Uzi Paz | 1 Apr 07:49 1998
Picon
Picon

Followup-to splitting

This message is related to the beginning of a thread under the name of
"21. Newsgroups and 23. Followup-To", which held around Aug 1997.

I wish to make a suggestion for the "Followup-To:" problem.

I shall start with an example:
---
Let's say I wish to recall the name of a movie about cats that I saw long
ago. either movie-fans or cats-fans may know about the movie. Let us see
what would happen if I crosspost my question to both newsgroups:

readers of (say) rec.fan.cats would mention some names of films and
immediately develop some discussion on how "real" is the behavior of the
cats there (ending up with discussion about the behavior of Persian
cats), while the readers of, say, rec.fan.movies would develop a thread
about other films made by that director.

In my experience (and this was mentioned in that thread) most people just
hit the Reply (or Followup) key, and whatever goes goes, so both threads
would usually come up in both newsgroups, adding-up to a nice party.

How can I split the threads of my post without posting two different
messages, each to each newsgroup? Here is my suggestion:
----

Suggesting a new optional header field, say:
"Followup-to-recent: yes"

Whenever a reader tries to reply-to-group (/Followup) to a message which
has such a header-field, the news-user-agent should act as follows:
(Continue reading)

Greg Berigan | 1 Apr 07:15 1998

Re: Followup-to splitting

Uzi Paz  <uzipaz <at> bgumail.bgu.ac.il> wrote:

>Suggesting a new optional header field, say:
>"Followup-to-recent: yes"

No, not another boolean header.  Rather than that, how about:

	Followup-To: recent

where "recent" is a new special keyword.  Or maybe some other keyword.

Or you could say in the message body that posters should not crosspost
their responses, or just multipost the original message.  An
"Also-Posted-To" header or something like it could be used to identify it
as such.

It sounds like a header in search of a problem.

Uzi Paz | 1 Apr 21:04 1998
Picon
Picon

Re: Followup-to splitting

> Date: Tue, 31 Mar 1998 23:15:27 -0600
> From: Greg Berigan <gberigan <at> cse.unl.edu>
> 
> Uzi Paz  <uzipaz <at> bgumail.bgu.ac.il> wrote:
> 
> >Suggesting a new optional header field, say:
> >"Followup-to-recent: yes"
> 
> No, not another boolean header.  Rather than that, how about:
> 
> 	Followup-To: recent
> 
> where "recent" is a new special keyword.  Or maybe some other keyword.

News servers which not yet identify such a keyword may reject the message
as "such a newsgroup is not supported/exists".

I find a necessity to limit required fields. To add a header field to
solve such cases, is not so bad. It is an optional field.

-Uzi

Charles Lindsey | 2 Apr 12:35 1998
Picon
Picon

Re: Encoding of newsgroup descriptions (was: <0.2> newgroup, rmgroup, ...)

In <6qtsFCcZcDB <at> faerber.muc.de> usenet-format-list <at> faerber.muc.de
(=?ISO-8859-1?Q?Claus_Andr=E9_F=E4rber?=) writes:

>BTW, it must be the description only that can be in RFC2047, not the  
>whole newsgroup line. The newsgroup name and the "(Moderated)" string,  
>which are not free form text, but have a meaning to implementations MUST  
>NOT be encoded this way.

But to have two different encodings within the same line is just CRAZY.
UTF-8 is going to have to work within the newsgroup lines because that is
what newsgroup names are written in (generally speaking). So much simpler
to use UTF-8 throughout the line. User agents will translate it to
something displayable insofar as they are able (user agents already need
to be able to understand UTF-8 - at least so far as translating into the
character sets known to them).

>Why should they? Those parts of control messages are primarily for  
>machines (i.e. servers that update their group configuration), not  
>humans. If a user views the raw headers of a plain message, s/he might  
>also see encoded RFC 2047 text fields.

Humans read control messages for all sorts of reasons. Sysadmins because
they want to decide whether to take the group. Other people who have a
look at the charter which is usually included (Named-Articles will not be
in widespread use for some years yet). Years later, people will look them
up on the Uunet archive to see who created a particular group, and why and
when. RFC2047 texts in the middle of the body of a message are unlikely to
be translated by User agents, so people looking at these control messages
will have to translate by hand (or whatever hacks they can devise).

(Continue reading)

Claus André Färber | 2 Apr 13:08 1998
Picon

Re: Followup-to splitting

Uzi Paz <uzipaz <at> bgumail.bgu.ac.il> schrieb:
> > Date: Tue, 31 Mar 1998 23:15:27 -0600
> > From: Greg Berigan <gberigan <at> cse.unl.edu>
> >
> > Uzi Paz  <uzipaz <at> bgumail.bgu.ac.il> wrote:
> >
> > >Suggesting a new optional header field, say:
> > >"Followup-to-recent: yes"
> >
> > No, not another boolean header.  Rather than that, how about:
> >
> > 	Followup-To: recent
> >
> > where "recent" is a new special keyword.  Or maybe some other keyword.
>
> News servers which not yet identify such a keyword may reject the message
> as "such a newsgroup is not supported/exists".
>
> I find a necessity to limit required fields. To add a header field to
> solve such cases, is not so bad. It is an optional field.

What about allowing multiple Followup-To fields, each giving a set of  
newsgroups the user sending a followup can chose from.

