Envelope-to: rlpowell@digitalkingdom.org
Delivery-date: Fri, 17 Dec 2010 07:12:38 -0800
Date: Fri, 17 Dec 2010 08:11:42 -0700
From: ".alyn.post." <alyn.post@lodockikumazvati.org>
To: jbovlaste@lojban.org
Subject: [jbovlaste] Re: defining "debug"
Message-ID: <20101217151142.GF32499@alice.local>
Mail-Followup-To: jbovlaste@lojban.org
References: <20101217133000.GA32499@alice.local> <AANLkTinZzGWyActPsETd=-PMGR+xRZhx_xZPDAy6QNJK@mail.gmail.com> <20101217142814.GC32499@alice.local> <4D0B7A8D.5060708@snell-pym.org.uk>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <4D0B7A8D.5060708@snell-pym.org.uk>
Sender: jbovlaste-bounce@lojban.org
Errors-to: jbovlaste-bounce@lojban.org
Precedence: bulk
Reply-to: jbovlaste@lojban.org
Content-Length: 1600
Lines: 45

On Fri, Dec 17, 2010 at 02:58:21PM +0000, Alaric Snell-Pym wrote:
> On 12/17/10 14:28, .alyn.post. wrote:
> > Is there a hint between these two at a more fundamental definition?
> >
> > If we have a reading fault finder and a computer program fault finder,
> > what is a fault finder?
> 
> I think there's a worthy distinction between fault *finding*, fault
> *diagnosis*, and fault *fixing*; "debug" in computer circles tends to
> refer to the latter two, and either users or QA/testers get the first
> job :-)
> 
> And all of these things can be specialised to computers with a bit of
> {sampla}.
> 

Good, yes.

I was thinking of debugging as the specific act of investigating a
program when the behavior is not matching one's expectation.

In that sense, I would be attempting to define fault diagnosis
according to the here presented lexicon.

Is fault fixing any different from the act of writing a program?
Can fault fixing and program writing have the same definition?

I can see that fault finding is worth having a working definition
for, at least so as to to compare any definition of fault diagnosis.
{cfisisku} most closely maps to the idea of fault finding.  In that
case, what would I use for diagnosis?  Something with troci seems
appealing in this case, "try conclude, try change, &c."  That begins
to fit nicely with Oren's suggestion.

Do we have several notions here, with cfisisku (or some variant)
expressing the idea of a QA/tester and {(rafsi "troci" *)}
expressing the idea of debugging/diagnosis?  Can you make a better
suggestion? 

-Alan
-- 
.i ko djuno fi le do sevzi