On 29/03/07, Russell Wallace <[EMAIL PROTECTED]> wrote:

I think there's at least one good practical reason to avoid doing that, or
at least to do it at arm's length in a "potential users discussing potential
features" mailing list rather than "here's our code as we write it". In the
early stages of something as bleeding-edge as this, it's normal to need
several rounds of scrapping and redoing major chunks of design; if you
don't/can't do that, if you have to go with whatever your first guess was,
it's easy to end up hamstrung later because the design doesn't really handle
the requirements and it's too late to rewrite from scratch. It's
psychologically a lot easier to do that sort of scrap-and-redo if the world
isn't looking over your shoulder.



The process of invention inevitably involves scrapping designs when they
reach a point where it's obvious that they're not going to work.  This is
especially a problem for AI systems, where even the theoretical basis
underlying the project is subject to uncertainty, whereas if you're just
writing a web browser all the theory of basically what it should do is
completely known from the outset.

I've lost count of the number of times which I've scrapped and re-written
some of my own projects, but by now I think I've made most of the mistakes
which its possible to make, and as they say "when you have eliminated the
impossible, whatever remains, however improbable, must be the truth".

-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?list_id=303

Reply via email to