Re: [Gnumed-devel] roadmap

[Horst Herb]

On Mon, 28 Jul 2003 21:07, Ian Haywood wrote:
> Perhaps we should consider odd-even versioning like the Linux kernel: evens
> are feature-frozen, it aims to
> to get it "out the door", odds are the cutting edge for trying out new

I would suggest to stick to our plugin concept: If somebody is bold enough to 
implement such XML parser, he can implement as many modules as he likes *as 
plugins* into the existing framework - if it performs well and if it is 
easier to write the XML files than coding the old fashioned way, evolution 
will set in and such concept will then gradually replace conventional 
modules.

This has the advantage that we can just continue using what we have, while new 
concepts can emerge and prove their merits.

In order to use gnumed as it is, all we need is the demographic "business 
objects" or a demographic service implementation. The rest is just connecting 
existing widgets to database or meta objects and pass a few messages around.

Horst