On Wed, 11 Apr 2001 13:02:24 Wayne Wilson wrote:
>pschloeffel wrote:
>> Database bindings are another big area of work. The only binding at present
>> is to the ODB Matisse which was really done as a demonstration and because
>> Thomas had previously used this database in a commercial (non-health)
>> project. Anyone volunteering to write bindings for PostGresSQL, Oracle,
>> DB2, Sybase, SQL Server etc will be welcomed with open arms.
I raised this issue at the AMIA 2000 meeting and I raise the same issue again. I must be missing something since I still don't know why the GEHR project continues to be stuck on the database issue.
Why do you have to re-implement database binding for the GEHR project? Are there no existing Oracle, PostgreSQL, SQL Server, DB2 etc database "binders" that you can use?
By using Zope, the OIO system enjoys database adaptors that include Oracle, PostgreSQL, etc. We use SQL within OIO to communicate with the database. When the database adaptors are improved, the OIO system becomes a better system without having to waste time writing and maintaining our own database adaptor.
Is it a good idea to rely on GEHR-specific and Brand-specific database routines?
What we have done is create a simple interface that can map to OODBMS or to RDBMS through several
Object-Relational mapping tools (ObjectBridge at Sourceforge is one we really like). So we simply implement
these interfaces which allows us to use a variety of databases, and we use Objectbridge to map to "any" RDBMS,
since that is its power. You simpy creaty an xml file that describes the mapping between the tables and the objects (including collections, foreign keys, etc.) This we have almost no database specific code in our OpenEMed application. We have found this to work very well. It may be harder if there are no Object-Relational mapping tools in Eiffel.
Besides giving database portability, it allows us to always work with objects (which ODBC or SQL has no knowledge of).
Why not re-use existing routines?
>These are the kinds of things that could be added as other tasks to the
>one's I
> suggested earlier be done by the Internet2
>consortium that Mary Kratz has formed.
I am not sure throwing more resources on this problem is the best solution. First of all, it may be a good idea to understand the development strategy that you are proposing.
Are you suggesting that we need to write an "binder" for each database that GEHR talks to? Could we possibly use a generic SQL gateway to solve the problem?
(cf. Objectbridge at sourceforge, and there are dozens of commercial products that do this. One really good commercial product is JDX from Software Tree). It seems to me that the limitation of using Eiffel is that there are not a lot of libraries being developed to do this. In Java, a lot of the work is being done for us without our even having to ask.
Dave
Please ignore me if I am not making sense,
Andrew
---
Andrew P. Ho, M.D.
OIO: Open Infrastructure for Outcomes
www.TxOutcome.Org
Assistant Clinical Professor
Department of Psychiatry, Harbor-UCLA Medical Center
University of California, Los Angeles
Join 18 million Eudora users by signing up for a free Eudora Web-Mail account at http://www.eudoramail.com
Computer and Computational Sciences
Los Alamos National Laboratory
505-665-1907
