David Forslund wrote:
> 
> Many of the healthcare problems involving linking heterogeneous systems
> together involve
> complexity.   Opensource projects need to address this
> complexity.   Writing a kernel
> in 5 languages is not the way to address this complexity and probably make it
> even worse. 

Perhaps, which is why my original question was about interfacing the
OpenEHR
kernel written in Eiffel to other languages, not re-implementing it. I
was
trying to get an idea of the amount of work involved in these
interfacing
tasks. 

By "interfacing", I mean the wrapping of OpenEHR (or whatever) to make
its
use seem "natural" from the developer's language of choice. Of course,
re-implementing OpenEHR in a range of languages is the ultimate method
of doing this, but such re-implemntation is only a good idea if the 
OpenEHR kernel is quite simple (where, all things
being equal, simpler is better). If it is not, creating
language-specific
wrappers for a single Eiffel implementation might be more sensible.
That's
what I was asking Thomas about.

Tim C


>             Although the OMG documentation may be long, it don't think it is
> any less digestible than the current OpenEHR documentation, just
> different.  This is
> not intended as a negative on the OpenEHR documenation, either.
> 
> Dave
> 
> At 09:39 PM 10/27/2002 +1100, Tim Churches wrote:
> >David Forslund wrote:
> > >
> > > I don't understand all this discussion about implementation languages.
> > > We need the API in a platform and language independent form and then
> > > we don't have to worry a lot about the language except to make sure
> > > that the implementation language works on the platform we choose to run on.
> > > The OpenEHR should be made to conform to open standard interfaces (such
> > > as those of the OMG HDTF, and which it is very close to anyway).
> >
> >I hope that the OpenEHR API is a bit more digestible than a typical OMG
> >document. Yes, I acknowledge that exhaustive specifications
> >which run to hundreds of pages do have some utility, esepcially for
> >large commercial system vendors, but most open source projects are
> >resource-constrained, and if OpenEHR is to be useful and used in such
> >projects,
> >it needs to be simple enough for individuals to get their head around in
> >a
> >few days of effort. Or at least, there needs to be a version of the API
> >documentation which fulfils this need.
> >
> >Tim C

Reply via email to