[log in to unmask],Net writes:
>Unless someone writes a piece of software specifically for you on a
>bespoke basis it is highly unlikely that it will be sold to you. 

Yes.  I was thinking of hiring a programmer or two, or perhaps a
software house, to write code to specification, for eg a PCG.

My inclination thereafter would be to place the source code in the
public domain, with suitable restrictions, and although the GNU licence
does look tempting something more like shareware is probably sensible.

Consider - a PCG needs certain software, as do the other 399 
Exposing it so that any passing graduate student or programmer can
refine it, sharing it with other PCGs - and looking for a return on the
implied social contract - seems a better way of sustaining development
and improvement that keeping secrets.

Meditel's practice of selling a perpetual non-transferable licence to
its software is entirely proper, and was sufficient for the days that
are passing, I se no reason to think that they will not evolve with the
times and carry on producing clever robust software that addresses the
needs of GPs.  I do think that the market will change as it becomes one
where the software houses are smaller firms than their customers
instead of the other way round.

Consider development
A software hous with a successful product will steadily become a
maintenance organisation, rather than an engine of innnovation.  THe
two roles could usefully be separated.

The needs of the end-users will be served well by encouraging software
houses to come up with something new to tempt us, and rapid efficient
mechanisms of change, rather than locking us into prorietary closed
ssytems and deploying fear uncertainty and doubt to prevent us moving
to different interface software.
THis demands the adoption of generic standards and THE MOVABLE PATIENT
RECORD and a major regulatory shift from the RFA detailed crap which
paralyses the market and the innovation in design, toward holding all
health orgs to defined data exchange and maybe presentation* standards.
 XML/SGML is the favourite candidate, but I wonder if it may be a bit
more complex than we needs given the speed of free text searches
nowadays.

Provocative stuff this, what?
--- footnote ---
* by presentation I mean that the form in which a company (Hi James)
stores its data is immaterial, if its engine provides the same
predictable response to the same function call/OLE verb/DDE string as 
everyone elses.  An engine returning HTML pages for instance (the
Meditel demo of searches popping up in Internet Explorer gave me a
shiver right down to my Achilles tendons, keep going.)

Given a presentation standard, and a standard for importing a partial
patient record 
	
	(expressed in RFA5 as "If a file of this - definition - structure
appears in this 
	directory the system will identify the patient, import the record
element into the 	
	patient record, and distinguish it from local native input in a
fashion acceptable to the 
	user group")

we can then get on with interoperability and with PCGs handling
information from hospital, HA, OoH Co-op and GPs in such a way as to
place the proper information in front of each person to do the job they
should for their patients.





%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%