Horst,
>
> And, in case of your choice of Eiffel flavour, which do you
> think helps GEHR
> more:
> - YOU being able to use the tool you are used to with maybe a
> few handy
> proprietary libaries not available as free yet and a probably
> nicer IDE, or
> - literally hundreds of developers joining in and helping out
> because THEY
> can use the tools they already have and trust? Tools they can share
> unrestricted?
>
There is more than the developers view (either Completely OS Correct or with
some slack) and that are the requirements from the organisations in which
the software has to be used. I'm affraid that many of the OS developments -
and I know there are exceptions like VistA - try to tackle a rather local
problem, without taking into account the larger picture of information
provision in the health care environment. In the Netherlands there are
serious problems with the connectivity among GP systems. Also Hospital
Information Systems may not be communicative among each other (perhaps among
HISes from the same brand, but not necessarily accross brands). Therefor, I
think that work like that within GEHR and in CORBAmed (with e.g. the OS
implementation OpenEMed) are necessary. Here the focus is on choosing the
tools that are most suited to get sustainable solutions, not necessarily the
tools that are best known among the programmers community.
I have been working at the department of Medical Informatics at the Free
University in Amsterdam. There were quite a few people that used Mumps
extensively and also the Medical Informatics group in Rotterdam is currently
still using that language since it is particularly suited for handling
sparse data bases. So there is a business case to choose MUMPs as a
programming language rather than C, C++, Perl, JAVA or Phython (they didn't
even existed in the early eighties when that group started using MUMPS).
My starting point is not the dictum of OS but merely the need in the health
care environment for a working and sustainable information infrastructure. I
advocate discussing the specs based on (emerging) standards like HL7 RIM and
CEN ENV 13606 in an open environment. The modelling should use preferably
UML although no OS software for proper UML modelling exists as far as we
know. From these specs a coordinated development should take place,
preferably using an OS model for that software. However to use GPL as the
licensing model may steer industry away from this market and we become
reliant on software developers that somehow can find the time and money to
devote efforts to further developments. Whether that is a model that will
also result in a proper implementation of the dirty bits has to be seen.
We are soon going to start such an activity and I will keep the list
informed about our progress.
Regards.
Jan Talmon
Dept. Medical Informatics/Maastricht University
Foundation for Innovation in ICT in health care
