nocem in Gnus: should it go away?

nocem in Gnus: should it go away?

[Ted Zlatanov]

/gnus/lisp> grep -i nocem *.el|wc
    165     754    9124

That's a lot of code for a feature no one uses.  Should it go away?
Or should it be revisited, updated, and made useful?  Maybe it can
integrate with spam.el for instance.

Ted

Re: nocem in Gnus: should it go away?

[David Z Maze]

Ted Zlatanov <tzz@lifelogs.com> writes:

> /gnus/lisp> grep -i nocem *.el|wc
>     165     754    9124
>
> That's a lot of code for a feature no one uses.  Should it go away?
> Or should it be revisited, updated, and made useful?  Maybe it can
> integrate with spam.el for instance.

My $0.02 is that "classic" Usenet spam has practically vanished (for
the...five? public NNTP groups I read, three of which are moderated).
I tried using NoCeM for a bit but the overhead of running it was just
insane, even given a news server pretty close to me on the network.
Setting gnus-spam-autodetect to scan NNTP groups too seems to give
pretty good results at a much lower performance overhead.

If you were trying to read "most" of Usenet using Gnus, it might still
be useful to you.  But that's an awful lot of traffic.  Personally, I
wouldn't miss nocem if it vanished, and I could see it wanting to
vanish if the code maintenance issues are starting to stack up.

  --dzm

Re: nocem in Gnus: should it go away?

[Katsumi Yamaoka]

>>>>> In <g69fybz1o6o.fsf@lifelogs.com> Ted Zlatanov wrote:

> /gnus/lisp> grep -i nocem *.el|wc
>     165     754    9124

> That's a lot of code for a feature no one uses.  Should it go away?
> Or should it be revisited, updated, and made useful?

I still use NoCeM.  It is mostly for the fj hierarchy, however
there's no useful article in fj.news.lists.filters recently (the
last one was posted at 29 Aug).  So, I'll not be troubled even
if the feature is removed.

Please note that gnus-nocem.el was incomplete until I fixed it,
as far as I knew.

> Maybe it can integrate with spam.el for instance.

That's great.

Re: nocem in Gnus: should it go away?

[Ted Zlatanov]

On  3 Dec 2006, yamaoka@jpl.org wrote:

>>>>> In <g69fybz1o6o.fsf@lifelogs.com> Ted Zlatanov wrote:
>
>> /gnus/lisp> grep -i nocem *.el|wc
>> 165     754    9124
>
>> That's a lot of code for a feature no one uses.  Should it go away?
>> Or should it be revisited, updated, and made useful?
>
> I still use NoCeM.  It is mostly for the fj hierarchy, however
> there's no useful article in fj.news.lists.filters recently (the
> last one was posted at 29 Aug).  So, I'll not be troubled even
> if the feature is removed.
>
> Please note that gnus-nocem.el was incomplete until I fixed it,
> as far as I knew.
>
>> Maybe it can integrate with spam.el for instance.
>
> That's great.

Well, spam.el can add an exit processor for reporting spam that would
post in the appropriate newsgroup.  Do you know how this works, or is
there a document?  I would also be interested in finding out how you
can undo such a posting, and if there's a register/unregister process
for ham through NoCeM.  If Gnus already has such functionality, I
don't know it.

Thanks
Ted

Re: nocem in Gnus: should it go away?

[Ted Zlatanov]

On  1 Dec 2006, dmaze@mit.edu wrote:

> Ted Zlatanov <tzz@lifelogs.com> writes:
>
>> /gnus/lisp> grep -i nocem *.el|wc
>> 165     754    9124
>>
>> That's a lot of code for a feature no one uses.  Should it go away?
>> Or should it be revisited, updated, and made useful?  Maybe it can
>> integrate with spam.el for instance.
>
> My $0.02 is that "classic" Usenet spam has practically vanished (for
> the...five? public NNTP groups I read, three of which are
> moderated).

It's terrible on Gmane, though, and I use that for many lists :)

> I tried using NoCeM for a bit but the overhead of running it was just
> insane, even given a news server pretty close to me on the network.
> Setting gnus-spam-autodetect to scan NNTP groups too seems to give
> pretty good results at a much lower performance overhead.

Yes, I like it too.  Very fast on Gmane because of their anti-spam
system.

> If you were trying to read "most" of Usenet using Gnus, it might still
> be useful to you.  But that's an awful lot of traffic.  Personally, I
> wouldn't miss nocem if it vanished, and I could see it wanting to
> vanish if the code maintenance issues are starting to stack up.

I'll see what Katsumi Yamaoka says.  My feeling is that it could be
useful, but not too many care.

Ted

Re: nocem in Gnus: should it go away?

[Katsumi Yamaoka]

>>>>> In <g69vekeffl8.fsf@lifelogs.com> Ted Zlatanov wrote:

>>> Maybe it can integrate with spam.el for instance.

>> That's great.

NoCeM posts offer Message-IDs of spam articles, which are
categorized into advertisements, binary posts in non-binary
groups, multi-posts, etc.  What users need to do is to tell
gnus-nocem which posters are trustworthy and which categories
are useful.  I only thought that we will probably be able to use
those MIDs as a blacklist or equivalents.  Though it is just what
gnus-nocem does.

> Well, spam.el can add an exit processor for reporting spam that
> would post in the appropriate newsgroup.

I've never thought about posting NoCeM messages by myself.  But
that's a good idea and seems to be possible since at least
alt.nocem.misc, fj.news.lists.filters, and news.lists.filters
are not moderated groups.  Moreover we can ask Lars to create
the gmane.nocem group when we have to.

> Do you know how this works, or is there a document?

Cancelmoose <http://www.cm.org/>, for example, describes it in
detail.

> I would also be interested in finding out how you can undo such a
> posting, and if there's a register/unregister process for ham
> through NoCeM.

I don't know except cancelling.

> If Gnus already has such functionality, I don't know it.

There's no such function AFAIK.  Anyway I will necessarily add
"Ted" to `gnus-nocem-issuers' if you post nocem. ;-)

Regards,