[sc34wg3] SAM-issue psi-generics (was: SAM-issue term-scope-def)

[Lars Marius Garshol]

(Again, please note thread change.)

* Lars Marius Garshol
|
| The worst example of this confusion in XTM 1.0 is probably the
| xtm:topic published subject. It seems to say that every topic is an
| instance of the class "topic", which sounds reasonable until you
| realize that it would mean that "tennis" and "Dutch" are instances of
| "topic". They are not; they are instances of "subject".

* Marc de Graauw
|
| I do not get this. Since the subjects "tennis" and "Dutch" are not
| the same as the topics "tennis" and "Dutch", it seems reasonable to
| claim that the topics "tennis" and "Dutch" are instances of class
| "topic", 

Yes, but instanceOf assignment is a statement about the subject
represented by the topic, not about the topic. (Just as with all other
characteristic assignments.)

So as long as topic "marc" does not represent a topic, and I say

  [marc : topic = "Marc de Graauw"
   @"http://www.marcdegraauw.com/aboutme_eng.htm"]

I would be claiming that you are a topic, right?

| So what is wrong with the xtm:topic published subject?  IMO it does
| not say that the subjects "tennis" and "Dutch" are instances of
| class "topic", only that that the topics "tennis" and "Dutch" are.

The wording of the subject indicator is:

  Topic: The core concept of topic; the generic class to which all
  topics belong unless otherwise specified.
 
Of course, this is not very precise, and so it is difficult to know
for sure what it eans. Had the topic/subject-distinction been
correctly preserved throughout I would have agreed with your
interpretation, but I read this as saying that the example below is
equivalent to the one above.

  [marc = "Marc de Graauw"
   @"http://www.marcdegraauw.com/aboutme_eng.htm"]

There's no way that can be right.

For more examples, look at the definition of scope. In 1.3 it seems to
consist of topics, while in 2.2.1.6 this is less clear (it might well
consist of subjects), in 3.1 it clearly consists of topics, while in
3.3.1 it is clearly said that it consists of subjects. So which is it?
This is all normative text...

Of course, part of the problem here is that the spec keeps repeating
the definition of scope, which makes this problem much more likely to
occur. 

For the record, I should add that the SAM may well contain similar
mistakes, as they are subtle and very easy to make, so bug reports are
much welcome.

-- 
Lars Marius Garshol, Ontopian         <URL: http://www.ontopia.net >
ISO SC34/WG3, OASIS GeoLang TC        <URL: http://www.garshol.priv.no >