Re: Will the UML ever become truly universal?


"Phlip" <phlip_cpp@xxxxxxxxx> wrote in message
news:YT8oe.16030$iA6.10896@xxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>
> UML is neither universal nor interoperable in real life. It is both in
> _books_.
>
We agree, Phlip. There are lots of things that appear in books that
fail to correspond the the reality of actual projects.

Every so often, some new fad for software practice appears. I have been
in this field since the 1960's, starting with assembler and slogging my way
through many different aproaches to taming the software devils. Along the
way, we learn to be skeptical (but not cynical), and eventually adopt those
aspects of a new approach that is applicable to the kind of software we
are building.

The problem is that, every new fad, from every new guru, is taken as a
revelation from deity and any criticism of it is regarded as the ravings of
the lunatic fringe. The fact is that these new methods, whether they be
Agile, XP, RUP, OOP, SASD, Jackson Development Method, JAD,
formal methods, generative programming, aspect-oriented programming,
or some wondrous new programming language, all have their limitations.

Every proposed solution introduces a new set of problems. None of
the latest fads is any exception to this rule. Failure to recognize those
problems will lead those who adopt the solutions to create software
that is just as problematic as the software built with the methods they
have summarily rejected.

Richard Riehle


.

Relevant Pages