Experimental Physics and
Industrial Control System

Oops, my bad, I mis-read the dbCommon.dbd file, UDF /is/ writable, but
writing to the UDF field of an ao record has little effect:

tux% caput anj:ao.UDF 1
Old : anj:ao.UDF                     0
New : anj:ao.UDF                     0

I think what happened is that the record processed as a result of the
put [since UDF is marked pp(TRUE)], and the process routine overrode the
1 that I set the UDF field to with a 0.

I /can/ set UDF to make an ai record go into UDF/INVALID alarm, but that
doesn't work on an ao record, where I have to write NaN to the VAL field
instead (the UDF field acts more like an output for an ao record).

My apologies to Andrew Starritt who was evidently using an ai record,
whereas Kellee and I were both using ao records.