[sc34wg3] Supporting variants in TMCL

[Lars Heuer]

Hi all,

The paper is not a proof that we need variants (section 4.3 describes
a model without variants), it describes another model "containers of
names" where each "container" represents a group of names which have
somehow a relation to each other.

Unfortunately, the paper does not elaborate in detail the differences
to TMDM scope and I wonder why Xuân stops at topic names and does not
propose to make occurrences also "culture-neutral": If a topic map
author says, IBM has the website "http://www.ibm.com/de/", the topic
map wouldn't be culture-neutral since another user may prefer the
website "http://www.ibm.com/nl/" or "http://www.ibm.com/fr/". If we
think the the idea of containers consequently to the end, we need
also "containers of occurrences".

And don't we have a model that supports these containers already? I
think TMRM supports Xuân's proposal and the containers of occurrences
directly:

  name#1 := {name: "Constantinople",
             name: "Konstantinopolis",
             name: "Konstantiniyye"}

Anyway, I don't want to discuss the paper in detail here since I think
neither the name-containers nor the occurrence-containers have a
connection to the discussion how we can support variants in TMCL
directly (note that *TMCL supports variants*, it simply does not
provide a CTM notation / a constraint for it).

If someone wants that TMCL supports variants directly, he should
follow Lars Marius' advice and come up with a solution rather than
pointing out how useful variants are.

Best regards,
Lars
-- 
Semagia 
<http://www.semagia.com>