response 1 of 60:

Jul 3 13:32 UTC 1995

(GUI stand for Graphical User Interface)

response 3 of 60:

Jul 4 06:00 UTC 1995

hmmm, i think that reserving grex and environs for the slightly
more competent is a GoodIdea. Kinda like having the tight internet link.

response 5 of 60:

Jul 4 20:07 UTC 1995

Am I wrong in saying that the httpd keeps a log of who does what anyway?
Like if I http://www.cyberspace.org/ isn't there a httpd log showing
avi accessing the HTML file ... I'm pretty sure that there is
something like that available on some http daemons.

response 7 of 60:

Jul 5 06:47 UTC 1995

I may get a chance to look at the hot pepper page, but I already know
it is possible for the server side script to look at files, and
those files can be updated. So to call http "stateless" is perhaps
naive. It is precisely because it is possible to write CGI scripts
that refer to saved-up info that I believe a solution *IS* possible.
No one has used this strategy to build a conferencing system, though.

Picture this.

An httpd server running on a machine that no one but an occasional
staff person logs into.

A file on that machine of conference participants, storing 
name, email address, conf passwd, and participation info for each one.
This info would be centrally located rather than in home directories,
because there *ARE* no home directories on this machine.

For each user, The first connection to the machine brings up an html 
page named "login".  The user supplies his/her name and conf pw.
This produces html conference pages with all kinds of nice buttons,
and each response has an encoded security key - it could even be a one-time
password - generated and verified by server script.

Security and Conference participation are both handled here.
New participants can be created rather simply via http, and thus
there is no need for a newuser program, or even for Jan's Webbish nu.

OK, now you can go ahead and let people log in to that machine.
I needed to keep them off only to make my point.

response 9 of 60:

Jul 6 08:15 UTC 1995

What about ansi graphics. Those can't be real expensive to have.
And almost any Comm program will run ansi.
I haven't used ripterm yet but I have it for my ppp connection.
I just haven't found a place to use it yet. But I have heard it
was pretty good.

response 11 of 60:

Jul 6 13:17 UTC 1995

The GUI intelligence would reside on the user's machine, not Grex.
Thus, unlike applications such as Pine and Elm, the GUI functionality
would not be a direct burden on Grex's CPU.  The small bandwidth of
Grex's internet connection would result in slow response times, but no
more so than in an ordinary telnet connection.  In fact, the slowness
would probably be less distracting to the user and more efficient for
Grex, since text would be sent back and forth in blocks rather than
charcter-by-character. (No more waiting 5 seconds for a character
to echo...)

I agree with Steve Weiss that a GUI interface to Picospan, with
participation files, the ability to forget items, etc., and that people
could access through their web browsers, could all be implemented with
CGI.  A cleverly designed html interface could actually make some
additional functionality possible -- for example, by customizing the
rseps, you could make responder's names "clickable" so that selecting
their name with the mouse would bring up a dialog box for sending them
mail.

I looked at sfgate.com.  They appear to be running a conferencing
system via a the web with structure very similar to Picospan -- a bunch
of conferences, each conference has a number of items, items have
responses, and the system keeps track of what you've seen.  Before you
can access it, you go through a registration procedure and pick a login
id and password, sort of like our "newuser" although much less
detailed.  My connection stalled before I could read any items, but
I'll try again and have a closer look.  At any rate, computer
conferencing more or less as we know does seem to be possible via the
web.

Implenting a web-based interface to Grex could have some non-trivial
side effects of a social nature, depending on how it's set up.  When
you log into Grex the old-fashioned way, you're in a Unix-based
"environment" where you can easily switch back and forth among several
modes -- conferencing, party, mail, chat, file-browsing -- by "shelling
out".  I think that this flexibility has a lot to do with making Grex
the kind of "community" that it is.  Designing a GUI interface that
promotes a similar "feel" and that allows things like live chat
sessions in addition to conferencing would be tricky to implement, but
if we're going to move Grex into the GUI age the issue deserves some
careful thought if we want to preserve our community dynamics.

response 13 of 60:

Jul 7 05:08 UTC 1995

Thanks a lot, Jan, for the SF Chronicle pointer. I will be checking that 
out. As you can tell from my earlier posts, I wasn't aware anyone had
cracked this yet. It sounds a lot like what I have been specifying mentally.

by contrast --

I tried the sportstalk section of the espn page at
http://espnet.sportszone.com/talk . It gets a lot of traffic, but
it seems to behave a lot like a web-based party.

response 15 of 60:

Jul 8 00:52 UTC 1995

John, you will be rewarded in Heaven - (or Zingermans) - 
pick one.

response 17 of 60:

Jul 14 22:08 UTC 1995

The following is a very nice directory of Web-based conferencing sites:

 http://freenet.msp.mn.us/~drwool/webconf.html

It even includes a read-only interface to M-Net's conferences.

response 19 of 60:

Jul 15 08:40 UTC 1995

        Hey, who did that?  Nifty; thanks, janc!

response 21 of 60:

Jul 15 21:54 UTC 1995

So far the Gate looks like the most interesting one to me.
It has the basic concepts of Picospan (security and participation files)
but is still missing a lot of functionality. I was not wild
about the presentation of the header screens either.

The problem I see is that the other systems are even more problematical.

response 23 of 60:

Jul 16 16:22 UTC 1995

I think that if we could write CGI scripts and didn't try to tie it
to the existing Picospan conference system, that it could be done.
We have the advantage of knowing what features we want to be able to 
choose. things like expurgate/scribble, find, browse, and other useful
picospan features, plus Fairwitness capabilities, are all missing from
the Gate, but we know how to do these things.

To connect it to existing Picospan conferences the problem would be
much trickier, sonce Picospan cues on account's uids. We'd have to
create an account on Grex for each participant. That's something
I'd like to avoid.

If Picospan could use email addresses instead of Grex accounts as
identification, then I would be very interested in adding an interface
to it, but if that can't happen, then I'd be more interested in seeing
a separate conference system.