Re: [Gnumed-devel] widget connecting

[Karsten Hilbert]

> Why not
> just convert gmTmpPatient to a remote object ?
I wouldn't mind seeing gmTmpPatient being worked into an
XML-RPC server part. I would just want to be informed about
things so I can understand what's going. This may form the
basis of gmDemographicsServer.py.

> Or maybe K and you write a business framework specification
> (no coding, just interfaces , in python , rpc IDL, corba IDL )
I can't since I am just a poor coder (but one with strong
opinions on how *I*'d like to see stuff done :-)

> then toss it around for committee review until everyone agrees this is
> what is wanted.
Will never happen I assume.

> (couldn't take much longer than the constant stops and starts 
> at hacking out connection code).
While the proposal may worm it's way to wider and wider
agreement the code evolution will worm it's way to widening
viability. Of course, as Hilmar would suggest, this needs to
be complemented by documentation.

> Why do we keep re-writing code that works in a basic sort of way?
This is evolution.

> It would be good if some sort of versioning control , or plan could be
> done e.g. 
A plan may help. But a plan is no good if no one is there to
stick to it. Or if it gets scrapped after yet another
non-surprising relative disadvantage over any of the other
possible plans is discovered.

>       cycle 1 : put out the poor man's MD. Stores and prints scripts, notes,
> and a few lists , prints letters, enforces a investigation follow up
> accounting.
Nothing more is desired at this stage :-))  But this requires
all the infrastructure to be just about right. That is the
hard part. I for one don't want to build another OIO, OSCAR,
etc (schema-wise) coz I could just download any of them.

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