I agree with Gunther's assessment of the implications of being dependent
upon MUMPS for VistA and the importance of building a system that is truly
open source from the ground up. On the other hand, there is a lot of work
that went into developing VistA, and it has been a successful platform. I
(not surprisingly) think it would be a shame to throw out all this
experience and start over from scratch. But is this really necessary? Even
if we end up building a system that does not involve MUMPS anywhere, we
still have access to the source code, the data model, the user environment
(where options, for example, may correspond roughly to use cases), etc. The
question is whether we can effectively reuse all of this analysis and design
work. I think we can (though that doesn't mean I necessarily think recoding,
say in Java, is preferable to building an open source MUMPS system). In
fact, one idea I have been thinking about pursuing is either exploring
implementation of familiar design patterns with Fileman, exploring new
patterns that are more VistA-specific, or both. I don't want to exaggerate
the importance of the pattern concept, but I do think it's an exciting idea,
and one which provides a framework for exploring existing designs in a
language independent manner and, more importantly, how those designs can be
applied to new situations or used in new environments

===
Gregory Woodhouse <[EMAIL PROTECTED]>
Financial Product Line
+1 415 744 6362
"Computer science is no more about computers than astronomy is
about telescopes." -- E.W. Dijkstra

Reply via email to