[mercury-users] Newbie problem. :)

[Peter Schachte]

On Thu, Jun 17, 1999 at 03:55:14PM +1000, Thomas Conway wrote:
> On Thu, Jun 17, 1999 at 02:45:50PM EST, Richard A. O'Keefe wrote:
> > In general, verifying that a negative rule is a consequence of
> > the positive rules may be impossible.  That's one reason why
> > you have to declare them.
> 
> Okay, so we're talking about two different, but related things,
> both of which are worthwhile:
> 	(i) checkable declarations of mutual exclusion.
> 	(ii) uncheckable assertions of mutual exclusion.

I think it might be better to distinguish

	(i)  statically checkable declarations
	(ii) dynamically checkable declarations

This view, I think, eases the task for language implementors, because
they can always decide not to implement static checking at all, and
just put in a compiler switch (similar to -DNDEBUG or its absence in
C) to turn on and off dynamic checking.  Generating code to check such
declarations at runtime shouldn't be too difficult (at least not
compared to the overall job of implementing Mercury :-).  On the other
hand, it leaves open the possibility of checking some of these
declarations at compile time, which is clearly preferable to the
extent it is possible.  Finally, it does not require the language
standard to establish the class of declarations that Mercury compilers
are required to check at compile time.

-- 
Peter Schachte                     History teaches us that men and nations
mailto:schachte at cs.mu.OZ.AU        behave wisely once they have exhausted
http://www.cs.mu.oz.au/~schachte/  all other alternatives.
PGP: finger schachte at 128.250.37.3      -- Abba Eban 
--------------------------------------------------------------------------
mercury-users mailing list
post:  mercury-users at cs.mu.oz.au
administrative address: owner-mercury-users at cs.mu.oz.au
unsubscribe: Address: mercury-users-request at cs.mu.oz.au Message: unsubscribe
subscribe:   Address: mercury-users-request at cs.mu.oz.au Message: subscribe
--------------------------------------------------------------------------