On Wed, 13 Mar 2002, Elaine -HFB- Ashton wrote:

> Greg London [[EMAIL PROTECTED]] quoth:
> *>
> *>blaming the current language is convient.
> *>"Perl encourages bad coding practices."
>
> Perl is easy to pick on because it's so often true that people have mounds
> of crappy, ugly, shoddy, quickly slapped together perl code to point at.
> I'm sure bioinformatics is a veritable treasure trove of such gems.
>
> This is endemic to the language so people either have to live with quickie
> ugly perl that will render itself unmaintainable after a while or find
> some way of defining standard coding practices for all projects including
> perl.

I guess that's the price to pay for the speed of development you can get
in Perl. I would think that the cycle development -> maintenance ->
re-develop has to be quicker in Perl than in other languages, unless you
take special measures (coding standards, peer reviews, XP...).

As for the fact that software engineering is not as advanced as other
branches of engineering, you have to take into account that when you write
software you just write the equivalent of the blueprints in other areas.
So you don't have the cool-off period that corresponds to setting up the
assembly line and getting feed-back, you just release the product and
users find the bugs. That plus other disciplines tend to be more
evolutionnary while software design is often revolutionary.

And to talk about a domain I know a bit, aircraft manufacturing might look
mature, but believe me, aircraft design is a huge mess and bugs/awfull
mistakes and such do happen quite regularly, from clogged toilets to ABS
that doesn't always work, to on-board systems that fail way to often.
Aircrafts fly only because extra special care and redundancy are used for
critical parts of the system.

I believe tire manufacturing for SUV's also has had problems in the past.

In short software is complex, hence it is failure-prone, just like any
other system design.

Which does not give you much hope for those late night !@#$%^&* incidents
;--)

Michel Rodriguez
Perl & XML
http://www.xmltwig.com


Reply via email to