[Operators] Annoying spam

Mickaël Rémond mremond at process-one.net
Thu Nov 12 21:22:35 UTC 2015


> On 09 Sep 2015, at 16:45, Peter Saint-Andre - &yet <peter at andyet.net> wrote:
> On 9/9/15 8:23 AM, Peter Saint-Andre - &yet wrote:
>> It might be time to revisit the old spam-prevention specs we started to
>> define years ago.
> For the record, those were things like:
> http://xmpp.org/extensions/xep-0159.html
> http://xmpp.org/extensions/xep-0267.html
> http://xmpp.org/extensions/xep-0268.html
> http://xmpp.org/extensions/xep-0275.html
> I haven't looked at them in a long time, so I'm not sure if I still think they might work, if they're too complex, etc.

I agree it seems a good moment to revise the spam-prevention spec. I am willing to help.

I was toying with various criteria to build local user reputation system that could be used for a server to flag possible outgoing traffic for example (XEP-0275 is a good start). It does not prevent totally bad domain to spam, but if you trust some server, you can for example flag users with good reputation to give a hint that the request could be legitimate. This could be handy for outgoing subscription requests.
To build that reputation system, we need a lot of cooperation tools in place, spam reporting, etc.

Anyway, probably not the best threat / ML to get into details, but I would be interested in helping out.

Mickaël Rémond

More information about the Operators mailing list