Footnotes

 

(1)

In fact Sig is a category, and Sen(.) and Mod(.) are functors. The categorial aspects of the semantics of CASL are emphasized in its formal semantics [CoF97e].  

(2)

The choice of a particular proof system for CASL has been investigated, but not yet decided.  

(3)

The translation of such constraints along signature morphisms adds a further component, for technical reasons.  

(4)

Mixfix notation is so-called because it generalizes infix, prefix, and postfix notation to allow arbitrary mixing of argument positions and identifier tokens.  

(5)

The occasional reference to the subsort and overloading relations in Part II may simply be ignored (or replaced by the identity relation) when the framework for basic specifications is restricted so as not to include these features.  

(6)

Declarations occurring anywhere in the enclosing list of basic items are taken into account when disambiguating the grouping of symbols in a term.  

(7)

It is unclear how to cope with the possibility of comments occurring everywhere using standard parser-generators, when the comments are not to be discarded by attached to abstract syntax trees. Should the positions where comments may occur be (severely) restricted, as foreseen for annotations? See the Tools Note T-7 [Sch99] with further analysis of the problem and for a report on how it has been solved in the INKA front end.