[Gnumed-devel] re: gui / schema

[sjtan]

The business, pycommon , backend postgres schema / functions/ triggers,  
are pretty reusable.

The specific uses of the client make it difficult to use the workflow in 
different environments:
(e.g.  the lab result widget ; how to get it working with test data  )


The schema is pretty understandable, except when triggers other than for 
something cross-cutting like auditing is operating
(I don't think there are any yet).  I'm not sure about how external id 
referencing is being used : is there already
a business object handling this ? There are also a lot of constraints to 
follow for some tables ( not null,  value domain restriction etc).

It's ok to reuse components for say custom clients too, or maybe 
customization of layout will be done?
Is Ian/someone adding extra customizability to  the client platform ? 
There's already the plugin load/unload
feature which  Karsten did recently, which makes things pretty 
customizable. 

Is it possible to generalize Richard's experience about workflow,  say 
using a specific behaviour on the editareas as an example,
citing the reason a certain workflow is recommended (speed, usability 
comparison to another workflow that didn't work as well)
,  and then make guidelines out of them , and perhaps examine 
applicability with different layouts. ( just discursively , at this stage)
For instance :
text completion usefulness for each case ; 2 tabbed panes layout, vs 
single tabbed pane; tabbed pane and tree; tabbed pane and list; soap widget;
4 pane view (epydoc style, but with one a edit pane); 2 trees and 2 
panes; layout for the mainly mouse-user, layout for the mainly keyboard-user

The other issue is forking / dilution of effort : is it really forking 
if some components of the current gui are put into a client customized for
a particular area/ purpose ?