[Wikimedia-l] PRISM

Tobias church.of.emacs.ml at googlemail.com
Mon Jun 10 09:18:10 UTC 2013

On 06/10/2013 08:49 AM, Federico Leva (Nemo) wrote:
> Benjamin Lees, 10/06/2013 08:13:
>> http://thread.gmane.org/gmane.org.wikimedia.foundation/49712/focus=49727
>> is
>> probably relevant (if what Domas said then is still true).
> While I'm not aware of privacy changing substantially, speaking of
> fantastic names, Kraken is going to change things a bit compared to 2010:
> https://www.mediawiki.org/wiki/Analytics/Kraken/Request_Logging
> https://www.mediawiki.org/wiki/Analytics/Kraken/Data_Formats
> I didn't find a human-readable overview but the gist seems to be that
> WMF will log the same (partial) data, but for 100 % of visits rather
> than 1/1000.
> More technical members of the list will be able to tell more from the
> specifications and source code.

Interesting... I couldn't really find much information about the privacy
concepts of Kraken, though the flow diagram suggests that the raw data
(which I suppose includes the kind of data we discussed earlier, i.e.
IP, time and date, accessed content, ...) is kept for 7 days until it is
anonymized. Is that true? If so, it seems like a huge mistake to me.

-- Tobias

More information about the Wikimedia-l mailing list