Conference noted::hackers_v1

	My guess is that it is a bug.  There is some bit you twiddle to
tell RMS to give you 'password' instead of the real password.  Off hand,
I don't remember what it is.

	I mistakenly tried the SET DEF scenario once.  Was it ever awful.
There are a ton of things you type that require bizarre network processes
to start up.  The overhead is enourmous.   Although since RMS provides
transparent network access, most things do work.

	KCz

I think I remember seeing somewhere that VMS does not support setting your
default directory to a remote node.  I think it was the QAR database.

				JENelson

It's not really a bug, it's just that SHOW DEFAULT isn't as smart as the
rest of version 4 of VMS.  It simply translates SYS$DISK, and tacks on the
current default directory.  What it needs to do is check for search lists,
and this and that, all of which RMS can be persuaded to do on a $PARSE.
It's not too hard, but neither is it trivial; and the payoff is low.

For what it's worth, after accessing a few files with that default directory,
RMS might remember the password, so try:
	$ SET DEFAULT XXX"USERNAME password"::DISK$:[USER..]
(Yes, "password" in lower case, just like that).

Also, what's so bad about SHOW DEFAULT showing the password?
SHOW LOG SYS$DISK would also display it.

I have at times set def'd over the NI from a Microvax I to a VAXcluster, and
although it *is* slow, it also is completely transparent (so far as I can
tell) and it can be a convenience at times.... unless you get
confused and then the world seems to end.
	
Why worry about little things like "sho log"?  Just uparrow to 
see the text (unexpurgated, too) of the last twenty things you did.
	
VMS may need a flag to tell it to automatically NOT remember for 
uparrowing certain commands (or to sanitize the displayed versions
of those commands).

	A slight digression, but a worthy one, as it describes a kludge:

	If you want to type commands that uparrows (and ^B) won't remember,
do them like this:

		    $ @tt:
		    _$ type FOO"BAR BAZ"::XYZZY:[PLUGH]HELLO.SAILOR
		    _$ exit	! Or ^Z.  Whatever.
		    $

The command recall doesn't recall anything inside the "command procedure," tt:.
		<_Jym_>

Hey, what a super kludge! Thanks, Jym!

Burns

Re "password": From TOPS-land, now that you can do a parse-only GTJFN%
on a filespec containing node and userid/password attributes, an

	@INFORMATION (ABOUT) FILES

command at EXEC level returns you the real password along with everything else.
I believe this (and the SHO LOG VMS problem) are security weaknesses. The 
argument that since you just typed the password it's ok to echo it back (it's
ok for YOU to know it, right?) could just as well work the other way (YOU
know it, so why splash it on your terminal for others to see?)

.3> . . . after accessing a few files with that default directory,
.3> RMS might remember the password . . .

	I've noticed this, and it's a real nice feature.  What
are the conditions for RMS to remember the password?  Are they
documented anywhere?
		<_Jym_>