those who can
read Mandarin Chinese, but it is nonetheless capable of fully and
unambiguously conveying the whole subject.  A subject descriptor's
special relationship to its topic node is not limited by scope.  It is
not limited at all; it's really its identity in every sense.  Every
subject descriptor is a descriptor of the whole topic node.
Therefore, the distinction of being a subject descriptor is not
conferred by adding a <topicRef> to a scope; the notion of scope does
not apply here.  The distinction is conferred by using a <topicRef>,
rather than a <resourceRef>, in the <resource> that specifies the
information as an occurrence.

So here's what happened to 13250's -identity- attribute, according to
me.

* To specify the subject descriptor of a topic, make the subject
  descriptor an occurrence of that topic.  However, instead of using
  <resourceRef> to point at the subject descriptor, use <topicRef>
  instead.

* If the subject descriptor *describes* the subject, set the
  -referent- attribute of the <topicRef> to "describesSubject".
  Alternatively, if the subject descriptor *is* the subject, set the
  -referent- attribute of the <topicRef> to "isSubject".

When the topic map graph is constructed: 

* the subject descriptor becomes an occurrence of the topic, and the
  scope within which it is an occurrence includes one of two special
  XTM-defined topics:

  * the one that means that the resource is a subject descriptor that
    describes the subject, or

  * the one the means that the resource is a subject descriptor that
    is itself the subject.

* all <topic>s that have the same subject descriptor in the same sense
  (i.e., either *describesSubject* or *isSubject*), will be merged;
  by definition they have the same subject.

-Steve

--
Steven R. Newcomb, Consultant
srn@coolheads.com

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

405 Flagler Court
Allen, Texas 75013-2821 USA

-------------------------- eGroups Sponsor -------------------------~-~>
eGroups eLerts
It's Easy. It's Fun. Best of All, it's Free!
http://click.egroups.com/1/9698/1/_/337252/_/973655661/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com