gnumed-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Gnumed-devel] Interactions and stuff: proposed rules database.


From: Ian Haywood
Subject: Re: [Gnumed-devel] Interactions and stuff: proposed rules database.
Date: Fri, 27 Dec 2002 18:48:55 +1100

On Thu, 26 Dec 2002 12:07:04 +0100
Karsten Hilbert <address@hidden> wrote:

> I may have my client configured to connect to clinical notes
> on A, drug data on B and lab data on C. I will base my
> personal decision for care on data retrieved from those 3
> locations. A trigger in B has no way of knowing where I take
> my data from. For all it is worth I could have connected to
> any number of arbitrary services.
> 
> Maybe this is to far-flung but I expect problems there.
> 

Servers can talk to each other via the dblink () function
[new in 7.3 I think]

However, Horst has described German hospitals that have installed 
physically separate networks for different computer services, so 
even dblink would not work. 

So in this case you are right: the inference engine has to run on the client, 
with
rule-originating queries flying back across the network: a potentially 
significant 
and pointless speed penalty for the poor GP with aging hardware. 
(I'm sure we've had this discussion before....)

It's all in Python either way, so I think we can cover both bases
easily.

Ian


Attachment: pgpB1Kq9yw4bf.pgp
Description: PGP signature


reply via email to

[Prev in Thread] Current Thread [Next in Thread]