|
|
| Author |
Message |
brighn
|
|
Backtalk
| 
Sep 4 21:52 UTC 1996 |
Jan is currently in the stages of testing Backtalk. As I understand it,
Backtalk makes the Grex conferences accessible from the Web. The concern has
been raised by several of us that Web-users represent a significantly larger
population of users, and therefore present a potential problem. The two
possibilities for access that I have seen discussed with any real seriousness
are:
-- Develop an interface (Backtalk or other) that would allow persons without
Grex accounts to read, but not write to, conference items, and allow persons
with Grex accounts to read and write.
-- Develop an interface (Backtalk or other) that would only allow Grex users
to access the conferences.
Thoughts? comments?
Clarifications?
|
| 177 responses total. |
janc
|
|
response 1 of 177:
|
Sep 4 22:17 UTC 1996 |
Some technical notes.
Backtalk allows people to login with their Grex login names and passwords
before accessing the conferences. You will be able to read, post, and do
most other things you can do in Picospan. (Fairwitness commands work, etc)
Backtalk supports several functions that Picospan doesn't. These are
configurable, and can be turned off or on.
- Anonymous reading. This would allow people without logins to read
conferences, but not to post to them. Anonymous readers also have
no participation files, so everything is "new" everytime they come.
- Retitling items. Backtalk can be set up to allow fairwitness, item
authors or both to change the title of an existing item.
Even with Anonymous reading turned on, web searchers like Alta-Vista and
Yahoo won't index conference items accessible under backtalk.
|
ladyevil
|
|
response 2 of 177:
|
Sep 4 23:56 UTC 1996 |
I'm all for it, except fopr anonymous reading.
|
scott
|
|
response 3 of 177:
|
Sep 5 00:13 UTC 1996 |
Anybody who is worried about having more readers should reflect on growth
*already* happened. If you're worried about spread, don't post. Spread in
inevitable.
|
brighn
|
|
response 4 of 177:
|
Sep 5 05:07 UTC 1996 |
I like the retitling of items option.
My stand continues to be against anonymous reading, in suport otherwise.
|
void
|
|
response 5 of 177:
|
Sep 5 05:17 UTC 1996 |
why are you opposed to anonymous reading, brighn?
|
brighn
|
|
response 6 of 177:
|
Sep 5 05:20 UTC 1996 |
I'm beginning to think that it's out of stubbornness and habit, Void.
=}
|
tsty
|
|
response 7 of 177:
|
Sep 5 08:10 UTC 1996 |
as long as readers/participants have to log into grex first ... backtalk
appears to be simply a different mechanism for local conferencing.
within those simple and consistant constraints i think backtalk is
a very good idea. there might be all those other potentials as well
from which janc could make a buck or two selling it (with or without
the extra bells and whistles) elsewhere. more power to him.
|
davel
|
|
response 8 of 177:
|
Sep 5 13:59 UTC 1996 |
I don't understand the retitling feature, Jan. Does it rewrite the item file
from the beginning?
|
janc
|
|
response 9 of 177:
|
Sep 5 18:01 UTC 1996 |
Yes. It locks the item file, copies it into another file, and copies it back
with a new title. It's actually major surgery, and I'm not 100% sure picospan
won't grumble and complain, so we may not be able to use it on Grex. But I've
seen other conferencing systems where this was used to good effect, and we
wanted it in Backtalk.
I should remind people that this is a joint project between Steve Weiss and
I. We do intend to sell it, though don't expect to make any fortunes.
In most installations we expect it will run stand-alone, without yapp or
Picospan, and this allows us a bit more flexibility. Some features that
are already on HVCN (or will be there soon) will be a long time coming on
Grex, including buttons to change your password and HTML in responses.
|
kerouac
|
|
response 10 of 177:
|
Sep 5 19:04 UTC 1996 |
I dont think anonymous reading is a concern...anonymous
readers are more likely to become active users if they can use grex
on their own terms.
The goal should be that in two years or less, most users (at
least non-local) should be accessing grex through the web and not
via slow antiquated telnet.
|
kerouac
|
|
response 11 of 177:
|
Sep 5 20:59 UTC 1996 |
Ofcourse grex's web page doesnt exactly setspeedrecords right now...
