[tm-pubsubj-comment] Comments on ontopia strawman - part 1

[Bernard Vatant <bernard.vatant@mondeca.com>]

Steve

Many comments on your strawman, so I've split the thread.
This part deals with general points and structure and PSI "typing"
I will address metadata later on in a separate thread.

> GENERAL POINTS
> --------------
>
> The first point to note is that I have chosen to express my subject
> indicators as topics in a topic map. I believe there will be substantial
> benefits to using a machine processable syntax and, in particular, topic
> maps, for published subjects. Although we have agreed not to REQUIRE any
> particular syntax, I think we should RECOMMEND using XTM and provide
> examples of how to do so.

> * Issue: What are the arguments for using a machine processable syntax?
> * Issue: Should we recommend use of XTM?

Issue 1. One argument for using a machine processable syntax: more and more topic maps
will be built through automatic process, that could include automatic attachment of PSIs
from a definite PSIDocSet. This seems only possible if the PSIDocSet as a whole is
machine-processable.

Issue 2. I would reword your last sentence, to make it more conformant to what has been
agreed upon in the TC, or at least the way I understand it: "We should recommend XTM as a
possible syntax, and provide exemples showing how to use it for PSI Doc Sets."
I agree basically with Thomas arguments, that XTM should not be *the* recommended syntax.
But I agree with you that we need people making proposals in other relevant syntaxes, RDF
and XHTML.

> However, in recognition of the fact that topic map software is not yet
> ubiquitous, I have also created a (machine generated) HTML rendition of
> the subject indicators, available at
>
>    http://psi.ontopia.net/ontopia/
>
> If we decide to recommend using XTM or other machine processable
> syntaxes, I think we should also recommend this approach.

Definitely. The PS Indicators for human use should show off as plain HTML documents.

> OVERVIEW OF THE PSI SET TOPIC MAP
> ---------------------------------

> * Issue: If we use topics as PSIs, should it be possible to distinguish
> topics that are *intended* to be used as PSIs from other topics (in the
> same map) that are not?

> I think it should. The mechanism I have used is a self-referential
> subject identity.

<snip/>

> I'm not sure this is a good solution for distinguishing between PSI
> topics and non-PSI topics, but I got it for free, because I am using
> that subject indicator reference to specify the identifier of the
> published subject. (See the subhead "IDENTIFIER" in the next section.)

Seems convoluted to me. Declaration of type looks simpler and possible (see below)

> I'm not sure how else we might differentiate PSI topics from non-PSI
> topics. Creating a PSI for PSIs and making the PSI topics instances of
> this type is not an alternative, because that would amount to saying
> that, say, (the company) Ontopia is an instance of the class Published
> Subject Indicator, which it obviously is not.

Nope. This is IMO clearly a topic type mismatch. The topic in the PSI DocSet should not
represent Ontopia company itself, but the *PSI for Ontopia*, which is a different subject.
And in fact you declare it somehow that way in your comment:

<!--  This is the topic that is the subject indicator for Ontopia  -->

If we have a generic PSI for the class PSI, at e.g.
http://psi.oasis-open.org/pubsubj/psi.html

Therefore every topic used as psi could indeed be declared like <instanceOf> the above,
providing you make clear that your topic represents the (addressable) subject *PSI for
Ontopia*, not the (non-addressable) subject *Ontopia* itself. If the topic element
represents the PSI, its subject is addressable, and the auto-reference - which I have no
objection to - should therefore be asserted by a <resourceRef> element. Note that I
changed the id wording to stress the point.

<topic id="psi-ontopia">
<instanceOf>
<subjectIndicatorRef xlink:href="http://psi.oasis-open.org/pubsubj/psi.html"/>
</instanceOf>
<subjectIdentity>
<resourceRef xlink:href="http://psi.ontopia.net/ontopia/ontopia.xtm#ontopia" />
</subjectIdentity>
 ...
</topic>

Bernard