>>It is possible that Rushmore is disabled when there
>>are buffered records?
>
>
>No, rushmore works regardless. When you try your tests in the Command Window, do you set buffering on there? Or, is buffering only active inside your application?


I was not using table buffering from the command line, but I will
try that for comparison.

I'm using optimistic table buffering via VFE/Codebook and there
is a lot I don't know about how the I/O works. I will try to
get a better handle on exactly when it "goes out to lunch".
What I experience is that sometimes when I change some code and
test I get to a point where the disk goes into this spasm
of activity. If I'm patient (several minutes) it will
finish whatever it is doing and come back to the active window
or the debug window.

I've just grokked the fact that when I say "append blank" in
this enviroment, the new record is anything but blank since
all of the default values in the DBC have been applied. That
has clarified things quite a bit.

Peter Robinson

p.s.
See related question about record numbers after table
buffer updates.