From Newsgroup: news.groups.proposals

In news.groups.proposals Rayner Lucas <usenet202101@magic-cookie.co.uknospamplease> wrote:

As a contribution towards the latter, I have ported PyModerator to
Python 3 (https://github.com/PyModerator/PyModerator). It's still rather elderly and clunky, with much work to be done, but is considerably
easier to set up than the other extant moderation software, STUMP. The development version now has support for secure POP and SMTP connections, making it more likely to work with modern email providers.

The only other way I can think of to lower the barrier to entry is some
sort of hosted moderation platform, but that would be a single point of failure just like Robomod was.

If I understand correctly, the moderation software just needs to
read mail from the newsgroup's submission email inbox and post
approved messages to a willing NNTP server. In that case you could
easily have instances of the same moderation platform running in
different places, similar to front-end websites like Invidious. If
one dies, moderators could make an account on another identical
instance and keep going. If it's open-source and well written in a
long-term stable language (I wouldn't choose Python on that basis)
then it shouldn't need much maintenance even if the original author
departs.

As, I gather, a closed-source service, Robomod effectively opted in
to being a single point of failure, but I think that approach could
be done much more flexibly.

The only issue, and I'm not sure if it's an issue, might be the
NNTP servers willing to accept postings from these distributed
neo-Robomod instances. I got the impression from past discussion
that some (most?) NNTP servers don't accept moderators posting
approved articles through them, or require personal requests to
allow it. If all the instances are pointing to the same willing
NNTP server then it becomes another single point of failure.
Ideally they'd all be pointing to different NNTP servers (_ideally_
many instances would be run by the same people who run those NNTP
servers).
--
__ __
#_ < |\| |< _#

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: news.groups.proposals

On 15.06.2025 00:20 Uhr Computer Nerd Kev wrote:

In news.groups.proposals Rayner Lucas <usenet202101@magic-cookie.co.uknospamplease> wrote:

As a contribution towards the latter, I have ported PyModerator to
Python 3 (https://github.com/PyModerator/PyModerator). It's still
rather elderly and clunky, with much work to be done, but is
considerably easier to set up than the other extant moderation
software, STUMP. The development version now has support for secure
POP and SMTP connections, making it more likely to work with modern
email providers.

The only other way I can think of to lower the barrier to entry is
some sort of hosted moderation platform, but that would be a single
point of failure just like Robomod was.

If I understand correctly, the moderation software just needs to
read mail from the newsgroup's submission email inbox and post
approved messages to a willing NNTP server.

Exactly.
But that still needs somebody who approved that posts.
webstump exists. :-)

In that case you could easily have instances of the same moderation
platform running in different places, similar to front-end websites
like Invidious. If one dies, moderators could make an account on
another identical instance and keep going. If it's open-source and
well written in a long-term stable language (I wouldn't choose Python
on that basis) then it shouldn't need much maintenance even if the
original author departs.

Which still has the issue that there needs to be a process for taking
over. If anybody can do that at any time, there is no real moderation.

The only issue, and I'm not sure if it's an issue, might be the
NNTP servers willing to accept postings from these distributed
neo-Robomod instances. I got the impression from past discussion
that some (most?) NNTP servers don't accept moderators posting
approved articles through them, or require personal requests to
allow it. If all the instances are pointing to the same willing
NNTP server then it becomes another single point of failure.

This is needed to keep the concept of moderated groups. Otherwise
anybody could post messages with an Approved: header.
--
kind regards
Marco

Send spam to 1749939628muell@stinkedores.dorfdsl.de

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: news.groups.proposals

Hi Marco,

The only issue, and I'm not sure if it's an issue, might be the
NNTP servers willing to accept postings from these distributed
neo-Robomod instances. I got the impression from past discussion
that some (most?) NNTP servers don't accept moderators posting
approved articles through them, or require personal requests to
allow it.

This is needed to keep the concept of moderated groups. Otherwise
anybody could post messages with an Approved: header.

Indeed. Also, one should also take into consideration anti-spam/abuse mitigation. Ideally, setting up pgpmoose or like.
https://wiki.killfile.org/projects/usenet/pgpmoose/

Tim mentions that he can watch newsgroups and send the related NoCeM
messages to cancel unwanted articles posted with an Approved header
field in moderated newsgroups, but I don't know how many such newsgroups
are watched and whether his robot still work.
I think this is a useful service to have in Usenet moderation. And I
agree there should be more than one watching robot. Adding new NoCeM
keys can take ages as they need being manually added by news
administrators, so the best is to have current NoCeM senders doing that :)
--
Julien |eLIE

-2-aYou know what I did before I married? Anything I wanted to.-a-+

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: news.groups.proposals

In article <684e1bb7@news.ausics.net>, not@telling.you.invalid says...

In news.groups.proposals Rayner Lucas <usenet202101@magic-cookie.co.uknospamplease> wrote:

As a contribution towards the latter, I have ported PyModerator to
Python 3 (https://github.com/PyModerator/PyModerator). It's still
rather elderly and clunky, with much work to be done, but is
considerably easier to set up than the other extant moderation
software, STUMP. The development version now has support for secure
POP and SMTP connections, making it more likely to work with modern
email providers.

The only other way I can think of to lower the barrier to entry is
some sort of hosted moderation platform, but that would be a single
point of failure just like Robomod was.

If I understand correctly, the moderation software just needs to
read mail from the newsgroup's submission email inbox and post
approved messages to a willing NNTP server. In that case you could
easily have instances of the same moderation platform running in
different places, similar to front-end websites like Invidious. If
one dies, moderators could make an account on another identical
instance and keep going. If it's open-source and well written in a
long-term stable language (I wouldn't choose Python on that basis)
then it shouldn't need much maintenance even if the original author
departs.

