Dear Jim and others,
One of our difficulties with the GEHR Ocean Kernel is the binding of our
kernel to database engines. We believe our appoach is a good one - offering
a standard open source component as middleware with COM and CORBA
interfaces - but building and optimising the backend bindings for all major
databases is beyond our resources and expertise. For that reason I would
like to try and set up an international project to write bindings to a
number of suitable database products - perhaps with vendors as partners and
the openSource community looking at Postges and mySQL.
It is possible at GEHR to remain compatible (we believe) with the HL7
efforts (at least drawing attention to when they have internal
inconsistancies) and CORBAMed interfaces. This is important in itself. I
kind of summarise the different approaches as:
_________________________________________________________
Approach Activities Transfer Implement
Design
_________________________________________________________
Messaging HL7 v2 ASCI ++++
+
HL7 v3 Any
++++ +++
API CORBAMed Objects +++
+++
Record Architecture PRA XML ++?
+++
GEHR kernel EHRs +?
++++
_________________________________________________________
This analysis is speculative (Gunther?) and obviously will have to be
proven - we hope out current trial of the Ocean GEHR kernel in Australia
will demonstrate some of the utility of using a common component - when 3 GP
systems start to share Health Summaries and Diabetic records.
I do see that binding the kernel to current DBs is required - even for
testing. So the question is - can we get some energy together to test the
kernel with MUMPS, Oracle....
While this may send horrors through the - I'll only use it if its free
community - clearly it is a reality in health care that people will have a
diverse range of DBs behind their EHR systems.
Any thoughts?
Sam Heard
