Grex Coop10 Conference

Item 11: The Medium-Range Planning Item

Entered by valerie on Fri Jun 27 06:34:09 1997:

valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 Valerie Mates valerie Jan  8 05:10:55 2004 Val
valerie Jan  8 05:10:55 2004 
54 responses total.

#1 of 54 by valerie on Fri Jun 27 06:46:37 1997:

This response has been erased.



#2 of 54 by mdw on Fri Jun 27 08:53:53 1997:

If you want more excitment, find a way to stick people into interesting
conferences when they join.  That why I fought hard on m-net, & to some
extent here, to have "general" or "agora" be the default.  What I think
would be even more interesting, and something I've always wanted to try,
is to have newuser read through what people say interests them, and
generate a .cflist file of conferences.

Generating more publicity, and getting grex better known, could mean a
*lot* more people, and a lot more demand for the system.  How do people
feel about having 200 people on at a time on grex?  Are we ready to
think about issues such as running out of UID space, kerberos, Solaris
2.6, DFS, etc?


#3 of 54 by aruba on Fri Jun 27 19:06:20 1997:

What I'd like to see most is more conferencers.  I think that will result in
more members, which is what I would like to see second-most.  In order to
pay for a lot of things Grex would like, we need more members.  Period.
(Oh, I suppose we could search for alternate forms of income, but I hope
we don't do that, because if we become dependent on such sources, then we
become beholden to them.)  The way to get more members is to have more
content, in my opinon, though perhaps we can make more of an effort to get
regular Grexers to become members.


#4 of 54 by dpc on Fri Jun 27 19:29:48 1997:

Let's get the new Sun up and running.  Since this obviously isn't
going to happen "short-term," I assume it fits under "medium-term."
Once we do that (or even sooner), I would *really* like to restrict
all the folks who use us as a remote mailer somehow.  There is no
easy way to do this, I know.


#5 of 54 by dang on Fri Jun 27 22:41:27 1997:

With Valerie, I would very much like to see mail limited on Grex, so that it
doesn't hog all the resources, especially the link.


#6 of 54 by orinoco on Sun Jun 29 13:37:47 1997:

Re:valerie's issue #4:
I know next to nothing about computers, and have very little money, but I
*would* like to help out somehow.  In the past, I've felt kind of guilty about
using grex so much and not becoming a member, but I really can't afford it.
What can I do to help out that doesn't require a high degree of computer
literacy?


#7 of 54 by steve on Sun Jun 29 17:20:20 1997:

   How about talking about Grex to people, when you think you might
have found a kindred soul who might like us?   That would be very
useful if everyone could find just one or two people and get them
into the Grex community.

   So, for social issues I agree with Mark.  On the technical side, Dave
is right.


#8 of 54 by krj on Sun Jun 29 22:27:35 1997:

Well, if you can't afford to buy a membership, you'll just have to
be interesting enough in the conferences so that *other* people will 
want to join.  Think of it as a form of "sweat equity."  
  (1/2  :)  )
 
I think getting conferencing speed up is important -- cutting out the
net link lag and the CPU lag so that conferencers don't get bored 
and go away.
 
I think that Grex's concept of "open-everything-for-everybody" is  going 
to need some reformulating, and there are going to be some political 
struggles with that.  (Hi, Mary!)  I think there are two problems:
1) As has been discussed at great lengths, many of us believe that 
free-email-to-the-world is becoming a black hole which will 
grow to consume all of the resources Grex will let it consume.
E-mail chews up lots of the link; it chews up lots of staff time,
from what I hear.
 
2) Regarding the minutes of the last Board meeting, where Steve Andre'
reports that Grex is under heavy pressure from vandals:  I am starting 
to suspect that offering open shell access to the world just makes 
us too tempting a target to the malevolent.  Again, here I suspect 
the issue is staff time.  Every hour Steve and others spend coping 
with system attacks is an hour not spent on getting the new Sun up.
 
