Archive for March 22nd, 2010

links for 2010-03-22

Monday, March 22nd, 2010
  • NyTimes article on how the control over what's private and not is slipping out of our hands:

    “Personal privacy is no longer an individual thing,” said Harold Abelson, the computer science professor at M.I.T. “In today’s online world, what your mother told you is true, only more so: people really can judge you by your friends.”
    Collected together, the pool of information about each individual can form a distinctive “social signature,” researchers say.
    [...]
    The F.T.C. and Congress are weighing steps like tighter industry requirements and the creation of a “do not track” list, similar to the federal “do not call” list, to stop online monitoring.

  • French research that describes how to infer search history from personalised search suggestions. Th issue is still not totally resolved.

    "The Web History is used to provide personalized results and keyword suggestions for searches that a user has already made. We design the Historiographer, a novel attack that reconstructs the web search history of Google users, even though this service is supposedly protected from session hijacking by a stricter access control policy. The Historiographer uses a reconstruction technique to infer search history from the personalized suggestions fed by the Google search engine. "

  • On the telco engineer that provided the evidence for EFF's lawsuit against AT&T for their complicity in illegal government spying:

    "Even though I'd heard Mark Klein's story before, I'd never considered just how frightening and surreal his experience must have been. His new memoir reads like something out of a kafka-esque sci-fi spy thriller — except that it all really happened right here in the USA, just a few years ago."

  • "Cookie Synching refers to the process of mapping user Ids from one system to another. The systems across which the user Ids are mapped could be Ad Networks, DSPs, Ad Exchanges or Data Providers."
    "Here is how it works:
    * System A (say an Ad Network) gets access to a user’s browser (as part of an Ad Serving request). It creates a unique user ID for the user and calls a pixel URL supplied by system B. System A includes the ID of the user as a parameter in the pixel URL call.
    * System B’s pixel URL server reads the ID of the user assigned by System A and creates an ID of it’s own for the user. System B can now store the user mapping for this user in it’s database or in the user’s browser using another cookie.
    * A slight variation of this process involves System A storing the mapping also: B redirects to a pixel URL supplied by system A and includes the system B user ID."
  • O'Reilly radar on the announcement of the Spotrank API:

    "This is a new type of data. Never before has something like SpotRank been released. It will be used mobile apps and mobile ads (as Stump pointed out, it could cost more to show an ad in a busy part of a city). I can also see it being used city planners and corporate real estate agents everywhere. It will also give us great insights into human behavior."

    Fascinating. This will enable anyone to do almost realtime crowd monitoring…

  • Based on the number of location requests Skyhook gets for a specific location, they calculate "spotrank", the density of people around a certain spot.
    Since Skyhook only gets requests in case wifi hotspots are present, the meaning of it is limited to urban areas (where you can and sometimes have to rely on wifi hotspots – with buildings often obscuring the GPS signal).
  • Making your browser the central hub where your contacts are stored, managed, and given in lease to web service would be a giant step forward for data ownership and privacy (and you can always have automatic encrypted backups in case your hardware fails…
    "Contacts prototype consists of these pieces:
    * A browser-based Contacts database that stays in sync with your address books (so far, it supports GMail, Twitter and Mac OS Address book)
    * A generic importer system for Contacts from desktop or web-based address books (so you can implement missing ones)
    * An email autocompletion feature, which demonstrates how the browser can auto-complete email addresses on any website. The autocompletion is performed entirely in the browser, without sharing the your list of contacts with the website.
    * A Javascript API that websites can use to access the Contacts database, with explicit user permission and filtering"
  • Hunch, the recommendation engine for simple life questions by Caterina Fake has built this taster for this service that will try to predict your answers on simple life questions, based on your Twitter profile.
    Quite disappointing however, as it simple seems to determine a rough profile (male/female liberal, well educated – which covers 99% of all Twitter users) and then goes on about it, exploiting the same pattern in all kinds of cliche ways…