RE: [colorforth] abort

[Fréderic DUBOIS <frederic.dubois@xxxxxxxxxxxxx>]

> -----Message d'origine-----
> De : Mark Slicker [mailto:maslicke@xxxxxxxxxxx]
> Envoyé : mar. 10 juin 2003 20:55
> À : 'colorforth@xxxxxxxxxxxxxxxxxx'
> Objet : RE: [colorforth] abort
>> 
>> Proovability on programs seriously restricts your possibilities as a
>> programmer. You cannot cut corners and that's definitely not the
colorForth
>> style.
>
>I'm not sure that is true. Provability involves writing a specification
>and proving the program meets the specification. It is an open technique
>and can be applied to any existing program. Programmers already reason
>about correctness of their programs, proofing is just a more rigorus
>approach.
>
>> If such a thing is possible, it must be done at design level, and this
leads
>> us to the 'right by design' concept/ideal.
>
>I'm not sure what is meant by design. Specification is one half of
>provability. Provability is an apealing concept to me, yet it may be too
>much work to put into practice or may be redundant effort in many cases.
>
>Mark

For me, specification is what the program must do; design is how it will be
done.
Quality standards say that this first step from specification to design is
also subject to checkings.
It seems to me that your approach is specification -> implementation ->
proof whereas to me provability must be obtained by using some method to go
from design or specification to the implementation.

I've thought a while about this subject and I realized that it was for me an
attempt to replace or automate tests. It appeared to me that if I was not
satisfied with tests then maybe I should review my test methods first. What
about you? 

 Amicalement,
  Frederic




---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com