|
Grex > Coop13 > #376: The problems with Grex, e-mail and spam | |
|
| Author |
Message |
| 25 new of 480 responses total. |
cross
|
|
response 305 of 480:
| 
Dec 14 22:19 UTC 2006 |
Regarding #302; So, is this your idea of ``encouraging words'' Mary? People
post honest criticisms of what they see to be problems and you post back
some snippy little comment like, ``But I understand why you can't do that.''
Do you really? Perhaps you could enlighten the rest of us, please? I mean,
I've witnessed you do your own fair share of misinterpretation over the
years. Perhaps you could give others the benefit of the doubt once in a
while?
Is this your idea of being more inclusive, of building a better community?
Is this your evidence that grex is doing just fine? Because from where I
sit, that sort of attitude reinforces my point that there is a serious
problem. If Todd misinterpreted you, why not go ahead and correct his
misinterpretation, instead of just sniping or poking with little barbs?
Do you have something to lose otherwise?
|
tod
|
|
response 306 of 480:
|
Dec 14 22:23 UTC 2006 |
I've got no problem with people pointing out my slaughter of punctuation.
|
cross
|
|
response 307 of 480:
|
Dec 14 22:28 UTC 2006 |
You quotation-killer! I spit at you!
|
tod
|
|
response 308 of 480:
|
Dec 14 22:29 UTC 2006 |
See, spitting at me is exactly what the staff does not need on its "TEAM."
|
mary
|
|
response 309 of 480:
|
Dec 14 23:12 UTC 2006 |
First off Dan, you need to calm down. Rage isn't healthy.
Todd has me saying something I wouldn't have said. Before I start
correcting folks, in detail, I think I'm going to wait just a little
longer until someone has me saying, in quotes, that Grex is perfect just
the way it is.
|
tod
|
|
response 310 of 480:
|
Dec 14 23:17 UTC 2006 |
"Al Gore invented the Internet"
|
keesan
|
|
response 311 of 480:
|
Dec 14 23:18 UTC 2006 |
We would like to be able to sign up a few friends with grex. One does not
like the email provided by his broadband connection (Verizon) and is playing
with sdf mail and would like to try grex. Another does not like the slow web
mail at the university over dialup, and besides dialup is ending Jan 2 and
they tell people to go find an ISP. He only wants email at home. And any
system which does not allow new users is going to die a slow death. A friend
of mine in Texas became a paying grex member this year before newuser
disappeared because he wanted a shell account for mail. Grex seems to be less
difficult to figure out than sdf.
There may be other U of M students and staff who want to use a shell account
from home without an ISP starting Jan 2. Could newuser be working by then?
Is anyone actually working on it?
|
twenex
|
|
response 312 of 480:
|
Dec 14 23:22 UTC 2006 |
Mary seems to be one of these folks who, unintentionally, just sound rude.
Considering htat in a high proportion of them that's caused by Asperger's,
and that a high proportion of aspergers' sufferers are technical people, I'm
surprised Cross isn't used to them.
|
cross
|
|
response 313 of 480:
|
Dec 15 00:33 UTC 2006 |
Regarding #309; Wow, that's really surprising. Calm down? Rage? I'm not
angry at all, nor have I been. Where did you get the impression that I was?
I *am* curious if comments like what you made to Todd are really what you
consider positive contributions, or the encouraging words you frequently
refer to. I mean, do you think that being rude is a way to push things
forward in a positive direction? Because you certainly seem to do things
like this often, at least in my opinion.
But for the record, I'm neither angry nor enraged. Just curious.
Also, maybe the one with the unhealthy rage isn't me, Mary. :-)
|
cross
|
|
response 314 of 480:
|
Dec 15 00:48 UTC 2006 |
Regarding #311; Like I said, there really ought to be a way for present
members to ``sponsor'' new user's (and thus create new user accounts).
