Re: [Gnumed-devel] Software Engineering

[Karsten Hilbert]

On Thu, Dec 15, 2005 at 08:42:32PM -0200, Daniel Minahim wrote:

> Let's take small steps, from my point of view the GNUMed seems to have been 
> put together using what in SE is called 'rapid prototyping'. 
> 
> 1) You've established the wishlist
> 2) Gathered some readily available technology to fit it
> 3) Implement something with it
> 4) Check what went wrong and go back to 1
Sort of but not strictly following the above. Especially,
the peer discussion and literature-related modelling has
been left out.

> The ER diagram should have been ready before the SQL, it seems to me you guys 
> took a mixed top-down/bottom-up approach and may get tangled with the 
> details. The developer manual is focused mostly on describing what you have 
> already done, not HOW it should be done.
There are quite a few academic papers on how various sorts
of things should be done. I have never seen any of the
papers do anything.

> The DB is growing a lot and getting more complex and confusing right?
Not that I have noticed. Actually, it's gotten less
confusing over the last year. Not to mention the recent
refactoring into topical schemata. One thing that may make
it appear confusing is postgresql_autodoc's lack of
recognizing PostgreSQL table inheritance which would be a
worthy peripheral project for someone.

Karsten
-- 
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD  4537 78B9 A9F9 E407 1346