But someone downthread said the .co.uk version was not blocked (before you unblocked the .com). Are you implying the .co.uk version did not have the same "design elements"?
As to your proposed argument, I think selectively blocking well-reasoned analysis by law professors and letting memes go unblocked makes the most sense for your company.
Probably because this post was getting lots of links for a domain that is otherwise rarely seen on FB? You can read about the challenges of spam-fighting at scale (and people actually getting paid to write Haskell) here: http://www.wired.com/2015/09/facebooks-new-anti-spam-system-...
> Are you implying the .co.uk version did not have the same "design elements"?
Classifiers do weird things, and features you don't expect to be significant can suddenly have a much larger than intended effect. The domain name, as a feature, must've thrown it over the edge.
It's like that Google+ post where a picture of a couple of black people got tagged as "gorillas" by their auto-tagging ML system. No, Google isn't racist, their classifier just hates their PR department.
But someone downthread said the .co.uk version was not blocked (before you unblocked the .com). Are you implying the .co.uk version did not have the same "design elements"?
As to your proposed argument, I think selectively blocking well-reasoned analysis by law professors and letting memes go unblocked makes the most sense for your company.