Example:
 Newsgroups: rec.pets.dogs,, rec.food.chinese
 Followup-To: rec.pets.dogs
 Followup-To: rec.food.chinese

--

-- 
(Continue reading)

Simon Lyall | 2 Apr 13:19 1998
Picon

Dan Ritter - MIA?

Has anyone heard anything from Dan Ritter in the last week or so? I have
emailed him to query on how his putting togeather of the draft was going
but I've recieved nothing for a couple of weeks. My email may be having
problems however..

--

-- 
Simon J. Lyall.  |   Very  Busy  |   Mail: simon <at> darkmere.gen.nz
"To stay awake all night adds a day to your life" - Stilgar | MT.

Uzi Paz | 3 Apr 01:45 1998
Picon
Picon

Re: Followup-to splitting

On 2 Apr 1998, [ISO-8859-1] Claus Andr Frber wrote:

> What about allowing multiple Followup-To fields, each giving a set of  
> newsgroups the user sending a followup can chose from.
> 
> Example:
>  Newsgroups: rec.pets.dogs,, rec.food.chinese
>  Followup-To: rec.pets.dogs
>  Followup-To: rec.food.chinese

This may serve for grouping. It doesn't use new types of header fields,
but would result in more header fields. I do think, however that my
suggestion is more compact and its use is more clear. It is not too hard
to implement grouping into it:
Followup-To-Recent: <rec.pets.cats,alt.fan.cats>,rec.movies.misc

If UA identify the user as reading the message from one of the *.cats
newsgroups,
then the "Newsgroups:" of the reply would be set to both of the *.cats
newsgroups.
If it identifies the user as linking to the message from the *.movies
newsgroup it would set it to that newsgroup. If it cannot identify from
which newsgroup the reader linked to the message, it would either prompt
the reader or just ignore the header field, acting upon other fields that
are on the message (such as the Followup-To).

Uzi

John Moreno | 3 Apr 02:34 1998

Re: Followup-to splitting

Uzi Paz <uzipaz <at> bgumail.bgu.ac.il> wrote:

> On 2 Apr 1998, [ISO-8859-1] Claus André Färber wrote:
> 
> > What about allowing multiple Followup-To fields, each giving a set of
> > newsgroups the user sending a followup can chose from.
> > 
> > Example:
> >  Newsgroups: rec.pets.dogs,, rec.food.chinese
> >  Followup-To: rec.pets.dogs
> >  Followup-To: rec.food.chinese
> 
> This may serve for grouping. It doesn't use new types of header fields,
> but would result in more header fields. I do think, however that my
> suggestion is more compact and its use is more clear. It is not too hard
> to implement grouping into it:
> Followup-To-Recent: <rec.pets.cats,alt.fan.cats>,rec.movies.misc
> 
> If UA identify the user as reading the message from one of the *.cats
> newsgroups, then the "Newsgroups:" of the reply would be set to both of
> the *.cats newsgroups. If it identifies the user as linking to the message
> from the *.movies newsgroup it would set it to that newsgroup. If it
> cannot identify from which newsgroup the reader linked to the message, it
> would either prompt the reader or just ignore the header field, acting
> upon other fields that are on the message (such as the Followup-To).

Much better in my opinion to simply have the UA include the ng that the
reader is reading and include that in the attribution and then allow the
users to decide how followups should be set (can't dogowners be
interested in chinese food?).
(Continue reading)

Jonathan Grobe | 3 Apr 04:23 1998
Picon
Picon

Do We Need a New Editor (was Re: Dan Ritter - MIA?)?

Here is a partial post from him from March 31:

Subject:      Re: Anybody out there.
From:         dritter <at> pasilla.bbnplanet.com (Dan Ritter)
Date:         1998/03/31
Newsgroups:   bofh.config,bofh.leeks

[as I ponder when the next fix of the new Usenet Article Format RFC will be
ready fro release...]

So he hasn't disappeared.

I feel that we need a different document editor though. Primarily because
(as I remember his earlier posts) Dan seems to feel that his job is a
copy editing type of job while I feel we someone need someone much more
involved in writing/rewriting the whole draft. If we have this then the
result will be a more internally consistant/cohesive product--instead of
simply the tacking together of the drafts of the various authors. 

Does anyone want to devote a major amount of time to doing this?
I think Charles Lindsey would be a excellent choice if he would
be willing.

Jonathan Grobe. 

On Thu, 2 Apr 1998, Simon Lyall wrote:

> Has anyone heard anything from Dan Ritter in the last week or so? I have
> emailed him to query on how his putting togeather of the draft was going
> but I've recieved nothing for a couple of weeks. My email may be having
(Continue reading)

Claus André Färber | 3 Apr 07:26 1998
Picon

Re: Followup-to splitting

Uzi Paz <uzipaz <at> bgumail.bgu.ac.il> schrieb:
> This may serve for grouping. It doesn't use new types of header fields,
> but would result in more header fields. I do think, however that my
> suggestion is more compact and its use is more clear. It is not too hard
> to implement grouping into it:
> Followup-To-Recent: <rec.pets.cats,alt.fan.cats>,rec.movies.misc

No this would be deadly. Old user agents would try to send news to  
"<rec.pets.cats" and "alt.fan.cats>".

With two (or more) Followup-To fields, there's a chance that old UA will  
just use the first FT field.

--

-- 
Claus Andre Faerber <http://www.muc.de/~cfaerber/> Fax: +49_8061_3361
PGP: ID=1024/527CADCD FP=12 20 49 F3 E1 04 9E 9E  25 56 69 A5 C6 A0 C9 DC


Gmane