I suspect that for Backtalk to reachits fullpotential, Grex is going to
need to dump ICnet for a better,faster ISP. Is changing providers
something being sconbsidered for long term?
|
janc
|
|
response 12 of 177:
|
Sep 5 21:17 UTC 1996 |
Actually, the web interface will probably be slower than telnetting in and
using Picospan. Given the quality of our internet connection, as you point
out, Backtalk is going to be rather turgid. However, some things, like
composing and editing responses, will be much faster, because they are done
on your local machine, not on Grex.
I don't have any goals for how many people will use one interface or the
other. We just offer alternatives. It's up to the users to decide which they
want to choose. Steve and I are trying to do a good job with Backtalk, and
hope users will like it, but we'd be the last to claim it is unambiguously
better. It's not, and the Web will have to evolve a fair bit before anything
really ideal can be done.
Yes, we are actively seeking to upgrade our internet connection. We've been
a bit distracted by the move, but that may prove to be an opportunity since
we may be able to move someplace where a better net connection would be
possible.
|
coyote
|
|
response 13 of 177:
|
Sep 5 22:01 UTC 1996 |
How much slower do you think BackTalk will make Picospan, if any?
|
russ
|
|
response 14 of 177:
|
Sep 5 23:37 UTC 1996 |
The HTML wrappers will take more bandwidth, for certain. And if anyone
adds a response while you're reading the item, or you decide to back up
a few responses, I expect that it willl be much slower than PicoSpan
with a pager. I don't think that a WWW interface is anywhere near as
bandwidth-stingy as telnet.
|
nephi
|
|
response 15 of 177:
|
Sep 6 01:15 UTC 1996 |
Kerouac, I'd rather say that it's not because of ICnet that our Internet
conection runs so slowly, but because of the *size* of our Internet
connection. There's only so much stuff that we can cram over a 28.8 modem,
and people like you don't seem to care enough to actually donate <gasp!>
money. If we had more money, we would have an ISDN connection for sure, and
not only would people like you not have to complain about having to wait to
log in, but you wouldn't have to complain about how slow the place is,
either. Hmph!
And regarding #14, I don't think that Backtalk will work the way you picture
it. I believe that it displays an entire item at a time, rather than just
one response at a time, so seeing previous responses will be as easy as
hitting the "up" arrow on the scroll bar, or as easy as hitting "Page Up" on
the keyboard. The HTML wrappers won't take up any bandwidth, actually.
What they will end up taking is CPU. What will end up taking lots of
bandwidth is the auto-browse thing that it does (so that people can just
click on the item they want to read), that I can think of. Personally, I
think that instead of being faster or slower, that it will end up being
about neutral, unless Grex's web server falters under the increased load.
Of course, no one will really know for *sure* until we test it all out. 8^)
|
srw
|
|
response 16 of 177:
|
Sep 6 01:16 UTC 1996 |
There are a number of reasons that the total amount of text shipped will be
slightly more via http than telnet. The HTML tags are one reason, although
it is a slight amount. Another is that the Backtalk interface provides more
options to choose from, and that menu of options uses a little more.
There are also reasons to expect it to reduce bandwidth. There are plenty of
times when I am using Picospan that I need it to redraw things for me. This
is especially true when I am formulating an answer in gate, or vi (and pico
is even worse), This interactive character slinging will go away with
backtalk, as it is all done on your home PC or Mac.
We were searching for a GUI interface for conferences 4 years ago, when I
started grexxing. We couldn't agree on a method to achieve that then. RIP was
proposed, but it did not have enough penetration. Finally the web has given
us a better RIP, HTML. It has the penetration, and that's why we did Backtalk.
It's not like we really love HTML or anything like that, it's just that
so many people have a client for it.
|
srw
|
|
response 17 of 177:
|
Sep 6 01:21 UTC 1996 |
Nephi slipped in. I'm not quite sure what he meant by "wrappers".
Backtalk produces HTML directly. It is written in C. There are no wrappers.
There are embedded tags, and they do take up a little space.
I am interested in seeing how it affects things. We really don't know.
We do expect it will increase the balance of conferencing to other things.