But given the problems grex has with email, do you really think it's a
good idea to suggest to these people that they use grex as their sole
email provider? Just tell them to login to sdf and run pine or mutt instead.
|
gelinas
|
|
response 315 of 480:
|
Dec 15 00:58 UTC 2006 |
(Re-enabling newuser consists of:
Defining a pf rule set for the newest members
Tweaking the current pf rule set to only allow the "grandfathered"
users
Setting newuser to use the pf rule set for the newest members
Creating a way to move people from the restricted group to the
permitted group
I can take care of the first three, but the last is more complicated, since
it involves changes to the password file.)
|
keesan
|
|
response 316 of 480:
|
Dec 15 01:00 UTC 2006 |
I have accounts at both grex and sdf. Once they were both down at the same
time for a day. I have a fastmail.fm for emergencies like that.
|
cmcgee
|
|
response 317 of 480:
|
Dec 15 01:41 UTC 2006 |
In reviewing the staff meeting information posted in the past 24 hours, I'm
struck by cross's statement that meetings were held that he wasn't asked to
participate in.
I don't agree with Mary that having Ann Arbor-based staff meet more often
would solve the problems. Who is going to get them to meet more often?
Especially if it's a team that needs more slack in their personal lives. Who
is going to create that slack?
And how can you call them a team if they don't even tell another team member
about meetings? That is hardly respect if they are making decisions FOR
another member. Did they decide he was too busy? Too sleepy to participate?
Too disinterested to contribute? To new to know how things were done?
This is not a functional team that is trying to include new members. It does
not appear to be a team that has a good set of processes for bringing new
members in and getting them up to speed. And it does not appear to be a team
that has good ways of working with non-Ann Arbor members.
I don't think that the current problem-solving strategies that staff has
employed over the past year are working. I do think that there are ways this
staff could make changes in its processes that would help them be more
inclusive.
Clearly current staff care a lot about Grex. And just as clearly, the way
they've always worked doesn't work anymore. I would like to see the process
evolve into a growth and renewal pattern rather than a spiral of death
pattern.
|
gelinas
|
|
response 318 of 480:
|
Dec 15 02:19 UTC 2006 |
Staff meetings are scheduled by e-mail. It is _possible_ that Dan wasn't
in the 'right' e-mail group somewhere along the line.
|
gull
|
|
response 319 of 480:
|
Dec 15 02:20 UTC 2006 |
Re resp:270: I haven't read "The Wisdom of Crowds." I got kind of
turned off on that phrase when it started being used to argue that
things like Wikipedia were going to revolutionize human existance.
Re resp:277: For someone who claims to be transparent, you've been
awfully coy about your ideas for fixing the spam problem. "Let me on
staff and I'll reveal my secret plan" is not a very attractive pitch.
While keesan's filtering scheme is a nice effort, applying it globally
(as someone back there suggested) is not an answer. It's too specific
to the email *she* gets. If someone else used it, it would probably
miss more spam and throw out more legitimate mail than they would
prefer.
I'm going to reiterate that the email problem would be an excellent one
to delegate, either to one staff member or to a committee. Email is a
fairly stand-alone service; changes to it don't tend to affect other
parts of the system. I think there'd be no harm in letting someone
knowledgeable like Dan Cross take it on as a pet project and implement
an improved email system, as long as it's documented so other staff
members can maintain it later. I don't think anyone has emotional
involvement in the current email system the way they do the password
hash, so this should be politically fairly simple. It's baffling to me
that no one seems at all interested in taking him up on his offer.
|
cross
|
|
response 320 of 480:
|
Dec 15 02:39 UTC 2006 |
I'm not sure that grex staff has what can be said to be clearly defined
processes. Rather, there's a set of general guidelines for doing things:
try to get a concensus, work on what you're interested in, unless it steps
on someone else's toes. Some things are always okay for all staff members.
For instance, cleaning up stale accounts, deleting hacker tools or copies of
BNC or eggdrop or other large files. Some are okay for certain staff
members at all times. For instance, Marcus can more or less make whatever
changes to the authentication system or password database he wants. Steve
can more or less do whatever he wants (for instance, repartition the disks
during an upgrade without, I don't think, consulting anyone beforehand or
really making a plan to make sure we don't lose data). Some are okay for
some staff members some of the time. Remmers can make certain changes to
system configuration files in an emergency; usually he will go out of his
way to let people know that he's done and and solicit feedback, etc.
