Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!rutgers!ames!ucbcad!ucbvax!rca.COM!EVERHART%ARISIA
From: EVERHART%ARISIA@rca.COM ("GLENN EVERHART, 609 486 6328")
Newsgroups: comp.os.vms
Subject: Loginout checksums
Message-ID: <8708111334.AA25811@ucbvax.Berkeley.EDU>
Date: Mon, 10-Aug-87 10:02:00 EDT
Article-I.D.: ucbvax.8708111334.AA25811
Posted: Mon Aug 10 10:02:00 1987
Date-Received: Thu, 13-Aug-87 01:37:18 EDT
Sender: daemon@ucbvax.BERKELEY.EDU
Distribution: world
Organization: The ARPA Internet
Lines: 16

I understand (could be wrong though) that LOGINOUT is the same on
all vaxen except for the n-user key stuff on microvaxen. Thus a
finite number of checksums should suffice...
I also understand the n-user microvax version is the same as the
"normal" large-VAX version on which the max. number of users is
unlimited. Surely about 4 checksums isn't too many to support with
manual (that is, human) comparison to see that your copy isn't
corrupted. That also doesn't seem like too many possibilities
for a DEC patch to give new "correct" checksums to, since patches
are supposed to be installed serially up to a level, not in a cafeteria
style "one from column A, two from column B" mode.
	In the future a checksum that permitted some known areas of an
image to be ignored (where they are known to change harmlessly) could
be a mighty handy thing to have. But DEC will have to supply it, as
a user-written one would be non-standard and therefore not usable for
security checksumming. VAX menu item, anyone?