So I would lay out an argument like this:  From this item, and other 
discussions, I feel a strong consensus that conferencing, and 
probably party and e-mail within Grex, are Grex's most important
functions -- these are the community-building activities.
But we have not been willing to act on this consensus: to direct 
our scarcest resources, the link and staff time, in support of 
this community-building.  Such a redirection of resources would 
imply some significant shake-up in how Grex operates, and what it 
offers to the wide world.


#9 of 54 by steve on Mon Jun 30 03:07:52 1997:

   The risks associated with an open system such as Grex are
real, but a couple of things need to be remembered.  The first
is that some largeish number of root-explotiable bugs in commonly
available software these days comes from complicated user programs
like elm, or lower-level things like sendmail which are still
accessable by people.

   The second part is that more and more attacks are from the
"outside", ie off-grex but still attacking the system.

   Thus a lot of #1's problems aren't solved by closing
down the openness, and none of #2's are.

   The worst part of Grex's openness is that we've become
something of a "magnet" amongst the vandal community.  I'm
not sure that we'd be better off in this regard if we shut
shells down, or not.  Unforunately, probably not.

   You are right however Ken, that this activity does take
away from working on the Sun.  And yes, at some point we
are going to have to come up with a new way to do some things.


#10 of 54 by mary on Mon Jun 30 15:09:40 1997:

(Hi, Ken!)


#11 of 54 by dang on Mon Jun 30 17:01:42 1997:

I, for one, would very very much like to see Grex keep open shells.  You can't
get shell access easily any more.  Most ISP's don't offer it.  Universities
do, but only to students and staff.  So, grex is one of a few places to get
shell access, and one of even fewer that are free.  I very much like that
about Grex.  The shell access that Grex offers has been very important in
deciding my career choice, and I'd like to offer that to others.


#12 of 54 by mdw on Mon Jun 30 19:05:38 1997:

People have been saying since 1983 that it was "not possible to run an
open access Unix system".  I think first m-net, and now grex, show that
this is just not true.  It does take some percentage of staff time, &
more importantly, "the right perspective", to do it.  This was known
from the start on grex.

Going to a more closed system would not reduce, in the slightest, the
amount of work that would be necessary here, nor, to any substantial
degree, would it increase security.  In the short run, it would actually
considerably *increase* staff work, because considerable extra effort
would have to be devoted to finding replacements or workarounds for
programs that inherently offer shell access.  For instance, programs
such as mail, party, vi, and picospan all offer this facility, and
derive part of their power and usefulness from including this facility.
In the long run, since clever users with socialization problems would
persist, the potential for misuse would still exit.  Since also grex
doesn't have the budget that NASA has for exhaustive proactive custom
"bug-free" design methodology and code development, it is inevitable
that there will sometimes be bugs to be found.  One need only look at
the typical freenet to find an example of a system that is considerably
less open, yet probably has just as many if not more problems from
vandals.


#13 of 54 by steve on Mon Jun 30 19:20:45 1997:

   Heh.  Thats an excellent point marcus, about various freenets
problems.


#14 of 54 by remmers on Mon Jun 30 20:35:17 1997:

I agree about the desirability of keeping open shell access.
Another reason to throw into the pot is that it's in line with
Grex's charitable mission. I can think of various people who
have benefitted education- and career-wise from the open shell
access on M-Net and Grex. From a practical point of view, it's
an avenue for recruiting new staff members. Folks like Ed
Anselmo, Marc Unangst, Mark Bobak, and Steve Gibbard come to
mind.


#15 of 54 by janc on Tue Jul 1 04:32:24 1997:

By all means, keep open shell access.

I would like to institute some scheme that would limit the preportion of our
total bandwidth that is consumed by mail.


#16 of 54 by valerie on Tue Jul 1 04:32:50 1997:

This response has been erased.



