Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uwm.edu!cs.utexas.edu!tut.cis.ohio-state.edu!rutgers!mcnc!duke!romeo!crm
From: crm@romeo.cs.duke.edu (Charlie Martin)
Newsgroups: comp.software-eng
Subject: Re: "Correct" IS UNDEFINED!
Keywords: proof, correctness, theory, balderdash
Message-ID: <17232@duke.cs.duke.edu>
Date: 1 Feb 90 14:52:29 GMT
References: <7598@hubcap.clemson.edu> <9630012@hpirs.HP.COM> <17504@joshua.athertn.Atherton.COM>
Sender: news@duke.cs.duke.edu
Lines: 73

In article <17504@joshua.athertn.Atherton.COM> joshua@Atherton.COM (Flame Bait) writes:
>> = crm@cs.duke.edu (Charlie Martin) writes:
># = munck@chance.UUCP (Robert Munck) writes:

>
>> Oh, crap.  I'm sorry, but this is simply the single most assinine
>> argument that I've seen proposed.  Never in twenty years of software
>> engineering have I seen a system where the question "does it work?" was
>> undefined.  Sometimes it's POORLY defined, sometimes the definition
>> turns out to have contradictions, sometimes it turns out to be vacuous,
>> but all of them have some statement of the expected behavior.
>
>The second part of you paragraph suggests that the first part of it is
>just wishful thinking.  I have seen definitions of software behavior
>which are poor, vacuous, contradictions, and incorrect.  What I have
>not seen is a definition which was reasonably correct and which could
>be used as the input to a software verification system, even a human
>one.

It just suggests that I can understand the difference between Not
defined at all and badly defined.  Can you?  Or is this a straw man?

>
># I'm afraid that belief that we will someday be able to prove our
># largest, most important programs correct has become one of my tests of
># native intelligence, along with belief in astrology and support of the
># flag-burning amendment.
>
>I would consider it a better measure of optimism and lack of real world
>experiance.

This is what is called "dick measuring."  So, okay, whip it out.  Let's
see your CV/resume, and we'll compare it to mine.  We can start with
date of first professional programming job:  mine was 31 October 1969.

>
>#ifdef PERSONAL_ATTACK
>
>If you go back through the "programs can be proven" thread, I think you 
>will find that 80-90% of the people who post in defence of proving
>programs correct post from academic sites.  You will also see that 80-90% 
>of the people who attack this idea post from commercial sites.
>I do not think this is pure chance.  

Okay, we can play this game too.  I'll just note that a friend left
Atherton because she was tired of trying to sell or describe systems
based on programs where no one could be bothered to write designs or
specifications at all.
>
>#endif
>

Now that we've gone through the obligatory male-dominance nonsense, tell
me: what is it that you object to in technology that IN INDUSTRIAL USE
(see the cleanroom references) resulted in extremely low defect rates?
















stupid postnews program
Charlie Martin (crm@cs.duke.edu,mcnc!duke!crm)