Tim Benson has written a letter to the editor of the British Medical Journal reacting to Douglas Carnall’s Medical software’s free future article. Benson argues that: ‘…An obvious route forward for the public sector would be to state that all software developed at the public’s expense be licensed as open source…’ He also weighs in on the commercial side: ‘…Unfortunately, the title of Carnall’s editorial gives the impression that open source software costs you nothing. This is not generally true. Open source software is “free as in speech, not as in beer.”…’

Not long ago I reflected on the open source medical possibilities of using the Veterans Administrations VistA project code. The VistA project has a large code base that is free and potentially offers the Holy Grail of open source medical computing: a working system. Thinking aloud on the Openhealth-list I asked the question:

Has anyone incorporated or otherwise used anything from the Veteran’s Admin. VistA project? Seems like they’ve written a lot of code already and perhaps solved many problems. I used it several years ago on DEC vt100’s and it wasn’t half bad. If no one is using or getting ideas from it, why not?

Little did I know how interesting and potentially important the answers that I would receive.

Some impressive screenshots of the system can be found here. What follows is a summary of the discussion:

Pro’s: With its new GUI front-end, it is said to be a favorite among physicians and nurses at the VA. It is apparently working and robust in a real-world medical environment. The source is free from the government under the freedom of information act. Anyone can get it and sell it. Bug fixes are potentially available courtesy of the government through NOIS and other entities such as HardHats.Org. An object oriented layer of M/MUMPS which VistA is built on is available from www.esitechnology.com and is open source. This not only updates it, but gives it a CORBA interface. VistA is being used in other places than the Veterans Administration such as state psychiatric hospitals in the state of Washington. An open source entry into VistA would be considered a major event in medical computing.

Con’s: No web interface. It requires closed-source components such as an M/MUMPS interpreter and proprietary operating systems. All implementations of M/MUMPS are currently owned by one company: Intersystems. Intersystems will probably move away from M/MUMPS in its latest product: Cache. Intersystems does not feel obligated to follow the ISO M/MUMPS standard. M/MUMPS is widely considered to be old or ‘legacy’ technology, however, an open source M/MUMPS may be available in the near future. Some of the client GUI code is said to be written in Delphi which is also not open source and not currently available for Linux although it is said to be under development for Linux by Borland as Kylix. The code is not licensed or copyrighted at all. Not all of VistA is available such as security and authorization parts which are rather important. VistA was developed with the Veterans Admin. environment and business practices in mind, so it is not always clear how VistA can best be used in a given environment. VistA applications tend to be so highly integrated that it is difficult to use only a small subset of the VistA components. This also makes it especially challenging to integrate VistA with non-MUMPS based applications.

There is a public mailing list devoted to lists associated with Hardhats.Org, the organization devoted to the VistA user community. The list address is hardhats@topica.com and the subscription address is hardhats-subscribe@topica.com