# [sc34wg3] TMCL: 4.4.6

Robert Barta rho at devc.at
Fri Feb 15 04:27:41 EST 2008

On Thu, Feb 14, 2008 at 01:37:38PM -0000, Graham Moore wrote:
>
> Agreed, a bit messy. The intent, is as you say, that there can be no
> direct instances of vehicle.
>
> We really need the notion of direct and indirect instances.

> Coming down on one side or other will lead to issues. i.e. if we go
> for, validation only occurs on the 'raw' tmdm before the entailment
> rules are applied it will resolve this issue but likely cause pain
> in many other situations.

Well, formally we do not have the "graph-view, raw TMDM" view. Yet.

> One options is that we have axis for ALL_INSTANCES and
> DIRECT_INSTANCES in tmql, and tmcl can simply use them to express
> what it needs. These axis are considered to be dynamic run time
> axis. Its just a view onto the instance. This approach would then
> work for other inferred associations.

I am not in favour to squeeze more into TMQL what actually should be
handled elsewhere (like basic types, axes, CTMishness).

One thing we discussed in Leipzig is that larsbot tries to finish the
TMDM -> TMRM mapping and that in there you can control how

http://psi.topicmaps.org/iso13250/model/supertype-subtype

is mapped to TMRM:isa and how that satisfies the type-transitivity, or
how (b) it maps to something else which does not show the
transitivity.

TMQL would not be touched and is completely oblivious whether it
operates on the original TMDM, or the TMDM-light. That is just
provided by the environment.

\rho

> -----Original Message-----
> From: sc34wg3-bounces at isotopicmaps.org
> [mailto:sc34wg3-bounces at isotopicmaps.org] On Behalf Of Robert Barta
> Sent: 14 February 2008 13:12
> To: Discussion of ISO/IEC 13250 Topic Maps
> Subject: [sc34wg3] TMCL: 4.4.6
>
> Relative to
>
>    http://kill.devc.at/system/files/tmcl.pdf
>
> 4.4.6 AbstractTopicType
>
> If I say
>
>    vehicle isa tmcl:topicType
>    isAbstract (vehicle) .
>
>    car iko vehicle .
>
>    my-mama-car isa car .
>
> then my mama's car is also an instance of vehicle. At least this is
> what TMDM says. Not in
>
>   http://www.isotopicmaps.org/sam/sam-model/#sect-types
>
> which remain silent on that part, but
>
>   http://www.isotopicmaps.org/sam/sam-model/#sect-subtypes
>
> in the 2nd NOTE is quite explicit.
>
> What the constraint should say is that there are no "direct"
> instances, but that concept does not exist in Topic Maps. Not
> officially. Like in RDFS direct and indirect instances are
> indistinguishable.
>
> We discussed this in Leipzig in the context of TMQL/TMCL.
>
> Not sure how to do this cleanly. Larsbot promised to think about it
> :-)
>
> \rho
>
> --
> Austrian Research Centers, Environmental Monitoring Systems
> http://www.smart-systems.at/rd/rd_environment_en.html
>
>
> _______________________________________________
> sc34wg3 mailing list
> sc34wg3 at isotopicmaps.org
> http://www.isotopicmaps.org/mailman/listinfo/sc34wg3
>

--
And then he said: "You should read my blog." http://kill.devc.at/