(Followup to part II)

Looking up assertions with match_forall works, but is cumbersome if you just want to find instances, types, subtypes or supertypes of a given toplet.

To experiment with this here, we need some topic map with just a few types, subtypes and some instances for these. Instead of creating them programmatically, we write the map in one of the convenient text notations and let a parser do the job.

For the notation we choose (the old and trusty) AsTMa:

The map features sacklpicka as an instance of cat and rho as an instance of human.

After loading a package supporting the notation, we instantiate our map object $tm. As we are too lazy to store the AsTMa text in a file, we use the option to specify AsTMa code inline. Still, we have to tell the map to read and deserialize the content now:

More about persistent maps later.

Taxonometrics

As you would expect, sacklpicka will be listed as instance of cat:

Since mammal does not have any direct instances, the following list will be empty:

To eventually reach also those instances which are instances of any of mammal's subtype(s), you will have to use the transitive version.

In a similar vein the TM package offers functions to learn about

• the subclasses, direct and indirect

• any superclasses, direct and indirect, and

• any types, also direct and indirect ones.

Boolean Versions

Since such taxonometric functions are pretty important, there is also a set of functions just for boolean tests:

Note how the tids function can accept lists.

Filter Versions

Finally, there are also variants to be used as filters for a certain class:

Sadly, at this stage the other filters are_types, are_supertypes and are_subtypes are not yet implemented.

Send more chocolate if you need them. Better than sending patches ...

Work supported by the Austrian Research Centers.