Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watmath!clyde!cbosgd!ucbvax!eg.ti.COM!BOLTHOUSE%MCOPN1
From: BOLTHOUSE%MCOPN1@eg.ti.COM.UUCP
Newsgroups: comp.os.vms
Subject: accounting and $GETs from source
Message-ID: <8705280149.AA03648@ucbvax.Berkeley.EDU>
Date: Wed, 27-May-87 11:02:00 EDT
Article-I.D.: ucbvax.8705280149.AA03648
Posted: Wed May 27 11:02:00 1987
Date-Received: Fri, 29-May-87 01:28:33 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 38


VMS print accounting is somewhat unusual in that it doesn't actually
count newlines or other 'magic' characters to make its decisions.  Your
problem with the number of pages printed is the exception to that rule.
If your device handles pagination itself, then the symbiont only sends
out one *very* large page (i.e., /NOFEED).  The symbiont duly reports to
the job controller that it printed one page.  That's what the job controller
then drops off into the accounting file.  Note that it's not accounting
that's messed up, but the symbiont!  This is fixable if you know the ins
and outs of symbiont alteration.  Those I know who have mucked with the
symbiont tell me it's not that tough, but I haven't done this myself.

Mucking with the symbiont will also resolve the problem of having the
number of lines incorrectly reported.  You can count the number of lines
yourself, and report that back to the job controller as the number of
$GETs from source.  Admittedly, this contradicts the intended spirit of
the number, but the point is, you can make accounting be what you want it
to be if you work at it hard enough.

The reason the number of $GETs from source is different with the 2 devices
is because (I think...) the DMF-32 based printer can grab/buffer more data
at a time than the LP11 based printer.  This, I believe, is also something
the symbiont can recognize.  The LP11 is the most inefficient way I know
of to run a printer (no...I suppose you could run one serially off a DZ11.
Yuck!).  Again, if you want to know how many lines were printed, you should
count newlines yourself in the symbiont.

Has anyone any ideas how we're ever going to handle accounting for printers
that understand page-description languages (i.e., PostScript)?  Page by
page seems easiest, but I don't know if that's possible.  Just a random,
passing thought...

David L. Bolthouse
Texas Instruments Defense Electronics Information Systems VAX System Support
McKinney, Texas

Ma Bell:  214-952-2059
Internet: bolthouse%mccore@ti-eg.com