Runtime Safety of Programs

[Zuurbier, E. - AMSXE]

David,

Exception handling is certainly necessary in Clean, but currently lacking.
Division by zero can happen,
addressing non-existing array-cells is possible, the program can run out of
stack or heap, to mention but a
few nasty things.

One lead, which is currently followed by Nijmegen, is trying to prove that
your program will never divide by zero,
trying to prove that your program accesses only existing array-cells,
etcetera. Then you don't need exception handling.

The other lead is trying to find people who want to study exception handling
and incorporate it in Clean, without
compromising Clean's mathematical properties.

Regards,

Erik Zuurbier, KLM,
co-mail: SPL/XJ-831
phone: +31 20 649 6255
email: erik.zuurbier@klm.nl <mailto:erik.zuurbier@klm.nl> 


		-----Original Message-----
		From:	David McClain [mailto:dmcclain@azstarnet.com]
		Sent:	zaterdag 29 januari 2000 2:40
		To:	clean-list@cs.kun.nl
		Subject:	Fw: Runtime Safety of Programs


		 I have a question regarding the creation of runtime safe
programs...  I
		have
		 a great deal of experience with Lisp, OCaml, and SML, and I
am just now
		 learning Clean. All of the aforementioned languages have
the concept of
		 exceptions and exception trapping. Clean has the ability to
"abort" but I
		 find no mention of exception trapping. How is this done in
Clean, or am I
		 asking the wrong question here? Perhaps with graph
rewriting it is
		 unnecessary? Any thoughts on this topic would be
appreciated.

		 David McClain,
		 Sr Scientist
		 Raytheon Systems Co.
		 Tucson, AZ