Let me reiterate because I think this is important: if you're one of the
what one might call ``principle'' staff members, you can more or less do
what you want and if someone objects you can later lay down an argument, but
do so knowing that, basically, there won't be any real consequences (for
example, under what circumstances would Steve ever get *his* root access
pulled? Now before someone jumps down my throat, I'm not saying that anyone
*should* pull Steve's access, but has anyone ever given any serious
consideration to what it would take for that to happen? Now what would it
take for someone to pull someone else's? We recently saw Mic's get yanked,
for instance. What would have happened if the roles had been reversed?).
It was also my sense that the idea of putting some processes in place just
wasn't going to fly. The usual argument against would be something along
the lines of, ``staff is busy; they have lives; they aren't paid; you can't
ask them to do something they don't want to do; be grateful they do anything
at all; why do you want to create more work?''
It's interesting to see some people other than myself say things that
basically echo my own sentiment of how staff operates: if you want to make a
controversial decision, odds are good you'll be told that it needs
discussion first, and then discussion will continue until you lose energy to
pursue it and just let it go. I feel like the password hash thing is a lot
like that: there's no good reason for sticking with what we have now, but
instead of changing, we're told we need to ``discuss'' it first and then no
one says anything until whoever proposed it (in this particular case, me)
gives up. If it's raised at some point in the future, the person raising it
(again, in this case, me) is told that it's been discussed, no one wanted to
do it, stop wasting everyone's time with endless debate, it creates friction
on the staff team to keep raising the same issues, etc. Eventually the
cycle repeats.
Here's another way to look at that particular issue: some people are
motivated to work on staff because sometimes they get to work on things that
interest them. That was my initial draw to grex: the technology that
allowed an open-access Unix system to actually work without imploding on
itself. When I first got on grex staff, I was sort of dismayed at what a
patchwork the software on the old Sun really was: in particular, the
password subsystem was *really* hacked together, very brittle, and very,
very easy to break (for example, making a textual edit to the password file
would *really* mess things up. Instead, Marcus or someone had written an
*interactive* tool for manipulating the user database. If you forgot and
ran ``vipw'' you would really, really screw things up). It was interesting
to me, with the move to the OpenBSD machine, to work on that problem, and it
eventually became clear that the BSD people had sufficiently evolved the
password subsystem in the standard distribution so that grex didn't need
customizations. To me, it was an interesting problem to figure out how to
get from our custom solution to the standard solution (which was necessary
on the Sun given the primitive nature of that system's password subsystem).
However, I found great resistance to doing that, because a lot of it was
Marcus's baby, and no one wanted to make him mad, even though he hadn't
really been active on grex in some time. Okay, fine, but *I* as a newish
staff member was still prevented from working on a problem that interested
me. Well, if that happens enough times, why would I *want* to continue to
do anything but the grudge work that no one seemed to mind anyone doing?
If every time you raise interest in working on some problem you're told no
because someone else ``owns'' it, even though they're not doing anything
with it, how long until you lose interest? Is that being thin-skinned? I
don't really think so, but I'm no psychology expert, so take what I say with
a grain of salt. I do think it's fair to say that it's not being inclusive
of the interests of ``new'' staff members. It is, in my opinion,
essentially saying that the interests of long-time grex staff members, even
if they're inactive, outweigh the interests of newer members, particularly
those that aren't local to Ann Arbor.
|
cross
|
|
response 321 of 480:
|
Dec 15 02:44 UTC 2006 |
Regarding #318; It's also possible that I was, but that the scheduling
notices got buried in the massive amount of spam that comes with the grex
staff email aliases. What discussions I *do* remember about the staff
meetings were usually along the lines of, ``can everyone meet at Steve's
place tonight at 7?'' ``Yeah, sure, I'll bring a pizza....'' Nothing about
how non-locals could dial into the meetings, no agendas, or anything. If
you don't read that email until 8pm, you're sort of screwed. There were
several times I recall seeing a post in the staff conference of the form,
``at last night's staff meeting...'' where there wasn't a lot, if any,
notice that there *would* be a staff meeting that night. I certainly never
participated in one, even though I would have liked to.
|
mcnally
|
|
response 322 of 480:
|
Dec 15 03:09 UTC 2006 |
re #318, 321: I was on staff for at least a year and I frankly
don't recall ever hearing about any staff meetings either. I
definitely never participated in one. I thought they were something
Mary imagined until comments from other respondents confirmed that
they believed in them, too.
