[527] in SIPB bug reports
One more thing....
daemon@ATHENA.MIT.EDU (daemon@ATHENA.MIT.EDU)
Sun May 14 02:16:06 1989
Date: Sun, 14 May 89 02:15:27 EDT
From: Jonathan I. Kamens <jik@Athena.MIT.EDU>
To: tytso@ATHENA.MIT.EDU
Cc: bug-sipb@ATHENA.MIT.EDU, sipb@ATHENA.MIT.EDU
In-Reply-To: Theodore Ts'o's message of Sun, 14 May 89 01:44:18 EDT <8905140544.AA07736@THOR.MIT.EDU.MIT.EDU>
First of all, most of those people agreed early on, before Stan and I
raised some problems with your proposal. You attempted to answer those
objections, but on one else was given a chance to give their opinions
about the flaws in your proposal.
My original message was sent out at 4:51 on Thursday.
Stan's original message stating his objections was mailed at 5:49 pm on
Thursday. Your original message was mailed at 7:19 pm on Thursday. I
installed the new binary at 5:12 on Friday, giving a little under 22
hours for anyone to agree or disagree with your objections or to
"change their vote."
Thus, I believe your claim that we
had reached a "SIPB policy decision" to be rather premature. (Is there
such a thing?) In addition, you didn't even breathe a word to any
mailing list after the objects were raised that you had single-handedly
declared the period of discussion over and that you were going to
implement your proposal.
I talked to Rob and told him that I was going to install the new
version, and that if the decision was changed at the meeting on Monday
night I would immediately put the old version back and make the
sources readable again.
I did not send out notice to sipb because I did not have time --
Friday night services started at 5:30, and, as I have already pointed
out, I installed the binary at 5:12. I didn't even have time and go
home and change for services. I apologize for not notifying people,
but I figured that they would notice when they ran the program....
Is the SIPB really going to be into restricting access to sources? Are
we going to censure anyone who transmits information we don't think is
"appropriate"? I hope not. If I had offered to go and find a version
of xscreensaver from the net, would you have been so incensed? Is the
SIPB starting to regulate the actions of its members? I hope not.
I *agree* that this is a short-term solution that the SIPB should not
seek to repeat in the future. I *agree completely* that a better
solution will have to be found, and that that better solution should
not in any way involve hiding the sources public.
Not one person argued with my contention that this is a time-sensitive
issue, and that the problem of people locking their workstations for
too long really does need to be solved now, as opposed to next week or
after the summer when I have time to implement some better solution,
like a logout button.
Yes, if you had gone to the net and gotten a version of the sources
from there, I would be just as angry, because that still, in my
opinion, goes against a decision reached by the SIPB. Certainly, I
would agree, less so than copying the sources out of sipbsrc, but I
would still be upset.
Jik, jik, believe me, I've been there before. I *worked* on the X10
version of xscreensaver. Back then, SIPB attempted to keep the sources
of xscreensaver private. It failed miserably. It just doesn't work.
It will, in my opinion, work for a week or a week and a half just this
once. That's all I intended, and I believe I've said that several
times in my messages to SIPB.
In any case, the case may be academic. Tom Copetto now has a copy of
the sources. *And* he has access to a locker with substantially more
than the 600k quota that you assume everyone has. I suppose I won't
have to give a copy of the sources to Ron; he's made my point for me.
No, but you said that you would, and that is what I objected to. The
ls locker is another issue entirely -- if Tom wants to make another
version of the xscreensaver available, that's his prerogative; if he
wants to publicize it widely, that's also his prerogative. But I
don't think that will make much difference in the next week and a
half, which is the time period in which I am expecting my solution to
do some good.
For heaven's sake, people can get around the timeout just by setting
a longer timeout (up to eight hours), and then they don't even have to
use Tom's version! I *know* that changing the binary won't stop
everybody, and I've *said that* before. I have only claimed, and
still claim, that a lot of the people who use xscreensaver will not
find out how to change the timeout before the end of the term.
There's no way we can no whether that's true without trying it, can
we?
