[CoFI] Re: Examples of concrete syntax for CASL basic specs

[Frederic.Voisin@lri.fr (Frederic Voisin)]

Dear all, three remarks on the examples sent by Peter (and thanks to him !)

[Thanks for the prompt response! --PDM]


1. The =! for existential equality really looks more like a typo than like
a symbol with a special meaning, not mentioning the poor of us who uses
!= in C for non equality !


2. I still do not like the = and > in the sort definition (as in ex. 1.4 SIG1
and 2.1 FILE). When followed by a similar symbol (like a < or { or ...) it is
not that easy to distinguish between the "concrete syntax" and the syntax of
the operation. See for example FILE in which you replace the = by a > ...

Why not "is" and "has" instead of "=" and ">" (OK it's verbose, my latine
origine surfaces :-))


3. the "end" in the "generated" clause does not look great when there
is only one sort. While the full form generated ... end may be needed for
the most general cases, maybe an abbreviation for this typical use
would help. Something like "sort Set[Elem} .... generated" to avoid any
misunderstanding of the scope of "generated" that could happen with
   generated sort S ... ;
             sort S' .... ;

[This last issue was also addressed by Peter in his comments of 20 Oct 97.
 He proposed to leave it open until after the discussion of structured
 specifications proposals (next week).  --PDM] 

frederic