D&C GLug - Home Page

[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]

Re: [LUG] Microsoft loses anti-trust appeal

 

On Mon, 17 Sep 2007 11:56:58 +0100
Simon Williams <systemparadox@xxxxxxxxxxxxxx> wrote:

> Tom Potts wrote:
> > On Monday 17 September 2007 09:24, Neil Williams wrote:
> > Can anyone tell me where I can apply for these NHS contracts - I guess all you 
> > have to do is make up a figure and then get a contract lawyer to pull holes 
> > in the agreement and you can be rich!
> > Oh hang on -  I can actually write software - I guess that will exclude me!
> 
> My question is where does all the money go? I can't possibly see how you 
> can spend 5 billion pounds on any piece of software. Or did it all go on 
> Microsoft licenses?

1. Time - government has no idea of how to manage something like this
so it goes on and on and on. Every time a new bod appears in the
relevant department, they want their "stamp" on it so they tinker with
it. Again.

2. Bad choice of mutually incompatible closed systems from suppliers
in direct, multi-billion dollar international competition with each
other as the starting point for a singularly *national* project that
cannot be easily rolled out to other nations, leading to zero
cooperation between developers or users. (As the NHS is unique, the NHS
system is not designed to cope with the needs of private medical
insurance companies and therefore cannot be used in systems without a
"single-client" (government) run health service.) Who's bright idea was
that?! Simply choose or develop a multi-user system and only configure
one user!! Duh!

3. Constantly rewritten specifications causing more rewrites when if
the problem had been resolved in good time there would have been no
need to waste development.

Proprietary software isn't cheap - especially when you want a quarter
of a million installations (or more?) of an ever changing specification
linking to dozens of independent proprietary packages that have no
common method of data interchange over a custom-built nationwide
network that has to be separate from the general internet, secure at all
points yet accessible to millions of "authenticated" persons and
millions more who work under the supervision of an "authenticated"
person yet completely inaccessible to everyone else (yeah, right) and
all for a single client who can't make up their mind and carrying scary
amounts of highly confidential data [1], running on an operating system
that is known to be insecure. Oh, and it was all meant to be
operational in October. (Yes, 2007.) It would have been far simpler to
trash the whole idea and start again by now.

[1] After all, what could be more useful in identity fraud than medical
notes?

All the GP surgeries have to be linked to the spine, all the hospitals
are meant to be linked as well and all the pharmacies. Every one of
those uses a different piece of software to access the data on the
central spine over normal broadband connections and not once had any
developer of those packages considered any method of data exchange
with their competitors until this whole farce started. Every pharmacist
and receptionist has to be authenticated and authorised for access to
the spine and all the old paper based prescription systems have to be
converted to digital. Over and above all that, the contracts governing
the entire system have all been rewritten and added huge amounts of
extra requirements that have consistently broken old versions of the
specification causing rewrites of the spec and delays in compliance.

Yes, it would all have been a lot easier with free software because all
those disparate packages would be able to use the same data exchange
protocols but that's just how government works in the UK. Too pally
with the patent trolls to see the reality.

Not to mention, of course, that the £5bn doesn't include the charges
that the doctors and pharmacy groups had to pay for the software
updates (which in many cases are ground-up rewrites so might as well
have been requested as entirely new software, i.e. free software).

It could have been a lot easier if the idea from the start was to use
free software for the data exchange protocols, force the rewritten
packages to be at least open source and just pay the developers direct
from the Treasury. Problem was that this whole thing was started so
long ago that free software wasn't as useful as today. The GLUG
archives are full of stories of how difficult it was to install RH5 or
earlier. This thing has been in the budget for a decade or more already
and there's no sign that those on that particular gravy train want to
get off - let alone allow others on for free.

The more you spend on proprietary the more you will continue to need to
spend because the interests of those who have already paid out (and
are looking for a return on their cash) will outweigh the needs of those
who cannot pay. You only have to look at Northern Rock for an example
of what happens when those who have paid in come to a collective
perception that they aren't going to get their cash back. Right or
wrong, the reaction is predictable. In the case of NHS software, those
who have paid in happen to be the only providers of the entire NHS
pharmaceutical and general practise IT provision. They are intent on
locking out anyone who can undercut them and as the client cannot
afford for them to go to the wall, the client listens to them and not to
the taxpayer. So it'll be another 5 or 10 or 25bn before the system
actually works. The software is relatively easy[2], what is locked-down
is access to the data updates on new products, new codes etc. that allow
the systems to place orders for stock etc. and which are solely
controlled by either the organisations themselves or quangos in the pay
of the organisations. It is these locked-down data updates that also
comprise the majority of the mutually incompatible data-exchange
bottlenecks because the GP system has to be able to be understood by
the pharmacy system by using the same codes. It would be so much
simpler if someone made the decision to dump all the existing code
systems by reason of mutual incompatibility.

[2] Heck, there are packages in Ubuntu today that could form the basis
of a replacement system. There were packages in Debian that could have
acted as the starting point before Canonical was even created.

One would have thought that the country that cracked Enigma (via a
government agency) could have created a new code system that replaced
all the incompatible systems but no, someone thought it would cost more
that way. Never - more than £5bn to create an open system of codes for
products that typically already have barcodes????

Absolute farce, the whole thing.

-- 


Neil Williams
=============
http://www.data-freedom.org/
http://www.nosoftwarepatents.com/
http://www.linux.codehelp.co.uk/

Attachment: pgpFHwK8YJ6dJ.pgp
Description: PGP signature

-- 
The Mailing List for the Devon & Cornwall LUG
http://mailman.dclug.org.uk/listinfo/list
FAQ: http://www.dcglug.org.uk/linux_adm/list-faq.html