#17 of 54 by krj on Tue Jul 1 16:44:02 1997:

I'm certainly not invested in the idea of restricting shell access if it 
won't buy anything.  Let me rephrase my goal:  We need to examine 
what ongoing tasks are devouring staff time; if these tasks are not 
central to Grex's community building purpose, we need to find ways 
to streamline or eliminate this drudgework.  We need to be able to 
redirect more staff energy from "maintenance" to "development."
 
The "fake POP server" is an excellent implementation of this idea; 
we need to find more and bigger ones.


#18 of 54 by steve on Tue Jul 1 23:30:39 1997:

  Heh.  Can't argue with that.


#19 of 54 by mdw on Wed Jul 2 05:40:28 1997:

I am currently thinking of several changes to newuser that may help in
this area, both to "discourage" not so good uses of grex, and to
"encourage" conferencing.

The first change I'm contemplating is to teach newuser to generate a
cflist based on people's interest.  This way, instead of seeing one
"introductory" cf, which may not be intrinsically interesting to all,
and which isn't really a "home", people will instead see real live
conferences right off, that are hopefully discussing things that will be
of interest to that person.

The second change I'm contemplating is to make newuser more nosy about
what people are going to use grex for.  One thing I want to catch is
people hoping to run "eggdrop" on grex.  This is definitely a big waste
of time, both for us and them, and I'm hoping by nipping it in the bud,
so to speak, that we can just avoid the whole mess.  Another thing I'm
interested in is catching people who are interested in mail.  What I
want to do there is basically a bit of "social engineering".  I want to
describe grex's mail in the least favorable terms possible (no modern
gui, slow, doesn't work to every system, etc.) and then to have pointers
to as many other mail systems as we can possibly find (hotmail, juno,
etc.) I also hope to deal with mailing lists, web pages, gif files, and
so forth in a similar way.  It'll probably take a bit of work to get the
wording in all of this phrased right, but I'm hoping that the result
will be a higher quality of user who makes it through newuser, the users
who basically *want* to become part of our community, and aren't just
looking for free pine.

I don't know when I'll get time to work on either of these.  I'm hoping,
though, that it will be soon.


#20 of 54 by scg on Wed Jul 2 06:52:01 1997:

I worry about adding more stuff to newuser without also taking some stuff
away.  Newuser already makes people go through a huge amount of stuff, and
it seems like while people are often willing to add something to it, people
are rarely willing to see anything taken out (kind of like the Federal
budget).


#21 of 54 by jared on Wed Jul 2 16:04:17 1997:

Yeah.. I wrote a very shortened version of newuser for nether.net
awhile back to make it much easier to get on the system, which took
out lots of weird little quirks that I did not think were very worthwhile,
and also had it do stuff like check the passsword against a wordlist,
and verious other things.. it's by no means perfect, but making it easier
to initally register, and later change the information is good.

I've got the eggdrop/bot problem also, but on a different scale since people
can compile it and make it owrk on my system, but I go and delete
their accounts.. there's stuff in the newuser text, /etc/motd, people
are just plain stupid.  Putting more text in there isn't going to be 
worthwhile.. they're already doing it and there's sufficent docs in the
ftpd motd to get the folks that are actually interested in caring.



#22 of 54 by valerie on Wed Jul 2 21:30:47 1997:

This response has been erased.



#23 of 54 by nt on Wed Jul 2 23:25:15 1997:

whenever I see .cflist I get hurt. guess why? :)


#24 of 54 by dang on Thu Jul 3 00:10:29 1997:

(Some of use acutally *like* to use strange things like vi for our .cflist
and .login, etc. :)
(Then again, some of us are total geeks...)


#25 of 54 by krj on Thu Jul 3 16:30:27 1997:

Let's have some more dreams about what Grex should be in 1-3 years.


#26 of 54 by valerie on Thu Jul 3 16:46:00 1997:

This response has been erased.



#27 of 54 by nt on Fri Jul 4 03:19:46 1997:

#26 thanx :)


#28 of 54 by valerie on Wed Jul 16 14:00:39 1997:

This response has been erased.



