Question on waterfall methods for SW Development

[Alex Martelli]

"Rusty Williamson" <rwilliamson at unno.gers.com> wrote in message
news:dxVl5.120$MJ.1606 at typhoon.san.rr.com...
> Hi!
>
> Under what circumstances would a 'waterfall' methodology be better or
> preferable to an 'iterative' methodology?  Any information on this would
be
> greatly appreciated.

I've been developing software for over a quarter of a century
and still haven't found one case where 'waterfall' could possibly
be preferable in the real world.

In theory, waterfall (basically, one single iteration in an
iterative cycle, as I see it:-) is better when: requirements
are fully known in advance, and perfectly fixed; developers
entirely master the application domain, and all aspects of the
technologies to be used, having successfully used those techs
to develop many previous, very similar applications in this
same application domain; the schedule is not going to be cut
short, resources available will not suddenly change without
notice (e.g., 2 key developers eloping:-), the scale itself
of the project is well inside managers' and developers'
previous experiences, there are no real opportunities for
early partial-product delivery, etc, etc.

But the difference between theory and practice, in practice,
is larger than the difference between them in theory...


Alex