|
Grex > Coop11 > #84: outgoing internet access for non-members | |
|
| Author |
Message |
| 25 new of 127 responses total. |
prp
|
|
response 93 of 127:
| 
Apr 2 06:12 UTC 1999 |
Re 91: Given the restrictions that would still be in place, I just
don't see the demand. Where would it be coming from?
Remember that Grex has 14.4K modems with an effective throughput of
between 2400 and 4800 bps.
|
keesan
|
|
response 94 of 127:
|
Apr 2 23:41 UTC 1999 |
What is an effective throughput? I find my 14.4 modem to run grex several
times faster than my 2400 bps modem.
|
prp
|
|
response 95 of 127:
|
Apr 3 00:11 UTC 1999 |
I think you will find that "several times" is about two times.
Effective throughput is 10*n/t where t is a time period in seconds,
and n is the number of characters actually sent. Result is in bps.
To be fair this should be measured over a time where there is no
waiting for keyboard input.
Note: My Modem shows "CONNECT 19200" when connecting, but then
stty shows "speed 4800 baud; line=0". So it maybe that there is
a 4800bps bottleneck somewhere.
You can test this with a stop watch and a medium sized file, say
20 screenfulls or so.
|
davel
|
|
response 96 of 127:
|
Apr 3 14:25 UTC 1999 |
I don't think so, Paul (the 4800 baud bottleneck). You connected at 19200
to the terminal server; from there it's ethernet, and the baud rate shown on
Grex itself is meaningless.
|
prp
|
|
response 97 of 127:
|
Apr 4 03:03 UTC 1999 |
Try "cat /c/p/r/p/screen/16" from a dial-up line. Whenever I do, it
takes about 110 seconds, which is about 2800bps; I'm ignoring the time
Grex got into 110bps mode, as I can not replicate it.
Interesting enough the bottleneck is only one way, from Grex to me. It
wouldn't be much of a problem the other way around; there aren't very
many people who can type 2800wpm.
|
rtg
|
|
response 98 of 127:
|
Apr 6 19:38 UTC 1999 |
re resp:92 I thought it was aruba who was frequently pushing recycled,
self-made envelopes.
back on topic: In general, I would like to support more open access.
However, I do not wish to become an anonymizer service, so I would support
blocks to any inbound/outbound use of the same protocol. I do not see a
legitimate use for even paying members to telnet in, then back out again.
Considering that we have such limited storage space, FTP use by a remote
telnet user, would result in a second FTP session to send it back out, So I
think that we should encourage people who have the capability to FTP directly
to their own machine, and not be relaying thru us. It is ONLY the dial-in
users who have a legitimate reason to use ANY outgoing TCP/IP service. By
definition, if you can make a TCP/IP connection to grex, you can run your own
TCP/IP clients.
What would it take to change our kernel blocks so that internet access,
including http, be available only to those connected thru our terminal
server, or directly connected to our local ethernet, regardless of membership
status? This is the proposal that I can support. This will focus our
resources on our mission: to provide an open-access conferencing system, and
to provide computer and internet experience and education to members of the
local community who cannot afford it elsewhere.
|
ryan
|
|
response 99 of 127:
|
Apr 6 21:27 UTC 1999 |
This response has been erased.
|
mary
|
|
response 100 of 127:
|
Apr 6 22:09 UTC 1999 |
I wouldn't mind losing membership money that is only
given for the perks. The fewer of those folks voting
on Grex policy the better for Grex, long haul.
Nice response, Rick.
|
dpc
|
|
response 101 of 127:
|
Apr 7 18:20 UTC 1999 |
I'm appalled to learn that I have the highest *total number* of
responses in Coop (at least of the top 20 "should index" people.
Hm. Maybe I should save up my words to make fewer, but lengthier,
responses...
|
prp
|
|
response 102 of 127:
|
Apr 7 19:38 UTC 1999 |
Re 99: The only limit that would be placed on members, is a prohibition
on loopback connections. That's no problem. Anybody who is using them
has either made a silly mistake, or is trying to generte load on the
Cybersapce Internet line.
