[ Home | Contents | Search | Post | Reply | Next | Previous | Up ]

Re: Methodologies

From: J
Remote Name:


I am sorely tempted to point out the original post contradicted itself, but that would be nasty, so I won't.

Interestingly, it could be argued that the ultimate methodology has already been discovered, at least as far as certain types of application are concerned. It's called the Open Source development model.

Before we get too bogged down in the question of whether this or that diagramming technique will help us write code, consider the fact that a bunch of random hackers, in contact almost exclusively via the Internet, have managed to concoct a full and high-quality implementation of Unix. Something tells me that drawing boxes was pretty low on these guys priority lists.

Before I get flamed (if only that were likely), I understand that "bought systems", to be deployed on corporate LANs and the like, are not exactly equivalent to an OS. BUT, the lesson that SHOULD (but won't, of course)be taken from this is that it is perfectly possible to build complex systems to an extremely high level of quality and stability without ever needing to bother with any of the following:-

A manager A "development methodology" A "life cycle" Arcane diagrammatical techniques etc, etc

Again, I think this observation provides a good basis for reasoning that methodologies per se are only actually useful in so far as they aid certain types of development model - the "Dilbert" model, if you like.

As soon as you expect any method to aid the process of conceptualising and solving the PROBLEM, you are mistaken. Linux shows us that it is not necessary to buttress pure thought with this activity - it is only useful as a means to exchange understood concepts, not to UNDERSTAND things.

Just my two quid (but I still want an answer!)

Last changed: October 05, 2001