I think there are (at least!) two ideas here Tom.
The first gets to the effectiveness of the approach I'm trying - *querying
a VistA about itself*. The second gets to the wider question of what's the
optimal mechanism for describing a VistA.
So first things first. What I want to see is *how much of a picture of a
system comes from querying it live for three things*: its gang-of-four
"Package-Build-Install-Routines", its FileMan's schema and its Concept
definition files (dictionaries).
Yes, you face the fact that a VistA's own picture will, in places, be