|
|
| Author |
Message |
| 25 new of 239 responses total. |
carson
|
|
response 100 of 239:
| 
Jan 18 06:34 UTC 1996 |
I forgot one day that I'd left mine on for when my messages are
also on. Usually, I have my messages off, which excludes me from
being a helper.
|
rcurl
|
|
response 101 of 239:
|
Jan 18 08:10 UTC 1996 |
Another data point tonight: all network ports were busy for several
tries, so I dialed in, and got connected first try. Now, why was it
we added more modems?
|
robh
|
|
response 102 of 239:
|
Jan 18 11:41 UTC 1996 |
Funny, I was in the Dungeon from 11 to 11:30 last night,
and at one point all eleven modems were in use.
|
carson
|
|
response 103 of 239:
|
Jan 18 12:48 UTC 1996 |
All network ports busy again this morning. It wouldn't hurt to add
more modems, if only to encourage the local base that the consensus
wishes to focus on.
|
popcorn
|
|
response 104 of 239:
|
Jan 18 13:30 UTC 1996 |
Re 101: I got e-mail last night from someone who had attack dialed for an hour
to get in. So the modems are definitely being used. <valerie contemplates
doing a "!ttyuse" to check in more detail>
Re 100: At some point, people were asking Jan if it was possible to be a
helper but not accept non-help-request messages. I'm not sure if that's a
current feature or not, but with that combination of settings you might one
day find that you receive help requests.
|
popcorn
|
|
response 105 of 239:
|
Jan 18 13:34 UTC 1996 |
By the way -- I've been bumping people with more than 1/2 hour of idle time,
once or twice a day, and I sent mail to jfk on Arbornet to ask about getting
a copy of M-Net's idle-killer.
|
carson
|
|
response 106 of 239:
|
Jan 18 13:50 UTC 1996 |
Is that to go along with the orphan killer?
You know: big guy, hates kids, wields axe...
|
davel
|
|
response 107 of 239:
|
Jan 18 17:33 UTC 1996 |
When I logged in this morning, I'd tried attack telnet for a while, then
dialed in. But there were only 39 people on, including me. I wrote Valerie,
who mentioned this item. She also mentioned killing a couple of people with
lots of idle time just before that. I don't know what the telnet-port limit
is, off hand, but was it really reached? I suspect some other problem.
I also note that a little while later Grex crashed. Just FWIW.
|
kerouac
|
|
response 108 of 239:
|
Jan 18 20:22 UTC 1996 |
I hate m-net's idle killer...sometimes I like to log on grex and leave
that window open while I'm waiting for someone to log on, and go do
something else. The idle killer is annoying. If you go to the
trouble of logging on to mnet or grex, waiting your turn to get on, you
ought to be able to go to the bathroom, or step outside for a smoke,
or take a phone call, without being logged off the system. I hope
idle killer is discussed at a staff meeting before being installed here.
I'd understand if it was set to kick those who have been idle for an
hour, but anything less is unfair.
Also if Im writing a letter or doing a file, I could easily be
procrastinating in front of my screen thinking of what to type, and some
idle killer would think I've left. Its just not necessary.
|
rcurl
|
|
response 109 of 239:
|
Jan 18 20:27 UTC 1996 |
Further to all ports being busy last night - when I got on via modem,
I saw three users with 2 to 5 hours of idle time. Sure, put in an idle
time killer (but with enough slack for kerouac's potty run). By the
way - I've noticed that UM (MTS) has inaugurated time limits on net
use. Two hours.
|
abchan
|
|
response 110 of 239:
|
Jan 18 21:29 UTC 1996 |
Idle killers are a good idea. Once I forgot to log off and went to
sleep and the next day I saw that I was on the "look who's logged on"
list at 3 am! I'd feel bad if someone else couldn't get on because
I had forgotten to log off. Of course it has to be long enough so
people don't kicked off when they're just lagging or running to go
get the laundry. Two hours is good.
|
kerouac
|
|
response 111 of 239:
|
Jan 19 01:30 UTC 1996 |
I guess if idle time is set at two hours or something it would be ok,
although I wonder if it would affect enough logins to be worth installing
at that setting. I just dont like how mnet has it where you start getting
bugged at twenty minutes and logged out at thirty minutes. I also dont
think staff should be manually logging out idle logins at thirty minutes
or whatever when there is no official policy.
|
janc
|
|
response 112 of 239:
|
Jan 19 02:10 UTC 1996 |
I think 30 minutes is very generous. Plenty for a potty run.
|
kerouac
|
|
response 113 of 239:
|
Jan 19 02:26 UTC 1996 |
janc, WHY is thirty minutes generous? specify and explain, if you
actually have an explanation. If someone spends twenty minutes trying to
get in, he ought to have at least an hour. Most free nets I know of
wont bother you for an hour. If there is a specific technical reason why
thirty minutes is better than an hour, or two hours, I'd like to hear it.
If I am in the middle of a letter, and I get a phone call and walk away
from the computer for a while, I should not be logged off. There should
be more leeway than a 30 minute idle time that starts paging you
constantly at 20 minutes allows.
