🚀 go-pugleaf

schiex@cert.fr (Thomas Schiex) writes:

[examples illustrating the usefulness of laziness deleted]

>All these GOOD thing have too be payed in terms of time (small overhead, so
>small it is negligible). But the main issue is DEBUGGING which is very
>difficult. Sequencing is completly disturbed and the poor human is lost when
>tracing ! (obvious relations with concurrent processing). However some
>papers gives interesting debugging issues for lazy languages (for Daisy, see
>"Lisp and symbolic computations", in the first numbers).

Our group has been involved in research on algorithmic debugging for some time
and recently we have applied this technique to lazy functional languages with
some success. However, several difficult problems remain to be solved before
this can be regared as a practical approach.

In algorithmic debugging, the program to be debugged is first executed and a
complete execution trace tree is built at the procedure/function level. This
tree is then searched top-down for the procedure/function invocation that
caused the entire program to misbehave, and once this invocation is found it
is concluded that the bug is inside the corresponding procedure/function.
The user must aid the debugger in this search by answering questions regarding
whether or not invocations behaved correctly. Thus the user need not concern 
himself with operational aspects such as execution order but can concentrate
on the declarative semantics of the program instead.

I'm going to present a paper on this topic at PLILP'92 next week and I'll be
happy to e-mail copies (postscript) of it to anyone who is interested.



-- 

Henrik Nilsson (henni@ida.liu.se)
Department of Computer and Information Science
Linkoping University
S-581 83 Linkoping
Sweden
--

Henrik Nilsson (henni@ida.liu.se)
Department of Computer and Information Science
Linkoping University
S-581 83 Linkoping
Sweden