Re: [Gnumed-devel] organizations

[Jim Busser]

On 2011-10-07, at 4:41 PM, Karsten Hilbert wrote:

>>>> Accordingly, can we have a nullable fk_source in
>>>> 
>>>>    dem.org
>>> 
>>> Seems sensible to me.
>> 
>> v16?
> 
> No, v17.
> 
> Karsten
> -- 

Presently, it is possible within a single

        org

to have multiple

        org_units

that are not just at the same

                street (id_street)
                'number'
                aux_street

but (moreover) share such granularly-identical values as

        subunit <--- the location within the site, for example the building or 
floor or suite number
        addendum <--- some additional information having the ability to be 
distinctive

In the case of a nursing unit, of which you could have many in the same 
hospital, is the thinking that theses might have the same

        subunit (perhaps building)

and the same

        addendum (perhaps floor within the building)

and would be adequately distinguished from each other purely by 'description' 
(name)? Why to have different physical and functional locations to share the 
same value for

        dem.address

although I suppose arguments could be

1) if a single floor within the same building served in common as the 'mail 
room' for the floor then maybe it would only be the description (name) of this 
unit which would (need to) distinguish one unit from the other.

2) I suppose similarly if you had

                one org praxis (say a group of cardiologists)

        whose constituent doctors were each

                one org_unit within the praxis

        then maybe each of these org_units would share the same address.

The alternative of course (*** and which we still did not yet figure out ***) 
was whether such individuals should instead be non-patient persons within 
GNUmed dem.identity, and each linked to the org. This latter approach could be 
better where a multi-site praxis (say a group of specialists) had two or more 
sites within the same city. Instead of each doctor being an org_unit, the 
doctors would be members of (linked to) the org via

        dem.lnk_person_org_address

Presently, the fact that the schema limits any one individual to be linkable 
within an org to just one of its addresses is maybe ok, because

        the address can be null (maybe the individual works at multiple among 
the org's branches) and
        it would be possible to look up, within the org, its various branch 
addresses

                org_unit.fk_address

Some time ago, Sebastian was thinking that maybe health professionals should be 
under some kinds of tables separately from orgs and identity but is it maybe 
better to stick to one and / or the other and not have a third construct?

-- Jim