This item text has been erased.

556 responses total.

This response has been erased.

This response has been erased.

Is there any easy way to forget linked items in conferences?

Ideally, I'd like to have some way to say "Forget <conf> <item>" to
forget the item in the least appropriate conf.

go to the conference where you don't want to see the item, find the item
number (by using "browse"), and then type "forget <item number>" at the
picospan prompt.

Or you can edit your participation file for the conference directly,
putting a -1 in the 2nd field of the line for the item in question.

This response has been erased.

What is the arithmetic to use at the Respond prompt to see a group
of responses which are not the last ##, nor a single response. I've
tried the    nn-xx  and receive   Bad parameters near "-"  as a
diagnostic. Let me try an    only nn-xx and see what happens.

same thing, hmmmmmm.

Hmmm...  I misread the first point in #6 as 'Type "enter"' rather than
"Hit Enter"...

Actually, typing "enter" is something you can do at the "Respond or pass?".
It lets you enter a new item, but when you've finished, it takes you back
to the "Respond or pass?" prompt at the item your were reading.

The intent of this was probably to encourage people to channel drift into
a new item, but I don't think many users know about it.  Maybe if the
prompt said "Respond, enter new item, or pass"...

When I first started using Picospan, several years ago, the prompt said
"Respond, forget, or pass".  The "forget" was dropped because people
thought it was anti-social.  (You can still do "forget", of course.)

My problem with breaking out of a response in mid-stream and then hitting
"enter" at the "Respond or pass" prompt is that the item does not get
marked as read.  We discussed this about a year ago.  I've been ctrl-c'ing
out and then typing "last" to see the last response on the item (hoping it
is isn't a long one).  That seems to mark the item as seen.   I think
there were some other ways to do the same thing but I can't remember them
now.

And also "Respond, forget, or pass" is the (anti-social) part of Confer(tm).
   
   sfsf

This response has been erased.

Agreed, inelegant but functional. Just looking for a better way ...

I wish there was a better way to track all those
linked items.  I hate having totry to keep track of them
manually.

This response has been erased.

I know what *I* would like, & it shouldn't be *too* hard: when I read
a linked item in one conference, it marks it as read for me in any other
conferences as well.  So I can read the new stuff in either place (not
forget it) but only once.

(Was that the kind of thing you meant, Dave?)

This response has been erased.

That would actually be quite cumbersome and expensive to do
without a major overhaul of the way Picospan organizes files.
Pico just uses Unix hard links to implement linked items, and Unix
maintains no list of what links there are to a file.  So every
time you read something new in a linked item, Picospan would have
to search through *all* items in *all* conferences to find the
links, then update every one of the affected participation files.
That's a lot of data processing.

I've just signed on with ProComm for the first time. It dials, I get
CONNECT, but then I get some garbage, and things hang up. By experiment,
I found that a "shift %" releases, and then its fine. There must be
something wrong in my Modem or Terminal Setup. Any suggestions?

Hmm.  I use ProComm (2 different flavors) all the time.  What terminal are
you emulating?  (I use ansi, for which I must use 8N instead of 7E.  This
doesn't always produce the optimum results for things like vi, but it's
easier for me than something else (for obscure reasons).)  I often do get
some garbage before the login prompt.

I set it to the VT-100 emulation: no reason, except that's what Merit
likes. (I know little about these protocol matters - just want what
works.) I wouldn't mind the garbage - its the hang-up I'd like to get
rid of (not that "%" will break a finger ;)).

Oh yes - I get garbage and a hangup after "quit" too, and it take "%" to
continue.

This response has been erased.

This response has been erased.

I'm using 1200 baud (all my old Tandy will crank out). So, that's 2400
baud garbage? Well, if I can't avoid it, SHIFT-% is sure fire (rather than
hitting returns). That is, just "%". (I don't know if any other combinations
will be as effective.)

This response has been erased.

For what it's worth, a backspace or del character sometimes clears
up the (optional) 1200 baud connection I make. Hadn't tought about
the Ctrl-Q option though ....... 
  
From Merit, the % cahracter is supposed to be a Merit command-prefix
which is trapped adn acted upon. Soometimes it is, sometimes it
isn't, but I haven't codified which is which - a double % is the
way to get the second percent sign into the following system. Does
anyone ahve any information about a Ctrl-% and what that does to Merit?