Like Dan, it's also possible that I, too, might have missed e-mail
notifications due to filtering the ludicrous amounts of mail that
I got deluged with after being added to the staff mailing lists
(I haven't mentioned this before but I almost resigned from staff
less than a week after volunteering to help because the crap from
the mailing lists so disrupted my personal mail account.) Instead
I diverted that mail to gmail and archived it. A search through
the archived staff mail shows only one conversation where the
phrase "staff meeting" was used in mail that I received and on
that occasion (in 11/2005) no staff meeting was called (that I
was informed of..) My contributions as a staff member were
extremely paltry, but if there was a loop I was so far out of the
loop that I apparently wasn't even aware of its existence.
|
cross
|
|
response 323 of 480:
|
Dec 15 03:22 UTC 2006 |
Regarding #322; Hey! You're still staff! At least, you're still in the wheel
group and thus have root access.
|
gelinas
|
|
response 324 of 480:
|
Dec 15 03:27 UTC 2006 |
To the best of my knowledge, the staff has not met since you were added to
it, Mike.
|
rcurl
|
|
response 325 of 480:
|
Dec 15 03:35 UTC 2006 |
Is it possible that the technology has moved so far beyond Grex, and most of
the staff has moved with it, so it just isn't *interesting* to most of them?
|
keesan
|
|
response 326 of 480:
|
Dec 15 04:36 UTC 2006 |
Is anyone currently working on restoring newuser?
keesan's proposed spam filter is simply using spamassassin to filter on
anything with three spam points, plus a sample of how to put your friends on
a whitelist, which is much shorter than what keesan is using to filter her
own spam (which got all but one spam today) which probably uses less cpu time
because most spam is WIndows charset or the 60 stock spams a day and I put
those filters first.
|
cross
|
|
response 327 of 480:
|
Dec 15 04:45 UTC 2006 |
Regarding #325; I don't know about that; a lot of the stuff that grex uses
now is fairly modern (relatively speak, of course). Moving off of SunOS on
the Sun 4 was a huge leap forward in technology, literally catching us up by
about 15 years. In many respects, it's now on par with other systems.
Judicious use of the money that grex has in the bank could further improve
the technology new-ness situation.
Actually, in some respects, I feel like staff is holding grex back
technologically: a lot of things are being done like it's still 1991,
sometimes paradoxically. For instance, we're told at once grex hits its
hardware really hard, but at the same time told that we can't justify
something like hardware RAID. Some of the things that have become almost
automatic responses as far as system administration goes are sort of shoved
out the window. E.g., someone says, ``highly available storage'' you just
sort reflexively respond, ``hot-swappable hardware RAID.'' ``Reliable
memory,'' ``ECC.'' ``Chasis profile for colocation,'' ``rackmountable
case.'' ``Reliable backups,'' ``tape stacker unit.'' Certainly, some of
these things *do* make grex less interesting to those who might otherwise
be able to contribe really positively.
|
cross
|
|
response 328 of 480:
|
Dec 15 04:47 UTC 2006 |
Regarding #326; I really doubt it.
|
gull
|
|
response 329 of 480:
|
Dec 15 05:58 UTC 2006 |
Re resp:326: Ah, okay. I had thought you were still doing simple
keyword matching.
Re resp:327: Having worked with a tape changer, I'm not sure I'd have
called it "highly reliable." In four years the Overland 10-slot unit
where I used to work was out for repairs at least three times. Oddly,
the problem was usually not the robotics, but the Benchmark DLT1 drive
they fed.
|