#29 of 54 by tsty on Fri Jul 18 04:15:38 1997:

one ignored medium range action would have beenfor the cfadm in
atrining and also new coop fw,   void   to have had the opportunity
and experience of changing the conferences and aslo changing the fws. 
  
but, noooooooooooooooooooo, the gun was jumped.
  
so now you have had the chance to train someone new ---adn stiffed her.
 
 ...and   ALSO  install the WRONG fws for coop.cf. 
  
don't we feel just so proud?
  
where the hell did 'respect go'?


#30 of 54 by valerie on Fri Jul 18 15:04:45 1997:

This response has been erased.



#31 of 54 by srw on Sun Jul 20 20:50:25 1997:

I'd like to see the dialup users have the (additional, optional) ability 
to access some of the services that only our internet users can access. 
This would require that we allow PPP connections upon dialup. We could 
make these non-routed connections, so that users would not be able to 
access services this way except those provided on our own subnet.

Advantages: 

(1) Dialup users could see our web pages, and use Backtalk if they 
preferred.
(2) It would be much easier for us to provide distributed services that 
way.
(3) We could add POP mail service for these users. It is still arguable 
whether it is a good idea to provide POP over the internet, but if we 
had implemented a mail bandwidth limiter, we could conceivably do that 
as well.

Users would thus have access to much better interfaces than they do now, 
providing that they have computers which can support PPP. Only really 
old computers can't do this.


#32 of 54 by dang on Sun Jul 20 23:45:52 1997:

I would very much like to see this too.  I prefer greatly to use backtalk to
do my cfing, but I can't as I dial up to Grexes modems all the time.


#33 of 54 by scg on Mon Jul 21 02:35:36 1997:

The Chase IOLan terminal server we have supports PPP connections.  That could
be set up fairly easily, I think.


#34 of 54 by remmers on Mon Jul 21 10:58:08 1997:

Interesting idea. Of course, dialup users can access backtalk
and grex's web pages now via lynx, but that ain't the same as
accessing them via their own browser, which PPP would make
possible.

How annoying would a restriction to grex's subnet be, though?
Dialup users wouldn't be able to follow any of the hyperlinks
that show up in backtalk and grex's or users' web pages. Would
they be able to see the graphics on the web pages (Grex logo,
buttons in the pistachio interface, etc.) if they continue to
be served from elsewhere?


#35 of 54 by janc on Mon Jul 21 14:33:34 1997:

We'd have to set things up so that dialup users get their button images served
from Grex.  At some point I hope to make this configurable, so that regular
backtalk users can unpack a collection of images on their home computers and
serve them from their own machine.


#36 of 54 by valerie on Mon Jul 21 14:34:39 1997:

This response has been erased.



#37 of 54 by valerie on Mon Jul 21 14:34:56 1997:

This response has been erased.



#38 of 54 by remmers on Mon Jul 21 18:20:13 1997:

I agree that the restriction is necessary. Without it, we'd
become something we can't afford to be and don't want to be.

This is the "medium-range planning" item, so maybe this next it
a little off-topic, but I just want to remark that the claim made
in #33 that Backtalk is a "nicer interface" than the standard tty-
based interface to Picospan is debatable. Without disparaging the
tremendous job Steve and Jan did in creating Backtalk, the fact
remains that there are things that I and others like to do in
Picospan that just aren't possible via Backtalk with the current
interfaces, e.g. various kinds of browsing (the equivalents of
"browse new", "browse since -1", etc.), doing a "!" escape to
run a Unix command (!tel, !mail, !pwho, !who, etc.), and piping
Picospan output through filters (I tend to do "browse all | 'grep string"
fairly often).

The lack of Backtalk equivalents to various useful Picospan
commands is remediable, since Backtalk has a very flexible
mechanism for creating new interfaces. The lack of connection
to the "rest of Grex" seems harder to address.


#39 of 54 by dang on Mon Jul 21 22:23:25 1997:

That's what your telnet window is for. :)  Afterall, if you can run your
browser through the link, you can telnet too.  I have no objection to
telnetting into Grex as long as I don't have to go through the link.


