Re: [Gnumed-devel] Drug Progress

[Karsten Hilbert]

Ian,

> I propose a coding system for drug database references of db:date:code, so
> 
> drugref.org:040503:324344
> 
> is drug 324344 in drugref.org released on 4th May. The code must be smart
> enough to deal with a change in drug database (so then it must search by the 
> redundant data 
> available) 
This is nearly identical to what I put into

"server/sql/gmclinical.sql/allergy.substance_code.comment"

:-)

Let's refine that:

- 3 fields
- field 1:
  - source identifier
- field 2:
  - source version
  - this might already contain colons, so we'd better use :: or so
  - this is the version as supplied by the source, not
    necessarily a release date
- field 3:
  - opaque drug ID (that is: non-checked foreign key) valid in
    source given by field 2

Applications are expected to deal with this correctly. The
information is taken to /not/ be authoritative.

This format is to be used in fields that hold an option
"foreign key" for drugs. Of course, a full-blown solution
based on tables is also possible but we likely don't want
those tables in each and every database, do we ?

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