[EMAIL PROTECTED] wrote:
>
>
> So, if one of the most modern hospitals in the world is
> implementing CorbaMed as we speak, I don't think anyone
> can say it's dead.
>
I really do think this comes back to the question I raised earlier about
what the target market is coupled with whether the approach is
organizational based on individual based.
This example of CorbaMed is exactly what I would describe as the large
organizational approach. IT makes sense in that context, especially if
you have the support of in-house developers that are not considered
seconday to the the vendors. With CorbaMed, the one exception I make is
Dave's work with OpenEmed but then Dave is exceptional.
So my argument goes like this:
1) for the small scale, individually started effort CorbaMed feels like
overkill.
2) In the internet application space the 'plumbing' behind CorbaMed,
i.e. Corba itself, also finds it's best expression in well funded and
large organizational settings. And the most widespread use of Corba is
just that, behind the covers, in ejb containers or the KDE/GNOME
desktop environments.
Thanks for the compliment. I do think that there is a business case for
CORBA HDTF specifications in the small practice. It could open up
business in another way by enabling a Doctor to provide their data
in a commercially viable way to larger organizations with little
effort on their part. I don't know how big this market is, and I think
it only needs the right support by a small company. I don't think
the maintenance is any different for a web-based 3 tier application to
use CORBA on the back-end. In fact, once you understand the technology,
I think it can be simpler, because of the nice encapsulation of objects that reduces
the number of ways things can go wrong. The sharing of data with referrals
and providing (and receiving ) data to and from others would be of great value
to the doctor and to the patient. Since the technology is pretty mature, the cost of
deploying it isn't really any different than other technologies. Borland demonstrated
at the last OMG meeting the ability to simply drag-and-drop build of a CORBA application,
for example.
Our goal in OpenEMed is to lower the entry level to the use of advanced technology in healthcare
so that it might become more ubiquitousl. The Brazil example shows that it works on a large scale
and we believe we have shown that it works on the small scale. We just need people who
are building and marketing desktop systems to implement their systems according to the
component specifications. At that point we can begin to have adhoc collaborations between
doctors that are essentially impossible today. End users need to demand it, too. It could
all be XML, but the capability for collaboration with XML is a considerable distance away because
of the immaturity of the technology. It also does a poor job of encapsulating objects, so
it tends to follow the old structured programming paradigm (although it doesn't have to).
Enough rambling.
Computer and Computational Sciences
Los Alamos National Laboratory
505-665-1907
