Re: [Gnumed-devel] basic test types

[Karsten Hilbert]

> further in the case where a lab might have supplied a code, but not a 
> coding system (suppose it is their own internal coding system, so 
> they do not bother naming it in their result) we might in the GnuMed 
> test_type retain the default NULL for the coding_system but that 
> would create a problem.
Why would that be a problem ? NULL = NULL is always NULL

> So we should require NOT NULL and if the lab 
> provides nothing we should assign to coding_system something to 
> reflect the source lab's arbitrariness like using the value in the 
> FK_TEST_ORG key, or else the org name, to convey "whatever unnamed 
> system Lab Org X must evidently be using to assign their code".
not a bad suggestion if we in fact have a problem with NULL

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