> I imagine that current use and technological possibilities for EMRs
> are heavily weighted toward data management solutions. Knowledge
> management seems much more complex, quite heady, *and*  somewhat
> further in the future before we see widespread applications.

I am not certain of that ; eventually, I believe that using knowledge
management
principles is the only way you can build good systems for health
practitioners.

Not only to build artificial intelligence systems - I believe you see me as
a dreamer ;-) - but because what is at stack
now is not just to have people exchange *all kind of garbage* through XML
pages, but to give them
systems to support continuity of care.
And in order to do that, you need to have the same kind of *concepts
accuracy* as for
AI programming.

> If this is true, notably that there is this lead time, couldn't that
> give server and network technology time to 'catch up' to knowledge
> management's computational appetite? Couldn't browser-esque
> interfaces on light clients (connected to heavy iron in broadband
> communities... my reve du jour) conceivably pack enough punch in 2-3
> years time to justify developping for light client applications
> today, even in intense environments like Philippe is describing?

2 problems :
1) you have a genuine "light client", and server's processor does the major
part of the job : you have better forget AI
2) you have a 3 tier app, and you hope that bandwith's increase will allow
intensive data processing at workstation level.
Why not ? but I honnestly believe I shall be retired before it comes - by
the way, I shall be 38 on saturday ;-)
To be serious, just using a simple ontology (Lexique + semantic network)
server side puts current networks on their knees (that's to say
have you wait... too much time to work properly).

Regards,

Philippe AMELINE
Odyssee project
www.nautilus-info.com

Reply via email to