Re: Business Week (12/6/99 issue) article on Software Quality

[Richard D Riehle <laoXhai@ix.netcom.com>]

In article <384ffd52.888393602@newsnew.draper.com>,
	rracine@myremarq.com (Roger Racine) wrote:

>Why don't you look up the meaning in Robert's reference?  There is no
>difference.  A bug is a mistake that has not been diagnosed.  The
>cause is not known, but the effect is seen.  As if there were a bug in
>the system mucking things up.

Ah, but we see the term being used, instead, for mistakes that are
diagnosed, simply not yet corrected, as in "known bugs."  Further,
you have just acknowledged that a bug is just another word for
a mistake.  Thank you.

>Exactly.  We would have to say "I have an unknown problem in my
>program" instead of "I have a bug in my program."

And the problem with that is?

>"Bug Report", "Problem Report", they mean the same thing.  I agree
>that no one should trivialize problems where the cause is unknown.
>That happens too much no matter what term is used.  For one example,
>the first Space Shuttle flight was delayed a couple days due to a bug
>(at the time the cause was unknown).  It had been seen in simulations,
>but so seldom that it had not been resolved.  The Shuttle flew with
>the error (it had been diagnosed by then, and was known to be an
>initialization timing problem with a probability of about 2%).  
 
So we use "bug" to characterize some error that is not yet diagnosed.
We use "error" once we understand the cause.  If that were a consistent
usage, no problem.  It isn't.  The "bug" report is too often a list of
problems "we have not yet gotten around to fixing."  This is certainly
true of a little software company in the Northwest corner of the United
States that some of us know and love.

>>Once it has been so identified, it is no longer a "bug" it is an
>>engineering defect.  
>
>Agreed.  And what is wrong with designating it that way?

Nothing as long as it is accompanied by the admission of the error
that caused it.

>Electrical Engineers also use the term.  My guess is that software
>folks picked it up from them.

I'm not sure about who used it first.  It is all too often used in very
sloppy ways.  Robert cites its use by Edison, but Edison was an inventor,
not an engineer.  His work was similar to that of very creative 
programmers of our own time who, clever as they may be, are not thinking
of software as an engineering process.  

In my own programs, the "bugs" are mistakes I have made.  You 
may call it a bug.  I still call it an error.  I correct my errors. You
exterminate bugs.  The fact that I have not yet discovered my error does 
not make it less of an error. I prefer my own terminology.  Sorry for being
a "shrill and crackpot" over this.  

The real disappointment for me is that no one appreciated my reply to 
the Business Week article.  Also, I suspect that those non-software
people who read it, if it is published, will understand the point I am
making without being defensive about the terminology.  Only the software
practitioners are likely to object.  Interesting...

Richard Riehle