At 20:43 +0000 on 15-12-1996, [log in to unmask] wrote:
> Many things to reply to!
>
>
> 2. Alan Capey says "From your approach I assume you have formal systems
> training" Not strictly true - I lapsed from GP into AI/Clinical
> Informatics research, but am interested in user centred design, formal
> methods and evaluation.
>
> It *is* hard for IT people and clinicians to communicate well, but they
> must. We demand a rigorous (+/-) process with drugs - design, trials,
> evaluation, evidence. Why not computer systems? The more safety-critical
> the system, the greater the need for confidence in its performance, surely.
> I regard information used for making decisions as very quality-sensitive.
> That is why we need Clinical Informaticians (self-interest declared).
I agree for more than 100%. For devlopment we need: technical people, users
and Industrial Designers in between is my firm believe.
> 3. Gerard Freriks asks why I am fed up. Many little things, not one big
> one. Someone who works with me "discovered" relational databases a couple
> of months ago. They will, of course, be the answer to everything. Expert
> systems were going to take over in the '80s...where are they now? (well,
> some of the bits that work are in things like Microsoft's Wizards). Many
> commented that many of the EU's Advanced Informatics in Medicine (AIM)
> projects seemed to be ignorant of clinical life (they now demand a clinical
> lead and demonstrators in real clinical establishments).
>
> My view is that we must understand what needs to be done before we can go
> and do it. New developments bring new opportunities - I'm at the more
> formal end of the spectrum when looking at these things, that's all.
True again. Before to start something new, start thinking and then think again.
Many systems I know were developed by teckies and never thought about all
kind of possibilities and exceptions. Or what realy the proces is that
users are performing using an application.
> 4. Mary Hawking says something very important. Get a lot of users on
> board the development process early and keep them in! For example, Pen&Pad
> in Manchester was user-centred - GPs came in and tried versions of the
> system and said what was wrong, right, missing, not needed. At least had
> something to react to - conventional design methods just ask up front -
> nothing to say "horrible!" about.
Involving users is necessary. But in general they know their workprocesses
only and have opinions about the Computer User Interaction. They lack
knowledge about computing, systems, software, hardware.
We need the middle-man.
> 5. The Internet is a messaging system. The browser/client sends a request
> message to the server. A reply message comes back. The protocol defines
> the form (syntax) but not the content (semantics in a way, I suppose). I
> think we need both - use the Web as the means, content structuring to
> prevent nonsense and errors, and to enable the fancy stuff. People are
> pretty good at picking up mistakes, but computers need a lot of help.
>
Therefore using HTML is not sufficient. We need a general standard to
describe Context. E.g. how do we indicate the the figures a a specific Lab
result?, name, diagnosis, coding system used.
The super-set of HTML is Standard Generalised Markup Languege which does
this trick of aplying context to content. But first we need standard Labels
to be used in documents.
Gerard Freriks,huisarts, MD
C. Sterrenburgstr 54
3151JG Hoek van Holland
the Netherlands Telephone: (+31) (0)174-384296/ Fax: -386249
Mobile : (+31) (0)6-54792800
ARS LONGA, VITA BREVIS
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|