> On Fri, 6 Jul 2001, Adrian Midgley wrote:
> ...
> > I wrote a tool some years ago for a particular purpose that offered
> > half a dozen phrases that separately or in combination served most of
> > its restricted set of purposes and were selected with a click on each
> > that applied. It was quicker and less boring than typing them.
> ...
It is only a question of elaborating the proper tool for the proper goal.
If you see your software as a medically specific word processor, it is a
good idea to have people quickly access the "most often used" sentences.
If you think you can allow decision support, it is not a good idea at all to
ease that, unless you can deal with these sentences. I can imagine two ways
you can do it :
1) natural langage analysis : you let people use free text, then try to
understand it. It is often impossible to do it in a way that is precise
enough to allow decision support (when I say impossible, I don't just mean
"so hard people can't do it, but shall do it one day", I mean that even the
brain of a man can't do it).
2) elaborate a structured representation of these sentences meaning, then
translate it in natural langage. I hope I will soon be able to offer tools
for it.
> > Imagine a daemon that watches what sort of things you type, and what
> > sort of events occur beore you type them, and stacks them up ready for
> > you. A hinting engine should be able to do that as well as be ready
> > to suggest what treatment or investigation is generally agreed to be
> > useful.
>
> Adrian,
> Thanks for clarifying. This sounds like Philippe's Odyssee system. I
> believe Odyssee auto-completes physician notes with terms from the
> semantic tree as text input occurs.
Andrew is right, that is what we are trying to do with odyssee :
If you start typing a word, we open a window that show you Lexique's terms
that fit. The more you are typing, the more the list becomes precise.
If you choose a word in the list (nice to do it), we look in which context
you are, and - if we can - propose the appropriate way to continue the
description (that's the Fil guides task). We can also open a dialog box with
your favorite "sentences" as pre-elaborated parts of trees.
Currently, I must confess that our natural langage elaboration engine is not
sophisticated enough to build the corresponding sentences, but we are
working hard on it : to be able to transform a part of a tree into a
sentence (or several sentences) when you "reduce" the branch (using the
minus sign)
Of course, this is much harder than a word processor with macros for the
"most often used" sentences. The companies we compete with used the "easy
way", but they went nowhere (thats to say they quite abandonned clinical
data administration to work on image processing and archiving systems - what
is funny is that both our competitors in the endoscopic report field and in
the echographic report field behaved the same way).
One of the reasons for it is that you need to work on the clinical meaning
of things to do good job over time.
It can be sumarized in the two sentences I could use to explain my current
behaviour :
- "Ars longa, IT brevis" stollen from Adrian Midgley
- "Data management leads nowhere, knowledge management is the answer"
Regards
Philippe Ameline
Odyssee project
www.nautilus-info.com
