Re: TMCL role combination question

Am 07.02.2011 um 10:37 schrieb Lars Marius Garshol:

> * Robert Cerny
>> 
>> Yet, it is still better to use distinct topic role constraints, because otherwise i cannot prohibit 'continent' from being a 'containee'. Because in the above model both topic role constraints (for container and containee) work on the same (abstract) super type A.
>> 
>> Correct?
> 
> Better than what? :)
> 
> The spec has:
> 
> contained-in isa tmcl:association-type;
> has-role(containee, 1, 1);
> has-role(container, 1, 1);
> role-combination(containee, city, container, province);
> role-combination(containee, province, container, country);
> role-combination(containee, country, container, continent).
> 
> That does prohibit 'continent' from being a 'containee'. It also prevents 'continent' from being the 'container' of anything but a 'country'. And so on.

But if:

abstract-region isa topic-type;
  is-abstract();
  plays-role(containee, contained-in, 0, 1);
  plays-role(container, contained-in, 0, *).

city ako abstract-region.
province ako abstract-region.
country ako abstract-region.
continent ako abstract-region.

How to i prevent continent to be contained in something? As far as i see, i can't. So it is better to put the 'plays-role' at the subtypes.

Robert