FLAP and Fileman Code Convergence

One of the two top strategic goals for FLAP Phase One (which ran from October 2012 through March 2013) was code convergence, to create a shared, common Fileman codebase that all VISTA dialects could run, to maximize our ability to exchange data and software with one another, to ensure that innovations anywhere became innovations everywhere as quickly as possible, and to reduce the amount of reinventing the wheel we had to do.

The main VISTA dialects we aimed at converging for Phase One were:



3) WorldVistA EHR

4) vxVistA, and

5) Medsphere OpenVistA

We were not aiming to converge the most diverged version of VISTA, (6) DOD's CHCS, because that was a larger project than we had time to undertake.

Of these six dialects, #5 was the most interesting, because at the time we began FLAP Phase One it included thirteen years of bug fixes and innovations by Fileman guru George Timson, much of it paid for by Medsphere Systems Corporation at a time when VA investment in infrastructure applications had dried up to just a trickle. Medsphere's investment was thus critically important and strategically brilliant - to shore up VISTA's foundation when it most needed the help. FLAP Phase One delivered on that goal, producing Fileman 22.2, which brought together all five target dialects.

This preparatory stage of FLAP Phase Two is also mainly about code convergence. It aims to converge four Fileman codebases to produce a fifth:

A) VA Fileman, which has been patched three times since FLAP Phase One was finished;

B) Fileman 22.2, the product of FLAP Phase One;

C) MSC Fileman 1051, with another year and a half of bug fixes and enhancements, as well as compliance with OSEHRA's new code-attribution policy;

D) The FLAP team's codebase from the Fileman Development Environment, which has fixes and enhancements beyond FLAP Phase One's work; and

E) The target version, MSC File Manager 22.2, which will be current and attribution-policy-compliant.

Part of the work of FLAP Phase Two, which falls within OSEHRA's Code Alignment workgroup, will be to develop patches as needed to help each of the five converging VISTA dialects install this new MSC File Manager 22.2, including any dialect-specific conversions needed to migrate to this new, standard MSC Fileman. For example, the Language file in Medsphere OpenVistA is a little different from that in MSC File Manager 22.2, because they had to innovate ahead of the FLAP team's standard solution, so this stage of our work together will be when we write a conversion to migrate their sites from their existing Language file and pointer values to the standard ones. We'll need to look at each of the other four dialects in the same way, to ensure we create any necessary conversions for each of them, to smooth the path for them to converge on this new standard codebase.

I've attached to this post a map of the VISTA dialects created by Carol Monahan and me and updated in recent years, to help you keep track of these dialects, their ancestry, and the relationships between them. This is among the subjects the FLAP Jordan Team reviewed this morning.