Royal Pain

[Peter L. Peres]

> > > Perfection is a waste of time...

Perfection is a waste of *all* the time, forever. Perfection is an
abstraction. If you wrote something and it appears perfect to you, then
someone else is going to have to find your bugs. This can be extended to
an arbitrarily large group of developers and customers. Proof is by
induction. Additionally, if the audience is large enough then there will
always be someone who will perceive a feature as a bug. The first step in
learning how to write code or engineer anything is to know your or your
processes error rate. Then you will know how large a project you can
manage, such that it will have a manageable number of bugs.

> Well, this is a cute thought, but it really hinges on your definition of
> 'perfection'. Is it code with no bugs? Or is it code that compiles correctly
> and just might be accepted by the customer? Is it code that meets every
> customer requirement? Is it code that can be upgraded in a modular fashion?
> And so on.
>
> In some applications you can let your customers find the obscure bugs but in
> others that's not acceptable.
>
> And to an engineer, perfection is attained by adding features and that has
> to be contained.

;-) You mean *bugs* are added with the features, such that the
features/bugs balance appears more appealing to the customer(s) ? I.e.
you can't hear the ghastly servo gear whining over the overly loud
elevator music ;-) ;-)

Peter
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml