Dr Adrian Midgley (In the office) wrote:
> Michael Hendry wrote:
>> Adrian Midgley wrote:
>>> Colin Brown wrote:
>>>> Atos Origin do have the renewed "Infrastructure" support contract
>>>> for NHS Scotland, which includes running the Managed Servers that
>>>> have been trying to host GPASS and GPASS Clinical, so on that
>>>> platform they inevitably have control over access for 3rd party
>>>> suppliers. 
>>>> 
>>>> 
>>> Not to the source code.  Not inevitably.
> (I should have said SNHS rather than SGPC, given where the decision
> is actually made.) 
> 
>> Because Atos Origin controls the Managed Servers, it also controls
>> the executables that are run on them.
>> 
>> With a local server you can install and run what you want, on a
>> managed server you've got to take what Atos is gracious enough to
>> allow you to use. 
> 
> 1.  Which doesn't in any way necessarily limit access to
> source code, and therefore has no relevance at all to who
> might inspect such source code for faults; write (or pay
> someone else to write) corrections; test such corrected
> executables on a test system (even in a single practice
> system with a local server, you test stuff...preferably not
> on the production server with the working copies of programs
> and data) debug them and present them (via CVS or in less
> easy ways) for merging into the production system.  IE there
> is an absolute disconnect between control of (access to)
> source code and control of operational servers conflation of
> them is misleading, and much if it is carefully crafted
> disinformation from companies.

I write as a retired 3rd party developer, who was able to do (most of) the
above.

I saw something that GPASS lacked, developed a program that would fill the
gap, tested it at first on a test system, and then as a beta-test in a live
environment in the practice.

At that time, I could also distribute the program to colleagues who were
willing to beta-test in their own practices, who would pass back
error-reports and suggestions.

GPASS Users who switched to the Managed Servers had no control over what
programs were available, and could no longer beta-test anything. Only
"approved" applications were made available, and approval never happened in
the eighteen months or so we waited for my eGPR program (which had been
commissioned and paid for by GPASS) to become available.

> 
> 2.  Servers.  It is highly likely that Atos has a (farm of
> servers constituting a cluster) server which runs only GPASS
> - in fact it is likely split ito three layers, a back end
> database, a layer running GPASS against them, and at the
> front either load-balancing or just possibly some caching to
> speed response.  To add another application one adds a
> separate set of servers, likely enough to be in a completely
> different data centre.  This is a description of the Internet
> if we take it to its logical end.  The implication that
> everything has to reside on one server, and thus one
> authority must control everything in case they upset that one
> server is carefully crafted disinformation which regrettably
> continues to circulate.  _On_ the Internet - see above.

Indeed, but the active assistance of Atos is necessary for the separate (set
of) server(s) to get access to the database, and experience of colleagues
who went on to the Manager Servers was that it was hard enough to get them
just to get the basic GPASS application to run reliably over the network.

Michael