[net.news] cross-posting restrictions

msb@dciem.UUCP (Mark Brader) (08/15/86)

Geoff Collyer (geoff@utcs.UUCP) writes in ont.general:
> utcs is running C news, which has implemented some of the policy in B
> 2.11 news.  One notable feature is that, in order to prevent
> cross-postings such as (in the brave new world) "comp.unix.wizards,
> rec.auto" ... B 2.11 and C news insist that all newsgroups in the
> Newsgroups: line must be in the same "newsgroup class" (i.e. must all
> start with the same word).  This means that cross-postings such as
> ont.general,tor.general will be quietly dropped ...

This is a FEATURE?  I can see that a site which receives only comp,
or only rec (!), might want to reject such an article, but why on
earth should a site that receives both comp and rec want to reject it?

If it is an attempt to keep articles that are really rec (or talk) out
of the comp classification where they will get wider distribution, it is
terribly misguided ... there is nothing to stop people from simply posting
them twice, and the same social mechanisms for control will apply as now.

On the other hand, when an article is legitimately cross-posted -- such
as this one, which is going to ont.general (ont = province of Ontario)
where the point was raised, and net.news where it should have been raised
FOR DISCUSSION BEFORE implementing such a radical change... well, it will
also have to be posted twice, and that's a pain to those reading both places.

Geoff's signature quote was:
> ``Lack of planning on your part does not justify an emergency on mine.''

I'm supposed to plan for quiet additions of misfeatures?  I want this
changed, FAST, before 2.11 and C news do become widely installed!

Mark Brader

jbuck@epimass.UUCP (Joe Buck) (08/19/86)

In article <1919@dciem.UUCP> msb@dciem.UUCP (Mark Brader) writes:
>Geoff Collyer (geoff@utcs.UUCP) writes in ont.general:
>> utcs is running C news, which has implemented some of the policy in B
>> 2.11 news.  One notable feature is that, in order to prevent
>> cross-postings such as (in the brave new world) "comp.unix.wizards,
>> rec.auto" ... B 2.11 and C news insist that all newsgroups in the
>> Newsgroups: line must be in the same "newsgroup class" (i.e. must all
>> start with the same word).  This means that cross-postings such as
>> ont.general,tor.general will be quietly dropped ...
>
>This is a FEATURE?  I can see that a site which receives only comp,
>or only rec (!), might want to reject such an article, but why on
>earth should a site that receives both comp and rec want to reject it?

This is a very bad feature; it prevents people from doing many things
that are quite legitimate.  The boundary between "comp" and "sci",
for example, is not hard: I can think of many legitimate topics
where a cross-posting between sci.math and comp.<various> is
appropriate.

One result will be an increase in net traffic, since the same article
will in some cases be posted twice, instead of cross-posted, to
achieve the same effect.

It also would prevent someone from cross-posting to a general local
list (say ba.general) and a more limited net-wide list.  Some
articles will receive too wide a distribution as well, since some
sites belong to two local distributions where neither one is a subset
of the other.

It's actually quite radical in that is removes any connection between
the various distributions; "comp" and "sci" will become completely
separate entities.
-- 
- Joe Buck 	{ihnp4!pesnta,oliveb,nsc!csi}!epimass!jbuck
  Entropic Processing, Inc., Cupertino, California

werner@utastro.UUCP (Werner Uhrig) (08/19/86)

Recently, I wanted to post an article to both mod.something (for moderated
netwide distribution) and to local.something so local users would not have
to wait the week or so until it would actually show up in the moderated
group.   Another time, I wanted to bypass the delay of moderation by posting
to both  mod.whatever  and  net.whatever.  I was convinced that this was a
good idea, as I felt certain that those people that wouldn't have to wait the
extra time until they'd see the message would appreciate the fact;  and those
who wouldn't get to see the article in  net.whatever  because some backbone
upstream only carried  mod.whatever  would say "better late than never".

I can't say that I appreciated the cast-in-software policy that prevents
us from doing this and without a warning-message at that.   The first time
around I decided to wait for the moderated article to show up:  well, 3 weeks
later the moderator posted a message saying that he had decided (based on a
misunderstanding) not to distribute the article in the mod.group.  I'm sure
you *KNOW* what I did when I had the second case occur ...

