| Hubris | Open GP Clinical Systems |
This may seem unimportant at first, but one cannot predict the future development of for instance a company marketing a product they have written to run under particular operating conditions.
Among the less obvious ways of undermining openness are legitimate programming techniques such as adding an index dependent on a function which is concealed or proprietary. This stops amendment or in some cases even reading of a data file by other programs.
Revisiting the subject in 1999 there are obvious developments toward the common standards of the Internet and the Web.
In response to a question "What is an Open System" from a worker at Apple/Newton, the best reply presented was :-
James R. Bell, Chairman of the Board of X/Open, a
founder of the Open Software Foundation (OSF), and former (and first) CEO
of The Open Group (the merger of the two organizations) --
"Open"/"proprietary" refers to where a technology falls on
an axis including the following ordered set of points.
Companies often choose to make a binary definition via
cutting the axis, always at a place that makes themselves
"open":
- product spec well documented to enable independent
implementations
- interoperable products allowed/encouraged
- interface spec differentiated from product spec
- compatible replacement products allowed/encouraged
- any privately held mandatory technologies available
on equitable terms
- control of spec by set of selected partners
- control of spec by de facto standards body
- control of spec by de jure standards body
The above is a personal definition, [however], an official
one is available as part of the POSIX work.
International Standards and EffortsThe Good European Health Record project (which may have bitten off more than it can chew - something similar was tried in Exeter some years ago - and HL7 are worthy efforts.
GPs should tolerate nothing which is directly opposed to these projects.
RFAThe requirements for Accreditation have concentrated more on the data to be held than on sharing it.This is a classic and massive missed opportunity where a little work at the beginning followed by a great deal of hands-off observation could have allowed great improvements. Unfortunately the reverse held true. Do we have Open Coding?Read is not very open. It could be, and if it were to be a candidate for a worldwide GP coding system it would have to be. The challenges are more managerial and administrative than otherwise, and some progress has been made toward using the modern technology Read is supposed to be a part of, for serving its customers. Many of the deficiencies alleged of the Read Code are in fact failures of suppliers to implement the user defined coding parts which are well defined. If you cannot readily add a code in chapter @.... to your list of codes, complain to your supplier now.Fact-checking with CAMS and the NHSCCC is advisable if a supplier indicates that the copyright des not permit new codes or that NHSCCC discourages the provision of @codes to users. Microsoft might well argue that Windows is an open system. Many features of it such as particularly OLE and OCX controls are features lending themselves to an open system approach to a clinical system, and the Component Object Model (COM) of programming suported by these and by the big Microsoft applications exploding into their exposed objects over the last two or three generations is definitely a gesture in the right direction. Windows TM - Some of the features of Windows would be useful in clinical systems, but which, and how to get them? ReportingThe system should only produce one report itself, but satellite applications might well handle particular tasks which depend upon information held by the main system, or require access to functions within it.An example would be an audit or other analysis program which ran overnight or inthe background, but which required access to the permission controlling engine in the main software. It is desirable both that third parties can produce analysis routines, and that access to medical records is controlled by a proper access list, and therefore the interface between the two programs must handle this co-operation. Since there is an arbitrarily large possible number of third party routines, and indeed the number of different main software systems is only weakly constrained to be less than the number of Practices using them, it is obviously economical to use a common interface. Effectively Microsoft and Unix have provided these, and they do appear to be converging.
1999: the Web is the means of connecting, and HTML and XML the ways to display or export structured information. Reporting has become easier. | |