Re: [Gnumed-devel] Some suggestions to make gnumed more efficient

[Christoph Becker]

Karsten Hilbert schrieb:

>Take your time but this would be really helpful. The problem is,
>however, that the schema graphics need to be updated
>regularly...
Hallo Karsten,
a regular update is impossible for free. I played arround a bit with the gnumed schema in dezign for databases. It will need several hours, to optimize the tables on the display in a reasonable way. This is a nice game to learn about how the tables in gnumed are organized, but to do this on a routine base would be hard work.

I is a big difference wether you design a database with such a tool as dezign for database, or wether you try to reengineer an non trival existing database with many dependencies, to learn how it is organized. You have to move the tables around many times until the wiring diagram becomes readable, and then you still have several view options to make it even more complicate, 
A solution might be to put gnumed into such a dezign tool and then keep it up to date by doing all further modifications inside the dezign tool.

Regards
Christoph Becker
 

Hallo Christoph,

  

Dezign for databases imports an sql-schema via ImporterScript tool.  
You can't run it against a running database.
    

I understand.

  

Even more, because of some bugs in their ImporterScript tool,  a 
schema-file written with pg_dump has to be run through a filter first. 
For example character varying (n) has to be replaced with varchar(n).  
    

I just checked, we don't have any "character varying" in use.
Also, I replaced most varchar(n) with TEXT.

  

It should be easy to install the gnumed backend on Windows when using a 
native Windowsport of  PostgreSQL 8
    

Well, our bootstrapper needs testing and likely some changes
for Windows. Sure it would be nice to have a point-click
installer but someone needs to do it.

  

and it should be not to difficult to 
understand how the gnumed backend works
    

Writing a good schema is hard. Understanding and using it
requires thought.

  

and how it can be interfaced by a GUI.
    

Ideally, programmers should use middleware/business APIs to
access it. Eg.:

patient = patient_searcher(lastname=..., firstname=...)
emr = patient.get_emr()
print emr.get_missing_vaccinations()
print emr.get_allergies()
print emr.get_last_encounter()
print emr.get_current_medications()

This is how the GnuMed business layer works.

  

If I get Gnumed running on Windows I will send you the scripts needed as 
well as the output of Dezign for Databases. But this may take some time.
    

Take your time but this would be really helpful. The problem is,
however, that the schema graphics need to be updated
regularly...

Karsten