Re: [Gnumed-devel] emr questions

[Karsten Hilbert]

> gnumed had a few quirks with  s o a p , is_rfe, is_aoe, ( is_rfe should 
> be only available with s , and is_aoe only available
> with "a"  ,  ) -  why can't s o a p     be pushed into the heirarchy of 
> a clin_root_item
1) People seem to want to make them *less* basic than they are
now, not more (not that I agree too much with that).

2) Exactly *how* would one do that ? Interleaving a
zero-columns parent table between clin_root_item and rfe etc. ?

3) If one makes soap a structural instead of settable
attribute it will be harder to reattribute rows and it will be
more involved to aggregate the complete narrative.

IOW what would we gain by inheriting SOAPness ?  One needs to
remember that PG inheritance bears little resemblance to
standard OO inheritance.

> clin_root_item references clin_health_issue directly, and perhaps a 
> sequential attribute / episode_number
> classifies the episode of the clin_health_issue it relates to,
Many episodes will not even be linked to a clin_health_issue.
The patient may not even have any clin_health_issues.

> instead of having to tag each episode with a string name ( this was an 
> annoying requirement when I tested
> the schema with the web client).
We are currently trying to replace having to explicitly tag
episodes with names by moving to marking clin_narrative rows
as defining the episode name implicitely. Those clin_narrative
rows will exists *anyways* because that's where people store
narrative.

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