[tmcl-wg] TMCL Comments
Robert Barta
rho at bond.edu.au
Wed Mar 28 23:40:48 EDT 2007
Hi,
These are mostly editorial comments/suggestions relative to the
document posted for the Oslo meetings. Post-mortem, so to say:
------------------------------------------------------------------
- Issue: how would TMCL harmonize with TMOL (TM Ontology Language,
to be)???
TMCL Schema <= TMCL Rule <= TMOL
- 1 Scope
... must exist in order ...
Is it also possible to constraint that some things do NOT exist?
- General
write "Topic Maps" for the technology, write topic map for one
particular map
- 3 Formal Semantics
I'm not too happy with the title, maybe split the TMDM and the TMQL
part?
- 4 TMCL
Title not too good. "Schema and Rule"?
last sentence a bit unclear
...constraints of these to kinds can be used simultaneously within one
schema
- 4.1
....used to constraint the TMDM .... -> to constrain TMDM instances
maybe change semantics to predicate only:
validate (m, s)
defined on all maps m and schemas s
- general:
When a URI is used, then maybe keep it on a separate line like this
http://somelonguriwhichmesses.up.line.breaks/
- 4.2.1 Metamodel
I would try to completely formalize this with CTM, much shorter
- 4.2.2
I would have prefered a bottom-up explanation:
A /constraint topic/ is a topic of type .....
A collection of constraint topics is a /schema/.
- 4.2.3 IsAbstract
This does not look as it would follow the structure of the others.
- 4.2.4
The TMQL semantics would look like this (@TYPES parameter)
not
some $TYPE in @TYPES, $TYPE' in @TYPES
satisfies // $TYPE == // $TYPE'
- 4.2.10
Section title Occurrence..., but following text
names /associationTypeScope/.
- 4.2.12
TMQL
every $o in // $OCCURRENCETYPE
satisfies $o >> atomify >> types == $DATATYPE
- 4.2.13
every $t in // $TYPE
satisfies
every $o in $t / $OCCURRENCETYPE
satisfies fn:regexp ($o, $REGEXP)
& at least $CARDMIN exist $t / $OCCURRENCYTYPE
& at most $CARDMAX exist $t / $OCCURRENCYTYPE
---------------------------------------------------------------
\rho
More information about the tmcl-wg
mailing list