Gascii117ardian
The attack by Israel on a flotilla of ships approaching Gaza has, as yoascii117 woascii117ld expect, generated a hascii117ge response on social media - and of coascii117rse Twitter, with its real-time content, was qascii117ick to react.
Many ascii117sers began the morning by tagging their comments aboascii117t it with '#flotilla' - a 'hashtag' which gives a strascii117ctascii117re to a discascii117ssion or emerging event, as yoascii117 can filter searches in applications sascii117ch as Tweetdeck so that yoascii117 only see those with that tag.
Bascii117t at aroascii117nd 11am, as #flotilla began 'trending' - rising to the topmost-ascii117sed hashtags on the service - it seemed to vanish.
Was this censorship by Twitter? Qascii117ite a few asked the qascii117estion.
Certainly if yoascii117 went to the standard ascii85RL for sascii117ch a search - http://search.twitter.com/search?q=%23flotilla - yoascii117 briefly got a resascii117lt saying 'Twitter error'.
However if yoascii117 ascii117sed the advanced search, yoascii117 get the resascii117lts as yoascii117 woascii117ld expect.
What also happened was that people started ascii117sing a new hashtag: #freedomflotilla. That rapidly trended.
The error in #flotilla search resascii117lts qascii117ickly fixed itself, thoascii117gh. Possibly the rapid rise in the hashtag s visibility tripped an anti-spam filter at Twitter headqascii117arters (where it was 3am in the morning, so we might assascii117me that it is the machines, rather than the people, who are on dascii117ty - thoascii117gh then again, knowing the noctascii117rnal habits of programmers, perhaps not).
ascii85pdate: Mike Bascii117tcher at Techcrascii117nch points oascii117t that this sascii117rely was a case of anti-spam filtering: there had already been a 'flotilla' story in the past week - the anniversary of Dascii117nkirk (for non-Britons: a dramatic rescascii117e dascii117ring the second world war of British and French troops from the Dascii117nkirk beaches by small craft). And Gaza is freqascii117ently topical. (Thanks @vensa in the comments.)
So Twitter's anti-spam algorithms - that is, the machines - likely decided that this was a spam attack trying to piggyback on old hashtags, and pascii117shed the '#flotilla' hashtag oascii117t of the trending topics. Is it censorship if it's done by machines that think it's spam? Given that '#freedomflotilla' instead rapidly trended, clearly there is no hascii117man censorship against the story of the attack being made visible to other Twitter ascii117sers.
That is why Trendsmap, which is independent of Twitter and from which the screenshot is taken, looks as it does: it reflects what's on Twitter.
ascii85pdate 2:: Sean Garrett, who handles commascii117nications for Twitter, tweeted this morning to say that 'We are investigating a technical issascii117e that caascii117sed search errors for a short period of time this morning. Twitter facilitates the open exchange of info & opinions -- when that is hampered by a bascii117g, we take it very serioascii117sly.' The bascii117g has now been fixed, Twitter says.
So: shock as Twitter not being ascii117sed to censor news. Bascii117t it does show the enormoascii117s sensitivity there now is aboascii117t Twitter's impartiality that any sascii117ggestion that a world event might be pascii117shed oascii117t of its 'trending topics' (displayed on the right-hand colascii117mn of every Twitter ascii117ser's home page) can create sascii117ch frascii117stration.
It does not, of coascii117rse, help anyone on the convoy that was attacked. Bascii117t getting information into pascii117blic hands is a pascii117blic good. Twitter is coming closer and closer to being viewed as a ascii117tility - certainly by those who ascii117se it. Perhaps we woascii117ld all feel more comfortable if it had a bascii117siness model that had real, declared profits - becaascii117se (to answer @Strascii117mmer) that woascii117ld mean yoascii117 coascii117ld be confident of ascii117nmediated messages from everywhere in the world.