Given that only 2 times during the last 6 months did I feel reason to cross-
post, I don't really feel it to be too great a burden having to post such
articles twice (seperately to both groups), but I *DO* object to not getting
warned about the "software block" - a warning with default "blocking" unless
some override action is taken, that would be preferable.  It would help if
there were some statistics indicating how many messages are actually blocked,
saving us the aggrevation of *HAVING* to read inappropriate articles or
duplicate messages.  I wonder if that really happens with any frequency
which would justify aggrevating the average, sensible net-user ....??!!

---Werner

msb@dciem.UUCP (Mark Brader) (08/20/86)

> Geoff Collyer (geoff@utcs.UUCP) wrote in ont.general:
> > ... in order to prevent
> > cross-postings such as (in the brave new world) "comp.unix.wizards,
> > rec.auto" ... B 2.11 and C news insist that all newsgroups in the
> > Newsgroups: line must be in the same "newsgroup class" ... cross-postings
> > such as ont.general,tor.general will be quietly dropped ...

And I, and now others, have complained about this as a serious misfeature.

I now add a couple of points that I didn't think of in the original article.
The first is to say "thank you" for Geoff for bringing this restriction to
public attention.  I can't say whether it was an oversight or was
intended to be brought out as a fait accompli like the net reorganization
(which I am all in favor of and which clearly had to be done that way), 
but the effect in this case is that of a BUG.

There is no way that news software should be QUIETLY deleting hitherto
valid articles!  Let alone beginning to do so without warning.  This ranks
with Jamie's Junker as one of the most misguided attempts to do good that
the net has been faced with.  And I would be proposing, as was the consensus
on the Junker, that any site running the change should be cut off by its
neighbors -- if not for the fact that in this case we would apparently
be losing part or all of the backbone by doing so.

Mark Brader
	Without the threat of frequent new releases of the system to
	enforce conformity, we have been free to modify and adapt the
	system to suit our own purposes.  ...  On more than one occasion,
	we have found it has been quicker to correct a newly discovered
	program error than to document its existence.  We feel we are in
	a relatively advantageous position compared with users of other
	brands of software.				-- John Lions

hoffman@hdsvx1.UUCP (08/27/86)

Why don't we write some new software for the net which would allow
"linked" cross-posting?  That is, cross-post only once, but only one
copy of the posting goes across, accompanied by pointers for the
cross-posted groups.  An important requirement of such a system would
be the ability to mark something read *once and for all* -- if I read
it in net.legal, I don't even want to see it when I get to net.politics.
I think that such a system would eliminate all objections to cross-posting
except for the one that cross-posting encourages people to post opinions
to inappropriate groups.  But people will do that no matter what we do.
-- 
 Richard Hoffman                | "Oh life is a wonderful cycle of song,
 Schlumberger Well Services     |     A medley of extemporanea.
 hoffman%hdsvx1@slb-doll.csnet  | And Love is a thing that can never go wrong
 PO Box 2175, Houston, TX 77252 | ... And I am Marie of Roumania." --D. PARKER

campbell@maynard.UUCP (Larry Campbell) (08/30/86)

In article <461@hdsvx1.UUCP> hoffman@hdsvx1.UUCP (Richard Hoffman) writes:
>Why don't we write some new software for the net which would allow
>"linked" cross-posting?  That is, cross-post only once, but only one
>copy of the posting goes across, accompanied by pointers for the
>cross-posted groups.  An important requirement of such a system would
>be the ability to mark something read *once and for all* -- if I read
>it in net.legal, I don't even want to see it when I get to net.politics.

This is already true!  So don't go implementing it, please!  When you
cross-post, only one copy of your article is transmitted across a link.
If you use "rn" to read news (why use anything else?) and have applied
the "Xref" patch to the news software (the patch comes with "rn") then
when you read an article, it's marked read, and you won't see it even
if it's cross-posted to subsequently read newsgroups.
-- 
Larry Campbell                             The Boston Software Works, Inc.
ARPA: campbell%maynard.uucp@harvard.ARPA   120 Fulton Street, Boston MA 02109
UUCP: {alliant,wjh12}!maynard!campbell     (617) 367-6846