No, I'm dialing directly. I have now found that "%" is not unique 0 by
far! While "return" doesn't unlock the hangup, most letters do - but some
don't. So, it won't be automatic, but now its only 1 finger push, rather
than 2! Economy of Effort! ;->. (Say, how do I dial through Merit? I've
wondered about that. I can't connect from my Office, unless it is through
Merit.)

To connect through MERIT, you need an MTS account or some other authorized
account.  If you have an MTS account, select "dialout-aa" at the
"Which host?" prompt, then type "um/<your account id>", then your
MTS password.  That will connect you to the Ann Arbor dialout modems;
you should get an "AT" from the modem.  Then do "ATDT7613000".
If you find you're losing characters, try typing "ATSMCSLCC0SMF1SPF0"
next time.  The modem prompt will change to "CDC>", and you type
D7613000 to dial.

I don't think the "%" thing has to do wit Procomm specifically.
I use MS-Kermit, and just tried dialing in at 1200 baud a few times.
I found that every character I tried (including <return>) synchronized
things correctly.  Can some communications type person enlighten us as
to what's going on?

Re linking:  Yes, keeping a list of the items that an item is linked to
with each linked item would speed things up, though you'd probably want
to put it in a separate file rather than the item file itself, so that
routines that access and update item files wouldn't have to be changed.
There'd still be an extra cost associated with every "read new" of a
linked item, though, so I guess the question would be whether that and
the additional complexity of the software is worth the benefit.

Re linking:  This would involve some modifications to picospan, but couldn't
some type of flag be set in the item header to show it is linked?  Then
upon seeing that flag, picospan could consult a link table of all items
in that conference.  (Or even just somehow editing the header of the
message itself to give a link history)

I don't mind forgetting the item manually, but it would be nice to be
able to read the the end of the item, type links at the Respond... prompt,
getting a list of the links, then being able to say "Forget in general, arts"
or something like that.

This response has been erased.

I assume it does that with any character typed, since if it hasn't found
the right baud rate yet, it doesn't know for sure if you typed a <return>
or something else, but can tell that there's been *some* activity on
the line.  That's consistent with my experimentation, but doesn't explain
why <return> won't work for Rane.

Rane, do you have any procomm control stuff set up which is intended (say)
for Merit - a script that runs on connect or something?  I've *barely*
played with this - I like to know what I'm sending, so I don't automate -
but Procomm lets you do pretty serious stuff.

If you don't know what I'm talking about, I'd be glad to come to wherever
your machine is & munge through it with you.  There's *bound* to be someone
more knowledgeable, but I'll happily do it if you want & if no one
volunteers.  (Nothing may come of it, but worth trying, I'd say.)

Re #30 (linking):  Making *any* changes to an item file, other
than appending to the end, that changes the size of the file, gets
to be very tricky business.  Remember that Picospan runs in a
multiuser environment where more than one person may be reading or
responding to the item at the same time.

Something close to the "links" command that you propose could be
done as an external program without modifying Picospan, although
it would run slowly.

I took what downloaded with ProComm, and only set 7E1 (and maybe a
couple of other things :-<). I don't know what all the other gadgets
do, but it is working OK for me now. Thank you for the offer of
 
that is worth your time, I'll take up your offer. Well, system down in
five minutes. Bye for now.

Re #34 (linking):  Is there any way to determine if a file has links?
If so, a link table could be devised and integrated into picospan some
how.  (Although I suppose that this is a moot point since I beleive I
read somewhere that picospan is not going to be changed much)

In Unix, it's very easy to tell if a file is linked (just look at the
link count in the inode -- if it's greater than 1, it's linked) but
essentially impossible without a global search of the entire file
system to determine where those links are.  So it's easy for Picospan
to tell you that an item is linked, but to go the extra step of
finding the links efficiently would require enhancements to
Pico's record-keeping of the kind you suggest.  

Actually, a separate program could keep track of links.  It wouldn't
have to be part of Picospan.  (Any volunteers to write it?)

I would if I could but I can't so I won't.  (I was the one who said this
would be easy, right?)

     What happened in the Arts/entertainment conference such
that I lost all my read pointers?  It's not a big deal, but
I was  startled to find out that I wasn't a member.

- Backtalk version 1.3.30 - Copyright 1996-2006, Jan Wolter and Steve Weiss