bye bye ViaCode

[Adrian Midgley]

On Thursday 06 June 2002 10:29 pm, you wrote:

> How would you use WoT in the NHS?  It would be fascinating if it could
> be made to work.

> The problem as I see it is that you are going to want to send an email
> to consultant X.  Gnupg will complain that it can't find reasons to
> trust his key.  Now, you might have been encrypting your email to other
> consultants quite happily, but it so happens that you don't have enough
> acquaintances in common with X to allow it to work.

Lets take Exeter as an example.

I'm in the LMC which represents GPs.
There are quite a few of us but we meet every couple of months, and all k=
now=20
each other.
We also know our constituents, whom we meet from time to time.
We live within 10 miles of each other.
So we can sign each other's keys - one step - or we can delegate one GP f=
rom=20
each practice to take our trust and exchange keys with the other delegate=
s.

(in June I go to London to the conference of LMCs, as do a handful from e=
ach=20
county...I know these people, that can spread the web from county to coun=
ty=20
if we choose)

There is a hospital.
I've lost count of the consultants, but I know at least one in each=20
department/specialty by sight and voice.

A new consultant joined the Gastroenterology department a month or so ago=
,=20
today I got an invitation, in common with all GPs, to a meeting at which =
he=20
will speak (on Hepatitis C as it happens). =20

There will probably be one or so from each of the 20 Gen Med Practices th=
ere,=20
these meetings are in order to get to know (syn develop a defined degree =
of=20
trust with) each otehr, specialists and GPs.

In theory I can refer a patient to any consultant in England and Wales.
In practice, I might refer two or three people a year to another place, I=
=20
sent one to Plymouth a couple of weeks ago, to see a pain clinic consulta=
nt=20
who as it happened I was an Anaesthetic Registrar with some years ago - i=
f we=20
had been using PGP/GPG then we might well have signed keys and swapped pu=
blic=20
keys.

I send people with ischaemic heart disease to our local cardiologists.  T=
here=20
are two consultants, David and John, both of whom I have had dinner with,=
 and=20
in fact one of whom lives across the road from the surgery and rents a ga=
rage=20
next to my parking space...
There is an associate specialist who I used to give anaesthetics for whil=
e I=20
was working in the hospital... and whom I can meet if I care to walk a mi=
le=20
=2E..

If those people need their coronaries replumbed, then John or David refer=
s=20
them to one of the three or four surgeons at either the Bromptom or the=20
Plymouth Cardiac units with whom they deal.  They meet them several times=
 a=20
year deliberately or by coincidence at meetings about hearts and surgery =
(I=20
expect)

If I sent a child with a complex ENT problem for surgical management, I'd=
=20
send them to a chap I was at medical school with for 5 years, met a while=
=20
back at the 25 year re-union (we both remain much the same, relatively) a=
nd=20
who has a boat somewhere near here and passes through Exeter on his way t=
here=20
- if we needed to meet to swap keys then this would not be hard.

This seems to me to be not in the least unusual, and to extend along the=20
lines which actually tie medicine together.  my year and my career are no=
t=20
exceptional in this connectedness, so each year's SHOs and registrars wil=
l=20
know someone who knows someone...

5 years in med school should let you sort out who to trust for what, as c=
an a=20
year of house jobs, or even being one of three on a 6 month rotation in a=
=20
single specialty.

Basic point, it isn't random where we send patients, and we don't usually=
=20
need to sign messages to unknown people in different parts of the country=
=2E

Thoe building a system that allows someone who I never met to tell me tha=
t he=20
is trustworthy because he has a certificate form a CA may not really=20
understand how our domain and its trust relationships works.





--=20
=46rom one of the Linux desktops of Dr Adrian Midgley=20
http://www.defoam.net/            =20