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

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

In article <82ochh$27p$1@nnrp1.deja.com>,
	Robert Dewar <robert_dewar@my-deja.com> wrote:

>You cannot change the language to be the way you want it to be.
>Sure, like Humpty-Dumpty in Alice, you can make words mean
>whatever you like, but if you want to be understood, you need
>to use words in a standard manner. The word "bug" has a securely
>established meaning (which incidentally is well described in
>the OED). We are not about to change the meaning radically
>because one person thinks it would be good to do so.

How then, Robert, do we distinguish between a mistake and a "bug?"
Is a bug simply some mystical entity that gobbles up tasty little
chunks of code, an arthopodic alimentary canal with no sense of
responsibility at either end?  Is a bug something we know is there,
but cannot yet identify?  Is a mistake a bug that we have identified?

What is so difficult about calling a bug an error, or at least a 
defect that originates in an error?   When we use the word bug, we
are suggesting that we have no idea what is wrong with our program.
I often hear it used to trivialize the presence of the so-called bug.

I agree that the term is in widespread use.  That does not make it
a correct choice.  I also agree that it is colorful and poetic. That
does not improve its descriptive properties -- does not make it a more
accurate representation of what is really going on.  

If we are really interested in promoting software practice to an 
engineering discipline, we will be more careful -- more precise --
in our terminology.  When there is a defect in a program, it should
be so identified.  When that defect originates in a mistake, it should
be so acknowledged.  Informally, "bug" continues to be an attractive
shorthand for some problem we don't understand.   From an engineering
perspective, we have a responsiblity to identify the causes of the
"bug" and give it a name that corresponds to the error that caused it.
Once it has been so identified, it is no longer a "bug" it is an
engineering defect.  

We recently had a police helicopter crash here in Silicon Valley. The
headlines this morning identify the problem in terms of some defect, not
as some kind of bug.  

Dennison takes me to task for being a "shrill and crackpot."  I hope
that is not true, but we shrills and crackpots rarely realize it when
we are correctly identified as such.  My point of view is simply that
no other branch of engineering, except software practice, consistently
uses the word "bug" to label its errors and defects.  It will always be
difficult to take software seriously as an engineering discipline as long
as its practitioners insist on placing the blame for its mistakes on 
some mystical creature.  It is a poetic appelation with little engineering
value.

>Personally I think Juliet had it right, and to paraphrase
>
>"a bug by any other name would stink as bad"

Again with the poetry. :-)  I love poetry, as you know, Robert.  
I have been trying, as I compose this, to think of a counter-quote
from Shakespeare -- there must be one -- but it does not come to 
me at the moment.  I'll probably think of it by the end of the day,
and it will then be too late.  :-)

Richard Riehle