Hi Will,

 WH> In another msg. you refer to the X, 1, 2, 3 sequences.  If you make
 WH> several archives you will find that the 1 -3 entries appear to be a
 WH> used to refer to 3 'buckets' where a new archive replaces the oldest
 WH> (3) one and becomes 1 while the remaining 'buckets' (1 and 2) are
 WH> incremented.  Sort of a first-in/first-out stack of 3.  Note: I may be
 WH> bass-ackwards on the order.  I say 1 is the newest but it may really be
 WH> 3; I haven't messed with this in a LONG time.

This sounds sort of like an old IBM mainframe terminology from the 1970's,
80's.....;-) GDG's, Generation Data Groups... The latest version of a file is
0, previous version -1, prior to that -2 etc, for as many archives as you wish 
to keep. When its time to update to the latest version, input file is 0,
output is +1. If the update succeeds, then EVERYTHING renumbers down 1 when it 
completes leaving the latest at 0... 

It was also possible to "cycle" the numbers early, so that 0 became the new
space allocation of the series that is empty, and updates are done FROM -1, TO 
0. While this was easier to handle re-runs (frequent in those days), it means
the latest file may not contain valid data until the update worked. 

Many years ago (before I moved to OS/2 even) I wrote 2 small DOS & OS/2
utilites that performed something similar (CYCLE and GDG, one Cycles DOWN in
number, the other GDG's UP) and I used those on the BBS to replicate the
Mainframe methodology. Each week (actually on the 1st, 6th, 11th, 16th, 21st,
26th of the month) I "Cycle" the BBS log files and keep 8 weeks of files as
backups. Each month I combine the oldest months data and archive it. This way
the BBS log files never get too large to manage and I have all the records of
each month archived for historical purposes (I run the BBS for a club).

Just another useless piece of info.........;-)


--- Maximus/2 3.01