Ah yes, Python. "Let's remove nntplib from the standard library, nobody
uses that any more". *sigh*

I like the idea of an open-source moderation platform. We have
STUMP/WebSTUMP, but it's a pain to set up.

We'd still need volunteers to run instances of the platform, but maybe
that way we'd only need a handful of technically competent people to
provide moderation services to people who are willing to do moderation
work but don't have the skills to set up their own platform.

As, I gather, a closed-source service, Robomod effectively opted in
to being a single point of failure, but I think that approach could
be done much more flexibly.

The only issue, and I'm not sure if it's an issue, might be the
NNTP servers willing to accept postings from these distributed
neo-Robomod instances. I got the impression from past discussion
that some (most?) NNTP servers don't accept moderators posting
approved articles through them, or require personal requests to
allow it. If all the instances are pointing to the same willing
NNTP server then it becomes another single point of failure.
Ideally they'd all be pointing to different NNTP servers (_ideally_
many instances would be run by the same people who run those NNTP
servers).

Panix and Eternal September are willing to allow posting of approved
articles, if the user can show they have a legit reason. If anyone knows
of other NNTP providers that will grant this permission, please let us
know, it's good to have more options we can recommend to potential
moderators.

Rayner

--- Synchronet 3.21a-Linux NewsLink 1.2

From Newsgroup: news.groups.proposals

On 2025-06-15, Computer Nerd Kev wrote:
In news.groups.proposals Rayner Lucas wrote:

I couldn't help but feel that the recent discussion at large
comes as somewhat discouraging towards prospective moderators.
And hence counter-productive, given that perhaps our best chance
at having moderator /teams/ at this point is to have more than
one person independently volunteer at the same time.

As well, criticizing someone's job without a particular tangible
goal in mind seems pointless at best. A possible such goal might
be, say, "5% increase of quality traffic across Big-8 groups in a
year." It is my personal opinion that the recent actions of the
Board increased the chances of that happening. It is also my
personal opinion that the chances are still infinitesimal due to
circumstances outside of the Board's control (such as the fairly
good so far performance of competing technologies, ActivityPub
and Matrix among them), rendering the point largely moot.

The only other way I can think of to lower the barrier to entry is
some sort of hosted moderation platform, but that would be a single
point of failure just like Robomod was.

If I understand correctly, the moderation software just needs to
read mail from the newsgroup's submission email inbox and post
approved messages to a willing NNTP server. In that case you could
easily have instances of the same moderation platform running in
different places, similar to front-end websites like Invidious.

The moderator is expected to check their mailbox and either
do that, or respond with rejection notices (or just discard
outright abuse); and what software they use, and how, to do
that, is entirely up to them. For instance, I'm going to rely
on a handful of Vim macros for the time being.

That's somewhat of a problem, as a moderation team will need
to, among other things, agree on what software to use, and the
preferences here might be highly subjective. Say, I don't mind
using a browser, Lynx mainly, to read the Web, yet using one to
do meaningful work is something I'd rather not do outside of a
paid-for job.

That said, at the core, a team needs a shared address or two
(one for submissions and another for reaching the team), /and/
one or more newsservers whose operators allow the members to
post approved articles.

It's very well possible for individual members to pick their
own software for turning an incoming email into an approved
article (or rejection notice.) About the only issue is
coordination; say, if one member rejects a submission, another
should at least be warned if they try to then approve it.
(If an article is approved more than once, it will be rejected
by the server due to a duplicate Message-Id:.)

One way to coordinate would be to use a shared incoming
mailbox: before dealing with a submission, you move it into
your own mailbox; and once you've dealt with it, you move it
into "approved" or "rejected." But there're just so many
ways of doing that that trying to market some sort of "single
best solution" is likely to fail.

A natural place to put the requisite functionality would be
a patch or an extension for a mail + news user agent: Alpine,
Gnus, Neomutt, Slrn, whatever. Some of them (Gnus, Slrn) are
readily extensible; others, I presume, will require patches.

Moderation being accessible from one's own preferred user agent
would likely increase the likelihood of one volunteering, and
yet maintaining all that code will be quite an effort. Not that
there has to be a single person doing it; anyone interested could
try implementing this for the user agent of their choice.

A particular impediment to that is the lack of standardization
when it comes to how an email submission is ought to be
transformed into an approved Usenet article (say, that the
incoming message Subject: is ought to be preserved, while
Control:, if any, must be rejected.) Researching existing
moderation software first may be necessary.

[...]

The only issue, and I'm not sure if it's an issue, might be the NNTP
servers willing to accept postings from these distributed neo-Robomod instances. I got the impression from past discussion that some
(most?) NNTP servers don't accept moderators posting approved
articles through them, or require personal requests to allow it.

There's indeed no easy way for the server to verify that the
approval is genuine, so it makes every sense for server operators
to only allow approvals to be posted from pre-verified accounts.

But perhaps the problem needs to be approached from the reverse?
Could we perhaps make a moderation platform that the newsmasters
wouldn't mind deploying as part of their news setup, given that
they already have a willing newsserver at hand, and presumably
also working email?

That said, Invidious can get away with working only with "big"
browsers, but I'm not so sure that the majority of Usenet users
will be eager to adopt some single solution for moderation when
they use such a diversity of newsreaders already.

--- Synchronet 3.21a-Linux NewsLink 1.2