#40 of 54 by scg on Tue Jul 22 05:05:27 1997:

I can think of a number of DNS or routing based hacks that would make
BackTalk's images come from different places depending on whether people were
dialed into Grex's local terminal server or coming in over the Net, which
wouldn't require any modifications to BackTalk at all (and could also apply
equaly well to other graphics on Grex's web pages).  Everything I can think
of for that sounds pretty ugly, though.  As far as BackTalk goes, it might
be better for BackTalk to look at where a user is coming from and give
different image tags accordingly.


#41 of 54 by remmers on Tue Jul 22 13:37:38 1997:

Re #39: You're right of course, and that addresses the "doing
Unix commands" part of things, although not the "integration of
Unix functionality with Backtalk" part. (That could be an
interesting exercise in interface design...)


#42 of 54 by valerie on Tue Jul 22 15:48:57 1997:

This response has been erased.



#43 of 54 by srw on Wed Jul 23 01:38:18 1997:

At some point in time, we'd like to go over the functionality missing 
from Backtalk that our most sophisticated picospan users (like John) 
take advantage of. Ideally, we don't want Backtalk to be giving up many 
features.

The backtalk's config file defines a symbol, /imghost, which is used as 
a base URL for the images. That symbol is set up in 
/usr/local/lib/backtalk/backtalk.conf. It is currently set up to point 
to hvcn unconditionally. It could be defined conditionally. It does make 
more sense for backtalk to make this distinction than to put in a DNS 
hack or some other kludge.

I think there are other benefits that will accrue to local users besides 
accessing Grex web pages and backtalk, btw. I'd like to see us support 
POP so that local users can use local interfaces available for their 
mail, too.


#44 of 54 by awijaya on Tue Aug 19 16:29:25 1997:

Hello, I have suggestion to offload the giant e-mail problems.
IMO the staff can put info about various info about free e-mail 
service such as Juno, Hotmail, Bigfoot etc.
There are several free faster Lynx text browser service
accessible using telnet. You can download files using Kermit (77 cps).

You can send e-nmail using the service ("mailto:....)


#45 of 54 by senna on Tue Aug 19 21:59:54 1997:

Have we already mentioned that?


#46 of 54 by dpc on Wed Aug 20 15:49:53 1997:

Has it been done?


#47 of 54 by dang on Wed Aug 20 20:28:47 1997:

Staff routinely points people to hotmail, juno, etc. for email only service.


#48 of 54 by tsty on Sat Aug 23 06:16:17 1997:

... as do helpers... fwiw.


#49 of 54 by lilmo on Tue Nov 11 02:44:09 1997:

Good to see planning has been alive and well for ht elast two and a half
months.  *ironic grin*


#50 of 54 by e4808mc on Wed Nov 12 22:28:20 1997:



#51 of 54 by srw on Wed Nov 19 03:35:19 1997:

I think we're too overwhelmed with work to do any planning. Once we et the
670 up, and the terminal servers are usable, we should really do some
planning. The first thing I'd like to plan is distributed function. But that
would be premature right now.


#52 of 54 by other on Wed Nov 19 06:30:55 1997:

by that do you mean, say using the old machine for mail processes, and the
new machine for home directories and conferences?


#53 of 54 by srw on Sun Nov 23 22:10:39 1997:

I just meant that we ought to get ourselves in the position where we can 
have several machines doing several different functions. *Some* machine 
could handle mail delivery, but it might not be grex, and it might not 
be the old machine. 

Other things besides mail could be offloaded to different processors 
too. One of the things that would make all of this easier is a 
distributed authentication system. This requires a little planning, and 
more staff time than we have free right now.

There is progress being made on the 670, but there is more ground to 
cover there before it will be visible.


#54 of 54 by lilmo on Tue Nov 25 22:55:56 1997:

I wish I could help with that, but I lack both expertise and proximity.  <:-T


There are no more items selected.

You have several choices: