gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnumed-devel] what happened to the GNUmed developers ?


From: Ian Haywood
Subject: Re: [Gnumed-devel] what happened to the GNUmed developers ?
Date: Mon, 27 Nov 2006 20:53:13 +1100
User-agent: KMail/1.9.5

On Monday 27 November 2006 09:10, Karsten Hilbert wrote:

> No one else has ever raised such concerns (to my knowledge).
> I encourage anyone to speak up and lend some credibility to
> Richards ongoing claim. Even more would I like to hear
> advice on how to achieve better results.
"iron hand" is unfair, personally I think the project has suffered for too 
little leadership rather than too much, which is nobody's fault.

One problem is the high level of orthogonality between German and Australian 
requirements (or rather, requirements priorities), which took us all a long 
time to realise, Karsten, being German, has worked on his own priorities, I 
recognise he has never blocked any of us from developing Australian modules, 
for which the codebase has full support.

My problem has been the internal structural design  (in this I blame as much 
my own decisions as anyone else's) which means 100s of lines of Python code 
to get even the simplest stuff done, given the amount of time I have 
available to code, it's impossible to get any meaningful functionality (from 
the AU perspective). I think Karsten understands this problem, but because 
his requirements are orders of magnitude simpler, he, undertandably, wants to 
stick with what we've got.

So, I freely admit it's largely my own fault Gnumed hasn't moved much on the 
AU axis (which lies largely at 90 deg from the DE axis).

I have some ideas around solutions, however these involve a total code 
revolution, particularly in middleware, less a fork than a completely new 
project. To illustrate how far down the garden path I've gone, I'm currently 
experimenting with Prolog, as this seems the only way I can jack up the level 
of abstraction high enough (plus it lets you do all sorts of cool 
decision-support stuff....), but I wouldn't advocate this on the list, as I 
don't want to detract from project that's otherwise doing useful work, 
unfortunately, elsewhere.

Ian




reply via email to

[Prev in Thread] Current Thread [Next in Thread]