[sc34wg3] TMCL: 4.4.1 Topic Type Constraint

[Robert Barta]

Hi,

I have worked through TMCL this morning and have many (!)
questions. Too many for one email. So this is against the draft

  http://kill.devc.at/system/files/tmcl.pdf

--

How is 4.4.1. to be understood? 

  "..only topics  ... defined as topic types can have instances"

What about association types, occurrence types and name types? They
definitely need to be "topics allowed to have instances".

So what it may mean:

  "Except association types, occurrence types, name types and one
   particular topic (which has been marked as topicType), no other
   topic in a map may have instances."

But that of course still would rule out a second topicType, right?

So what it probably should mean is:

  "Except association types, occurrence types, name types and a
   list of topics marked as topicType, no other topic can have
   instances."

Which implies that such a constraint has to parameterized with
a list, not a single topicType, right?

A TMQL expression would then look like this

  uniq ( // tm:topic >> types ) -- // tmcl:topicType == null

It finds all instances of topics (so not assocs, ....), finds their
type(s), makes everything unique, reduces the list by those things
which we defined as topicType and NOTHING (null) is allowed to be
left.

This assumes

  (a) that we adopt the meta-ontology of the TMDM->TMRM mapping which
      defines subject, topic, assoc, occurrence, name

      http://kill.devc.at/system/files/2007-11-20-tmrm.pdf (page 13)

  (b) the tmcl:topicType and the marking that some topics are instances
      of it are in the 'same' map. But TMCL stays silent about that.

      So maybe

       " TMQL expression will be evaluated on the merged .... yada
	yada ..."

Else:

- first sentence has two dots ..

- ad example: ...and [is] added ... as it [is] considered [to be] ...

- ... if it needed to be ... this is just cryptic

\rho