Comments on Vista Software Ecosystem Report by Herbsleb, Müller-Birn, and Towne

I was pleased to get a very thorough and thoughtful software engineering assessment of the VistA system by James D. Herbsleb,   Claudia Müller-Birn, and  W. Ben Towne of the Institute for Software Research School of Computer Science Carnegie Mellon University.

It is encouraging to see such a thorough study of the issues relating to the future of VistA.  Here is the original report. and here are my comments interspersed in it.  This comes on the heels of the Industry Advisory Council Vista Report advocating an open source future for VistA.

I was pleased to get a very thorough and thoughtful software engineering assessment of the VistA system by James D. Herbsleb,   Claudia Müller-Birn, and  W. Ben Towne of the Institute for Software Research School of Computer Science Carnegie Mellon University.

It is encouraging to see such a thorough study of the issues relating to the future of VistA.  Here is the original report. and here are my comments interspersed in it.  This comes on the heels of the Industry Advisory Council Vista Report advocating an open source future for VistA.

I think that this is an excellent report, thoroughly researched and documented – from the software engineering perspective.  I think that is important to understand VistA and open source dynamics from this perspective, but we also have to realize that VistA was/is more than just a software engineering exercise.  It was a much larger community of tens of thousands of people pulling together in an amazing feat of collective intelligence.  It included physicians, nurses, clerks, patients, programmers, managers, and many other talents who all contributed at clinical, administrative, and technical levels.  It developed a lot of content – closer to Wikipedia than Linux -  and developed a network – closer to Facebook than  the VA’s operating manuals and procedures.  VistA is an exercise in innovation in the face of stiffling bureaucracy, of user involvement and “de-professionalization” – bringing users directly into development process.

It is also an exercise in an evolutionary network approach to complex systems design.  Starting with a remarkably simple initial kernel (MUMPS language, using a single data type, 19 commands, and 22 functions) in 1978, it has undergone continuous evolution over the years.  Evolution, as Jonas Salk liked to say, is an error-making and error-correcting process.  What is critical to evolutionary systems is the fitness function – the metric by which the “survival of the fittest” is defined.  Unfortunately, many health care institutions’s survival is based on their billing system.  Clinical computing takes a back seat.  Fortunately, VA’s incentives are generally positive – it succeeds by creating a healthier veteran population.