[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Gnumed-devel] Mock up of Potential Patient Data Entry screen
From: |
Karsten Hilbert |
Subject: |
Re: [Gnumed-devel] Mock up of Potential Patient Data Entry screen |
Date: |
Thu, 9 May 2002 11:39:35 +0200 |
User-agent: |
Mutt/1.3.22.1i |
> > 1) I type the last name (auto-completion) - possibly
> > word-wheeled to a query
> > "SELECT unique(last_name) FROM PERSON"
>
> gnumed=# select count(distinct lastnames) from names;
> 1017
> gnumed=# select count(distinct substring(lastnames,1,2)) from names;
> 157
>
> With two characters typed in already it starts making more sense, as for
> any combination of two letters it will always be only a a few - but
> unfortunately rarely only one
>
> Thus, I would not start the word wheel befiore the third character has been
> typed in otherwise we would be very wasteful with ressources.
Absolutely. That's why I introduced the thresholds into the
match provider which I showed to Ian and discussed with
Richard:
threshold 1:
- if substring has this many characters -> find only
phrases that start with the substring
threshold 2:
- find only phrases that have a word starting with the
substring
threshold 3:
- find phrases that have the substring anywhere, even inside
words
These thresholds need to be very different for different input
fields:
A patient search field will likely require higher thresholds
as you point out.
A context-dependant field (say, referral target doctor) will
likely require rather low thresholds.
> number of chars to be typed before the word wheel loads from the database
> based on individual database statistics (just run a few count(distinct
> ...)) whenever you start the client)
Good idea.
> > | postcode : ___________________________[v] |
> > | street : ___________________________[v] |
> > | suburb : ___________________________[v] |
> > |
> > | street2 : ___________________________[v] |
> > | street3 : ___________________________[v] |
>
> Instead of street2, street3 I would definitely go for a multi line text
> control.
Duh ! Of course ! Didn't think of that.
> I don't think we should fall back to "modal" behaviour of entry / display
> widgets unless there is an outstanding benefit. The user should always be
> able to edit data in the widget that displays the data, otherwise it gets
> too complicated for most computer naive users
Well, sure, Richard also made some good points why it may be
Good to have the same widget for _display_ and _editing_
patient details.
However, I still think it quite beneficial (and I am
thinking of our receptionists) to have a super-efficient
widget for entering _new_ patients. One could, in parallel to
entering data, fill up the "normal" widget displayed next to
the entry form.
Karsten
--
GPG key ID E4071346 @ wwwkeys.pgp.net
E167 67FD A291 2BEA 73BD 4537 78B9 A9F9 E407 1346