[sc34wg3] A few comments on the CTM draft (Was: New CTM draft for Leipzig)

Andreas Sewe sewe at rbg.informatik.tu-darmstadt.de
Mon Oct 2 05:59:33 EDT 2006


I didn't expect my mail to end up on the list without moderation; I hope
the suggestions were welcome nevertheless.

At any rate, I just realized that one of my suggestions is not as
straight-forward as I made it sound:

I wrote:
> The hierarchical syntax of Example 24, while slightly more useful than
> the '*'-syntax of Example 25, could be emulated as follows:
> 
>   person {
>     supertype-subtype:
>       musician {
>         supertype-subtype: composer, conductor;
>       },
>       writer {
>         supertype-subtype: librettist, playwright;
>       };
>   }

Unfortunately this syntax proposal is, in a sense, ambiguous; when one
unravels the supertype-subtype assertion's value list as suggested by
section 4.6.2, Note 8, it becomes the following:

   person {
     supertype-subtype: musician {
         supertype-subtype: composer;
         supertype-subtype: conductor;
     };
     supertype-subtype: writer {
         supertype-subtype: librettist;
         supertype-subtype: playwright;
     };
   }

This was my proposal's intention. But since a value list is interpreted,
as per my suggestion, as a list of subjects, and an assertion block, as
per section 4.6.1.1, applies to the entire subject list, the following
will be interpreted contrary to author expectations:

   fortytwo {
     divisible-by: seven, eight {
       divisible-by: two, four;
     };
   }

Here the inner assertion block is supposed to have only eight as its
subject, not seven. (Being a prime, seven does not include an assertion
block of its own.)

But this is inconsistent with the interpretation of subject lists:

   seven, eight {
     divisible-by: two, four;
   }

Unfortunately I can't think of a way to resolve this inconsistency in
interpretation (of commas) in a satisfactory manner -- but maybe you can
at the Leipzig meeting. ;-)

Regards,

Andreas Sewe


More information about the sc34wg3 mailing list