Re: ::scr open sores

["Simon Kinahan" <simon@xxxxxxxxxxxxxxxxxxx>]

> Old Skool is dead.  Long live New Skool!

I think you're conflating two differents things ...

On the one hand, it seems to be widely accepted - and I agree - that Big-M
Methodologies that proceed smoothly from analysis, through design to
implementation and delivery work quite badly. Instead, short hops between
releases, incremental evolution of design, and feedback from users, seem to
produce software more reliably. Rather than proceding smoothly one after the
other, analysis, design, implementation and maintainance all go on
concurrently, and feed into one another. In this sense, open source is just
one aspect of a phenomenon in software development started shortly after
"The Mythical Man Month", and also includes XP, "Worse is Better", and
Microsoft's own development processes. Who said this business moved quickly
? The Methodology Snake Oil salesmen still entice many in their various
forms, but they are weakening, and the forces of Eris will triumph !!!!

... ahem, sorry, got a bit carried away there. Where was I ? Oh, yes, on the
other hand, rapid, incremental development is not a call to Just Let It All
Hang Out. Although analysis now procedes in parallel, or at least
interspersed, with the implementation, it doesn't mean you don't need to
think about what your software will do at all. You can't, for instance, sit
down and decide to write a programming language without knowing whether it
will have classes, whether it will be functional or imperative, whether it
will be strongly or weakly typed, and so on, because once you've make these
decisions, they tend to stick.

Perl, for instance, was initially quite carefully thought through by a
single mind. Its had various things - like declarative programming, and OO -
that don't really fit it very well grafted on, but these really just prove
my point. They don't fit well, and they don't look as if they fit well,
because they're in conflict with the rest of the language.

Gnome, on the other hand, had no conceptual vision at all, except to be a
bit like KDE, only freer. Whereas KDE was to be a bit like Windows, only
freer. As a result, the first versions were a random slinging together of
ill-fitted bits. Its only very recently that serious attempts to impose a
unifying model have been made, and these are (I predict) unlikely to work.
Similar things can be said anbout C++.

Simon