You are right about showing one item at a time. Responses are separated from
each other by horizontal rules (or graphical elements) and clearly marked.
Scrolling is very easy and requires no interaction with Grex or bandwidth to
be consumed.
|
janc
|
|
response 18 of 177:
|
Sep 6 02:40 UTC 1996 |
Displaying items via Backtalk will also require a bit more CPU than the
equivalent action via Picospan, for a couple reasons:
- Backtalk is memoryless. Between one query and another, it completely
forgets you exist. That means that every time you click a button,
Backtalk will (1) check that your login and password are those of a
legal user, (2) find your current conference, and (3) load your
participation file and the conferences sum file. Picospan does all these
things much less often.
Note, though, that you'll never notice this. Yes, Backtalk rechecks your
password with each transaction (actually the httpd does it), but you don't
have to type it in again. Your computer automatically remembers it and
resends it. The same goes for your current conference, which item you are
currently at, and all other state information. Your computer sends it
to Grex for each transaction, and Grex sends the updated information back.
It is possible to avoid this (WebCaucus does so), but other problems come
up, and things get much more complex. In the future we may study ways to
improve performance here, though as WWW applications go, Backtalk is really
pretty fast already.
- Backtalk is an interpeter. We wanted to be able to easily redesign the
look and feel of the screens backtalk puts up, as well as rearrange the
functions of the controls. So backtalk's behavior is actually controlled
by scripts. Since it has to interpret those scripts, this makes it a
bit slower than Picospan. However, we are using an RPN (Reverse Polish
Notation) language similar to Forth or Postscript, and those are really
quite fast to interpret.
|
ladyevil
|
|
response 19 of 177:
|
Sep 6 04:01 UTC 1996 |
Fine- I'd really rather that anon posting not be allowed.
At least, not in Sexuality II, okay? Thanks.
|
scg
|
|
response 20 of 177:
|
Sep 6 04:17 UTC 1996 |
Anonymous posting, with an anonymous account, has always been allowed. Selena
keeps telling us that's the only reason she's willing to use Grex. Backtalk
will still allow anonymous people to get Grex accounts and post. We're not
planning on changing that.
|
janc
|
|
response 21 of 177:
|
Sep 6 05:54 UTC 1996 |
To post with Backtalk, you must have a valid Grex account, and that name will
appear on the posting, just alike a PicoSpan posting. In fact, it should be
impossible to tell for sure how any given posting was entered.
The only thing Backtalk could possibly allow people without Grex accounts to
do would be to read items.
|
janc
|
|
response 22 of 177:
|
Sep 8 02:36 UTC 1996 |
A test installation of Backtalk is now running. We've still got a lot of
testing and development to do, so it has some problems, but seems mostly
OK.
It is at: http://www.cyberspace.org/cgi-bin/bt
Notes:
- Log in with your regular Grex ID and password.
- You have a choice of two interfaces, neither of which is "done".
Vanilla is in many ways the better design, but isn't very flashy.
Pistachio is a playground for netscape stuff, and has some features
vanilla lacks, but is overall kind of awkward.
- Only two conferences will be accessible: "backtalk" and "backtalk2".
- You can join these two conferences with Picospan too. They have no
very interesting content though. They are just test conferences.
- Much of the .cflist stuff doesn't work yet.
- Displaying lists of users and information about users doesn't work yet.
- It's slow, especially when the net is slow.
|
tsty
|
|
response 23 of 177:
|
Sep 8 03:27 UTC 1996 |
item 5 has response 4 listed twice in vanilla backtalk.
interesting, however i was under the *impression* that the
anonymous stuff was not going to be implemented on grex? none of it.
an anonomous login is a different situation entirely, and of course
i support that in the fullest.
|
janc
|
|
response 24 of 177:
|
Sep 8 03:36 UTC 1996 |
No decision has been reached over whether any anonymous reading will be
allowed. For the moment, Backtalk is configured to allow anonymous reading
of the deeply boring "backtalk" and "backtalk2" conferences. This is for
testing purposes only. I want to make sure it works. This software is not
being written exclusively for Grex, and I am taking the opportunity to test
several features that may well be turned off when we go live on Grex.
|