[sc34wg3] TMCL and MAX_INT

Lars Marius Garshol larsga at garshol.priv.no
Mon Feb 23 06:55:24 EST 2009

* Graham Moore
> If we go for card topics, how do we write the processing prose?

Same as now, except instead of finding the card occurrences on the  
constraint topic you have to step along an association to the card  

> If someone creates a topic '1-7' will other processors know how to
> deconstruct the 'mini-markup'?

Instead of

constraint-x isa tmcl:something
   tmcl:card-min: 1;
   tmcl:card-max: 7 .

the structure will be

constraint-x isa tmcl:something .
tmcl:has-cardinality(tmcl:constraint: constraint-x, tmcl:card: c-1-7)
c-1-7 isa tmcl:cardinality;
   tmcl:card-min: 1;
   tmcl:card-max: 7 .

So everything works as it does now except there's a jump to the  
cardinality topic.

> Isnt this kind of the same as just using string occurrences? But  
> with more plumbing.

It is almost the same, but the difference is that we can do:

c-0-inf isa tmcl:cardinality .
ph:photo isa tmcl:topic-type ;
   plays-role(ph:categorized, ph:in-category, c-0-inf) .

so that we need only one plays-role template, and we don't have to  
mess around with optional parameters and IFs etc.

> What I do like about this approach is that for 99% the basic provided
> card topics will be enough.

Yes. And for the last %1 it's not really much work to define the card  
topic you need.

> What I don't like is that it feels heavy to have topics for what are  
> very small 'inline' properties of a constraint.

True. That is the drawback.

--Lars M.

More information about the sc34wg3 mailing list