|
aruba
|
|
response 103 of 127:
|
Apr 7 22:07 UTC 1999 |
Actually, Dave, you need to look at the full list to see who has entered the
most total words in coop. Here is the first page, when sorted that way:
User Resps Lines Words Shlds Shoulds/Word
-------- ----- -------- ---------- ----- ------------
steve 355 3627 31176 69 0.0022
aruba 301 3247 26277 74 0.0028
janc 171 3050 24668 89 0.0036
rcurl 305 2149 19092 56 0.0029
mdw 122 1678 16650 34 0.0020
keesan 161 1612 16263 19 0.0012
remmers 207 1730 13367 37 0.0028
scg 137 1264 13121 37 0.0028
devnull 89 828 7485 18 0.0024
scott 97 1011 7356 11 0.0015
valerie 95 768 7111 11 0.0015
mary 80 855 7001 19 0.0027
mta 104 873 6910 12 0.0017
jep 64 641 5627 20 0.0036
cmcgee 88 582 5261 11 0.0021
davel 97 548 4711 15 0.0032
richard 54 504 4391 27 0.0061
krj 62 559 4363 15 0.0034
rtg 38 443 4212 5 0.0012
srw 45 447 4090 21 0.0051
dang 52 563 3987 3 0.0008
dpc 110 503 3965 23 0.0058
Now try it on the last Agora - three guesses who's first. :)
|
hhsrat
|
|
response 104 of 127:
|
Apr 8 01:54 UTC 1999 |
You mean I don't even make it onto the first page of the should list? I
thought I'd said should at least 5 times in this coop.
When I tried to run the little program, it froze on me. How should I do
it (and when) to avoid problems?
|
aruba
|
|
response 105 of 127:
|
Apr 8 03:25 UTC 1999 |
It takes a long time to run. What I do is run it in the background and pipe
the output to a file, like this:
!~aruba/bin/wordcount should coop11 > shoulds &
|
prp
|
|
response 106 of 127:
|
Apr 8 03:54 UTC 1999 |
I get "/bin/sh: ~aruba/bin/wordcount: not found"
|
davel
|
|
response 107 of 127:
|
Apr 8 12:05 UTC 1999 |
"~" isn't supported by sh. Try
!/a/a/r/aruba/bin/wordcount should coop11 > shoulds &
|
aruba
|
|
response 108 of 127:
|
Apr 8 20:43 UTC 1999 |
Sorry - Dave's right.
|
devnull
|
|
response 109 of 127:
|
Apr 11 04:46 UTC 1999 |
Re #98: I find it convinient that I can use cvs for outgoing internet
access for grex. Your proposal would cut off a service that I find
useful.
|
keesan
|
|
response 110 of 127:
|
Apr 11 15:50 UTC 1999 |
Unlimited outgoing access for paying members, and for dial in nonpaying users?
|
ryan
|
|
response 111 of 127:
|
Apr 11 17:16 UTC 1999 |
This response has been erased.
|
prp
|
|
response 112 of 127:
|
Apr 14 23:26 UTC 1999 |
Re 109: How does the Concurrent Versions System provide outgoing Internet
access?
|
prp
|
|
response 113 of 127:
|
Apr 14 23:42 UTC 1999 |
Re 97: There was indeed a 4800bps bottleneck somewhere. It turns out
it was my modem.
|
lilmo
|
|
response 114 of 127:
|
Apr 16 22:39 UTC 1999 |
Re resp:113 - funny how that works.
Re resp:103 - I wonder if you ran that on certain older editions of ccop if
I would show up as fairly "response"ive. :-)
|
aruba
|
|
response 115 of 127:
|
Apr 17 04:39 UTC 1999 |
Feel free to try. :)
|
devnull
|
|
response 116 of 127:
|
Apr 30 06:30 UTC 1999 |
Re #112: CVS can be used in client/server mode. I find it convinient
for keeping my dotfiles in sync across different machines.
|
dang
|
|
response 117 of 127:
|
May 6 02:50 UTC 1999 |
I quite like how Grex is now. I like that there is a definate time
limit on outoging access, after which a person has to prove they're
still alive (and if they pay by check, they are re-verifying). I like
that outgoing services, unattractive as they may be for various
technical reasons like limited bandwidth, are made even more
unattractive by the necessity of sending ID *AND* paying. I don't do
very much of the security work now, but I'd hate to see what it would be
like if we were a more attractive way-station. I'd hate to see how our
link usage would be if more people could fairly easily run ftp from
here. Finally, I know that sending money is a *much* larger deterant
for me trying something that just sending information. I'm glad, as I
don't think giving access is a large part of Grex's mission. I like
that Grex is a destination, not a pitstop.
|