(j3.2006) Generic interface question

[Malcolm Cohen]

>This illustrates precisely why I asked in 1986 for S8.99 (or so), and
>again at meeting 142, to include a definition of "entity."

I don't think so.

>Either we ought to stop using the term, or define it.

I do not agree.  There are many ordinary English worms we use that could be 
wilfully misinterpreted (and that if sufficiently perversely misread would 
result in contradictions).  Such misreadings are not plausible.  We are not 
about to embark on a quest to define the English language axiomatically.

> Malcolm wants

This is not about "Malcolm wants", and I'll thank you kindly not to pretend that 
it is.

>Acting on that advice,

I disagree that that characterises your misinterpretation.

> I concluded that where 16.2p1 says "entities with
>binding labels" are global global entities of the program, I thought it
>included interface bodies with binding labels.

Interface bodies *Do Not Have Binding Labels*.

I cannot imagine why you continue to assert this obviously false statement after 
your mistake has been pointed out to you.

Interface bodies with appropriate BIND(C) clauses declare procedures with 
binding labels.  An interface body does not have a binding label any more than 
it has rank, shape or type.  (That sentence brings forth the image of a 
rectangular pattern of logical interface bodies...)

>  After all, "everything"
>includes interface bodies, right?

Interface bodies are not entities *Of The Program*.

If one is going to stop reading sentences halfway through there is no possible 
change to the standard that could prevent misunderstanding.

Yes, this point is more subtle than the interface body one, but ignoring 
qualifications like "of the program" although easy to do on a quick read-through 
does not make a good basis for interpretation, especially when the result 
doesn't make sense.

>Could we [make a mess of what is already clear]

No.  Your misinterpretation draws on two false-to-fact conjectures, identified 
in my previous reply and again above.  Just the very first one is enough to 
settle the issue.  It does not depend on any subtle wording!

If one assumes things that are false it is not surprising if one subsequently 
draws an incorrect conclusion.

Cheers,
-- 
................................Malcolm Cohen, Nihon NAG, Tokyo.