Wrolf, Facebook (et al.) already have extremely powerful engines and many engineers working on anti-spam/anti-fraud technologies. They're quite good at keeping most of the spam out of your Timeline. They don't need "our" help.
The same techniques could plausibly be used to block ISIS propaganda, or the propaganda of any cause or organization deemed harmful. Whether this should be done is a topic out of scope for this forum. Regards, --Jered ----- On Dec 16, 2015, at 1:34 PM, Wrolf <wr...@wrolf.net> wrote: > Thanks Bill. I guess I should restate my question. > Would it be practical for > Twitter/Facebook/SnapChat/WhatsApp/Microsoft/Telegraph > to use SpamAssassin like techniques of Bayesian filtering and RBL lists to > block ISIS on social media? > This is an invitation for discussion, not a rhetorical question. Feel free to > repost to more appropriate forums. > Wrolf > Wrolf > wr...@wrolf.net > On Wed, Dec 16, 2015 at 1:22 PM, Bill Cole < > sausers-20150...@billmail.scconsult.com > wrote: >> On 15 Dec 2015, at 23:19, Wrolf wrote: >>> Stop me if you've heard this one. >>> Would it be practical to use the Spamassassin techniques of Bayesian >>> filtering and RBL lists to block ISIS on social media? >> I've definitely heard similarly unfunny and poorly thought-out jokes before. >> Bill Gates had one called "Penny Black" which seems to have taught him that >> his >> wealth would would be more helpful if spread out across a larger set of >> hands... >> The only reason email spam requires tools like DNSBLs, Naive Bayesian >> classification, shared hash databases (DCC/Razor/Pyzor/IxHash) etc. is that >> email is handled by many thousands of autonomous mail systems using a >> protocol >> that by spec and default falls back to a protocol defined and designed for >> the >> 1980's Internet. The collaborative tactics that are assembled by SpamAssassin >> only make any sense to use AT ALL because email uses an open protocol where >> no >> single operator or end-user sees all the traffic and no two see the same >> slice >> of traffic. SpamAssassin in combination with other tools is pretty good >> considering that challenge, but it totally sucks compared to what is possible >> for communication systems with unified administrative control. >> Excluding any set of @BadGuys from Twitter is within Twitter's power NOW. >> Excluding any set of @BadGuys from Facebook is within Facebook's power NOW. >> Excluding any set of @BadGuys from SnapChat is within SnapChat's power NOW. >> Excluding any set of @BadGuys from WhatsApp is within WhatsApp's power NOW. >> Excluding any set of @BadGuys from Skype is within Microsoft's power NOW. >> ISIS uses any "social media" where the proprietors welcome them. That is a >> business decision of for-profit private enterprises based in >> lightly-regulated >> jurisdictions (mostly the US and EU) who mostly have not thought about that >> choice in those terms. Whether they should is a complex question for other >> fora... >> Put another way, in specific terms: >> I use SpamAssassin in conjunction with Postfix, MIMEDefang, various OS >> components, and my own chicken-wire+duct-tape code in C/Perl/Shell to exclude >> spammers *AND* ISIS (incidentally) from my own mail system. I use other >> overlapping toolsets to do similar exclusions from mail systems whose owners >> authorize and mostly pay me to do so. Twitter does not authorize or pay me to >> manage their systems, so I won't try to do so. I'm quite sure that if they >> were >> crazy enough to engage me for that role, I would use stronger tactics than I >> CAN with email, using almost entirely different tools. On the other hand, the >> policies and tactics needed to exclude spammers and/or ISIS from Twitter >> might >> conflict with Twitter's business plans or the ethical principles (LOL) of >> their >> corporate owners & leadership. Twitter has people whose jobs are policing how >> Twitter is used. They do a shit job, because Twitter CHOOSES to not fund and >> require the performance of a good job.
