From: Charles Lindsey (chl@clw.cs.man.ac.uk)
Date: Sat Jun 08 2002 - 14:30:44 CDT
On Wed, 05 Jun 2002 04:26:42 -0700
Russ Allbery <rra@stanford.edu> said...
> > But tell me, how many groups are actually hand moderated these days?
>
> I'd say most of the moderated groups are still hand-moderated, although
> there are a growing number of groups moderated by nearly as many different
> automated programs as there are groups.
>
> > Is it not much more common to use a bot (e.g. stump) which applies some
> > filtering rules (crosspost limits, blacklists, etc) and then posts it
> > automatically?
>
> Nope.
That surprises me somewhat. In uk.*, it is almost universal for groups
to be moderated by means of a filter-bot such as stump.
Anyway, time to summarize the discussion on the moderators' list. Cc to
the Usefor mailing list.
1. Usefor is proposing that articles sent to moderators should be
encapsulated within the email as a complete news article, ready to
post. The content-Type should be set to application/news-transmission.
However, moderators SHOULD be prepared to accept articles in the style
that predominates currently (sending the complete article as a mail
message) for a long overlap period (a couple of years at least, and
likely much more).
2. The benefits of this style are
a) Message sent by mail are subject to mangling on route. It is
quite legitimate for headers to be reordered, news headers added or
deleted, Content-Transfer-Encoding to be changed, RFC 2047 invoked,
etc. etc. For Netnews, Usefor requires a strict "no tinkering"
policy.
b) An encapsulated article can be "tunneled" though a mail system,
being Content-Transfer-Encoded en route and reconstituted in exactly
its orignal form at the far end.
c) With the advent of news headers using UTF-8 (so as to accomodate
newsgroup-names, names of senders, etc. in strange foreign character
sets), a news article could not in general be sent as an email
without first encoding using RFC 2047, which then MUST be undone at
the far end (since newsgroup-names MUST NOT be in RFC 2047 format,
and other stuff SHOULD NOT).
3. This style was strongly promoted in Son-of-1036, and Henry Spencer
wrote CNews to with the intention of using it. Even Russ agrees that it
is a Good Thing (his reservations are concerned with the transitional
period).
4. The actual behaviour of CNews depends on the mailer it feeds into. If
the mailer expects a complete email message on stdin, it encapsulates.
If it expects just a body, it does not. Which kind of mailer you
get depends on the operating system, how your Path is set up, and
whether there is an 'R' in the month. Nevertheless, some CNews systems
undoubtedly DO encapsulate at the present time, and I have never had
an article rejected or fail to be posted for that reason. Whether the
moderators concerned curse and swear at me behind their closed doors, I
wouldn't know :-( .
5. If it is accepted that the new style is cleaner, better and
desirable, then the question to be asked is "how do we get to that state
eventually, given the state we are currently in". There seems to be
no magic bullet to ease it in gradually. People will start using the
new style, moderators will start seeing it, and they will have to cut
and paste, or re-hack their scripts, or whatever (but at least that is
possible with hand moderation). I think we have to accept that there
is no Gain without Pain, but at least once a moderator has learnt the
trick, it should be easy to apply it routinely in the future.
6. It may even be the case that having it encapsulated with the
special Content-Type will make it easier to process than having it
not encapsulated (depending on the MIME features provided with the
particular user agent). Note that the proposal is to provide that
Content-Type for the email message as a whole, though I have no doubt
we shall also see mails sent as multipart/mixed with a single part with
that Content-Type. I reckon that should be acceptable also.
7. The question remains of the effect on filter-bots. There are
no reports as to whether stump, for example, currently accepts
encapsulation CNews-style. It could easily be rewritten to do so, of
course, so the problem would come down to propagation of the revised
script.
Charles H. Lindsey ---------At Home, doing my own thing------------------------
Tel: +44 161 436 6131 Fax: +44 161 436 6133 Web: http://www.cs.man.ac.uk/~chl
Email: chl@clw.cs.man.ac.uk Snail: 5 Clerewood Ave, CHEADLE, SK8 3JU, U.K.
PGP: 2C15F1A9 Fingerprint: 73 6D C2 51 93 A0 01 E7 65 E8 64 7E 14 A4 AB A5