Re: Is this bad design ?

From: ben brugman <ben_at_niethier.nl>
Date: Fri, 12 Mar 2004 15:43:49 +0100
Message-ID: <4051cca4$0$282$4d4ebb8e_at_read.news.nl.uu.net>


The actual world I am living in is healthcare.

The example comes from a >>dossier<< which belongs to a patient. The dossier here has been mapped on a mother.

It is not my task to model and it is not my taks to understand the domain of the problem. But it is my task to implement the datamodel (logical) and help with making design decisions.

The dossier consists of a collection of treatments, appointments, bservations and results.
(Although these are also linked to a person). The dossier hold them in a certain context.
The dossier can be empty, open, closed, valid and non-valid. (This summation is not complete, not ordered and some are exclusive and some are not).
A patient can have a number of dossiers, sometimes it is decided that parts of a dossier do not belong to that dossier but to another dossier so those parts are moved. Sometimes a dossier is based on something which is not anymore valid. The parts of the dossier are moved to the relevant dossiers and the dossier is made non-valid. A non-valid dossier can not contain anything. So to make a dossier non-valid it must be emptied first and then made non valid. After that no parts can be added to that dossier. To a closed dossier nothing can be added but it does contain parts.

Even if a dossier non-valid it still has to be kept.

Part of the dossier forming is for medical reasons, but part for the dossier is also for financial reasons. Because of the dossier the medical personal can have a clearer view of the patients condition. There are rules how dossiers are formed, but almost all of the rules can be bent by medical personal, there is no restriction to the medical personal on which observation they can do or which treatment they can do with a certain diagnoses. It is assumed that they have finished their medical study and they are responsible for their decisions, the dossier is an aid to limit the visible amount of information from a patient to come to better decisions.

This is offcourse only a part off the total design, and probably not totally correctly represented.

ben brugman

"mountain man" <hobbit_at_southern_seaweed.com.op> wrote in message news:Pvj4c.100346$Wa.2780_at_news-server.bigpond.net.au...
> "ben brugman" <ben_at_niethier.nl> wrote in message
> news:40517714$0$281$4d4ebb8e_at_read.news.nl.uu.net...
>
> > Just read what you wrote.
>
> Ben, could you provide an actual or appropriately analogous
> physical organisational environment within which you "mother-
> centric" application might exist? eg: a registrar of births?
>
> Many applications use for example a 'person' table, which
> can be mother, but also child, father, teacher, etc. This one
> establishes unique identities. A second table 'mother' might
> hold the tuples 'person1' and 'person2'; 'person 1' and 'person 3'
> where 1 is mum and 2 and 3 are children.
>
> Other tables, eg 'father', 'teacher' might be similarly structured.
>
> However you sound like you are seeking a solution within an
> existent and specific data (schema) environment. If so, can
> you give some background to it.
>
>
>
>
>
>
> --
> Pete Brown
> Winluck P/L
> IT Managers & Engineers
> Falls Creek
> Australia
> www.mountainman.com.au/software
>
>
>
>
>
>
>
Received on Fri Mar 12 2004 - 15:43:49 CET

Original text of this message