>Hi Al,
>
>And thanks for taking the time to do the experiment and write me back.
>
>You're right, there may be some sort of justification for that FINAL behavior. But was it your experience (it WAS mine!) that the wait window showed movement from the top of the table to the bottom, then from the bottom to the top, and it displayed twice for records that passed the selection criteria, and it also displayed random movement before it settled on the top record of the browse.
>
>Weird!
>
>David

Our tests were different in some areas... you appear to have used a table with a PK, I used a free table without one. You used VFP6, I used VFP5. The "pattern" of record pointer movement was different, which will ultimately reflect differences in VFP internals. IMO, these internals aren't too important here. Suitable for conversation at a VFP DevCon cocktail party, but that's about it! :-)

One thing I wanted to mention earlier but forgot: if you want to use UDFs in that fashion, it's better to pass parameters, e.g.

BROWSE ALL FOR MyUDF(RECNO())

so that if you inadvertently move the record pointer in MyUDF(), you'll still get the correct results. This is similar in concept to the use of parameters in UDFs called within a SELECT - SQL. However, in that case, passed parameters are about the ONLY things you can rely on.