Also, just on principle, this is like the situation with the changed
party settings that got people upset. Things like this should be discussed
and if necessary voted upon before being implemented.
|
kerouac
|
|
response 114 of 239:
|
Jan 19 02:54 UTC 1996 |
I have an idea. Is it possible, if Idletime is implemented at 30
minutes, to put a command line in a file that would cause idletime
to read your login as active and not bother you? That could be an
option for those of us who feel responsible enough to log off our
logins when we arent using them without being babysat.
|
scott
|
|
response 115 of 239:
|
Jan 19 03:21 UTC 1996 |
Well, we likely wouldn't have quite as much contention jover prots if people
weren't idle for so long. Check out the "w" listing sometime, which gives
idle. Some people are stuck, or just plain forgot.
|
popcorn
|
|
response 116 of 239:
|
Jan 19 04:15 UTC 1996 |
There was a discussion a while back about idle timeouts on the modems. The
new GVC modems all use a timeout of 20 minutes of no activity to bump people
off. People discussed that a fair amount, to get to the 20 minutes number.
My personal feeling is that it's not very hard to hit Enter once in a while
if you want to stay connected. If you're not even *at* the computer for 30
minutes, or if you're spending 30 minutes thinking about what to say, you're
using up 30 minutes of a scarce resource that someone else could have put to
better use.
|
scg
|
|
response 117 of 239:
|
Jan 19 05:06 UTC 1996 |
re modem usage:
I'm sitting here Grexing from the Dungeon, since I stopped by and
decided I'd never be able to get in from home, considering how busy Grex was.
Looking at the modems, all but one are in use. All were in use a few minutes
ago.
|
popcorn
|
|
response 118 of 239:
|
Jan 19 05:34 UTC 1996 |
Ok, that explains why hoot had 3 hours of idle time on console and then
suddenly became active again. Hey, I take it this means you've now got
keys -- cool!
|
tsty
|
|
response 119 of 239:
|
Jan 19 06:23 UTC 1996 |
this situation, from the earliest entries til now, has had the
unfortunate overtones of ... turning off newuser ... which grex
will not tolerate.
Open newuser is a philosophical bedrock of grex. grex is a unique
system on the planet which allows a comfortable home, a friendly
place to learn and grow. Everyone has to tolerate the burden of
an open login policy even as it affects the maximized efficien
efficiency of the system. Forever-open newuser is far too importnat
a philosophical underpinning to mess with.
Just think, cyberspace for everyone who is willing to try and wiling
to learn and willing to help provides a freedom and an independence
unequalled since the Declaration of Independence
Constitution
Bill of Rights
were enacted - and those philosophical underpinings strain the
society as a whole justas well.
|
scg
|
|
response 120 of 239:
|
Jan 19 08:15 UTC 1996 |
Actually, the hoot on the console that suddenly became active again was robh,
who was there too, but I do have keys now.
set drift=off
|
remmers
|
|
response 121 of 239:
|
Jan 19 11:52 UTC 1996 |
When you do a !finger at a random time, what's the distribution
of idle times that you see, typically? I did it just now and got
the following:
time idle % of jobs
--------- ---------
> 2 hours 6%
> 1 hours 15%
> 30 minutes 24%
> 15 minutes 27%
(Each % figure includes all the earlier ones, which is why they
increase.)
I don't know if that distribution is typical or not, but looking
at such distributions would be a way of arriving at a reasonable
setting for an idle killer. For example, if usually 15% of the
have been idle for over an hour and 24% for over 30 minutes, and
reducing the number of telnet logins by 15% would be enough to
keep people from getting the "all ports busy" message very often,
then there'd be no particular gain in setting the idle time at
30 minutes over setting it at an hour, so you might as well set
it for an hour.
|
scg
|
|
response 122 of 239:
|
Jan 20 07:01 UTC 1996 |
Could the idle killer be set to exempt staff people? I ask this not to get
a perk for the "inner circle", but because staff people sometimes have
legitimate reasons being idle, like compiling something that takes a while.
Another approach to that might me a command that staff people could run to
exempt themselves if they were doing staff things that required idle time.
|
carson
|
|
response 123 of 239:
|
Jan 20 09:31 UTC 1996 |
Why just staff?
|
gregc
|
|
response 124 of 239:
|
Jan 20 10:30 UTC 1996 |
Because sometimes something goes wrong with the system and if a staff
person can't get access to the machine, who's going to fix it?
I've been thinking about this one, and I had an idea on how to do this
somewhat more fairly. Instead of exempting all staff, we could make
a system that would leave at most 2 staff logged in.
It would work like this: when the idle killer decided it needed to bump
idle peple off the system, it would make a list of people to bump and
then it would exempt any staff person if there was only 1 or 2 staff
logged in. If there was more than 2 idle staff people it would randomly
select who gets bumped and who stays. The reason I want 2 rather than 1
is that it's possible the 1 you leave logged in may not be available.
Leaving 2 leaves a better chance that 1 of them will be available to
fix a problem.
|
