Minutes on Moving VistA Forward - a Modest Proposal for Reengineering VistA, K.S. Bhaskar:
Thanks to Bhaskar for an interesting discussion on reengineering VistA. Although the title says "A Modest Proposal", we know any reengineering of a complex large scale enterprise system is never modest and always a huge undertaking - requiring a long term vision and large financial commitment over many years. That said, Bhaskar presented the lessons learned from a successful reengineering of the "FIS Profile" - a large scale enterprise M-system based application that runs the three largest real-time core banking systems in the world.
Back in the late 90's the "FIS Profile" application faced similar concerns to those that apply to VistA today, and over a period of several years, the FIS team slowly evolve the "FIS Profile" from M language to a more maintainable higher level scripting language called Profile Scription Language (PSL). During the reengineering process, the team also made many improvements such as separation of UI from business logic, ensure application consistency after a crash by enabling transaction fences, support stateless message processing, etc.
Some of the key lessons learned:
Horizonal reengineering instead of vertical reengineering!!!
- Revolution through evolution
- Avoid encapsulation
- Choose a language and compile it into M
- Re-architect as you go
- Integration beats Balkanization
- The value of M technology is as an execution engine
Links:• FIS GT.M: http://fis-gtm.com• FIS Profile: http://fis-profile.com• FIS PIP: http://fis-pip.com• Anti-patterns: http://sourcemaking.com/antipatterns/software-architecture-antipatterns• K.S. Bhaskar / email@example.com / +1 (610) 578-4265
We will not having any meetings until the third week of January.
Tentatively scheduled presentations:
January 10th - Special session on VistA Novo Demo, Salim Samy, Mitre
January 14th - Socratic Grid - Open Source Clinical Decision Support, Emory Fry, MD, CMIO, Cognitive Medical Systems, Inc.