I won't belabor the arguments that have already been made against your
changes to xscreensaver, but I think they deserve to be listed once
more:
* It makes SIPB into a fascist organization. You've already
started dictating what members can and can not do. It means we have to
be fascist with sources. Sort of makes us hypocrits, when we've been
pressing other people (NeXT, Apple) for an open source policy.
I have not "started dictating what members can and can not do," give
me a break. It is not I who says that sources that are only readable
to gsipbbin should only be read by gsipbbin people, it is the *fact*
that gsipbbin exists that says this.
* It's not fair for people with private workstations. Remember
SIPB != Athena. The users of public workstations are not the only
people we provide software for. For example, media-lab used to receive
the SIPB locker as well (which is why everything is compiled for
/usr/sipb). Your changes certainly would not be appropriate for them.
(Although granted I don't think they've updated their software from us
in a while.)
The problem of people occupying workstations in the clusters by
locking them is a serious one. I acknowledge that people with private
workstations will lose for a week. As I have said before, I believe
that we have to accept this in order to solve the aforementioned
problem for that aforementioned week.
If you really feel that the private workstation users will lose
greatly because of the timeout, I will remove the maximum timeout,
since an 8 hour maximum timeout is just about the same thing as having
no maximum timoue, this making it possible for private workstation
users to unset the timeout, and making the xscreensaver in place now
effectively the same as the old one except for the default timeout.
Is that an acceptable solution?
* It encourages people (such as Tom Coppeto) to keep private
copies of xscreensaver around, thus fragmenting our control of what
people run. Don't assume that MIT students are dumb. They aren't.
I have made no such assumption. People are *always* going to find
ways around things like timeouts (like the "Elapsed: 0:05" thing that
I described in my original message about that), there's nothing we can
do about that. However, for the n'th time, I believe that putting in
the time out will solve enough of the problem to make it worthwhile.
That contention has yet to be disproven, despite repeated claims that
"lots of people have copies of the binary." I don't believe it.
* Because of the above, it doesn't work! After all of these
disadvantages, it doesn't do anything but provide bad PR and spread
dissention.
Prove it. Go into the w20 cluster on a night when it is full and find
out how many people are running a version of xscreensaver other than
the one in the SIPB locker, or are running the one in the SIPB locker
with a larger timeout. Permit me to say that if the timeout keeps
even *one* person from locking the screen for too long, then we have
provided one more workstation for someone to do real work on during
the last week of class. That's better than leaving things the way
they are.
The most distasteful part of this is that your arrogance about the whole
thing --- your arrogance about not giving any notice that you had
considered the discussion closed; your arrogance in how you presume to
speak for SIPB. If I thought your views reflected the views of SIPB I
would have left this organization a long time ago.
I have already apologized for not sending out a message when I
installed the new binaries, so I won't go into that. As for my
"arrogance in how [I] assume to speak for SIPB," what else speaks for
SIPB than a majority view of its members? I put the question up to
discussion because I *would not* make the decision on my own -- if
there had even been a *close* vote in the discussion, I would not have
installed the binaries. 9 to 6 is not a close vote, in my opinion.
Furthermore, non-members on Athena staff whose opinions I value (and,
I feel, are relevant) also were in favor of the timeout.
Oh, and one more thing which you have not addressed at all. You
accused me of implementing "SIPB policy," and then said that we should
not implement policy that Athena hasn't even made. I responded that
Athena *has* made policy about this. You ignored that. I assert that
the fact that Athena has a 20-minute lock rule goes quite a ways
toward justifying the installation of a timeout.
- Ted
P.S. I suppose that if you were in charge of Kerberos you would have
used xor encryption and then persecuted anyone who had the audacity to
copy the sources or who tried to make a replacement program which broke
your inadequate implementation?
This analogy is flawed in so many ways that I won't bother to list all
of them. However, let me just point out that if I were in charge of
Kerberos and one of my *employees* copied the sources and/or tried to
make a replacement program which "broke [my] inadequate
implementation," I would fire him. How does that analogy strike you?
