[Freedombox-discuss] Friendica Red
----- Original Message -----
> From: Bob Mottram <bob at sluggish.dyndns.org>
> To: freedombox-discuss at lists.alioth.debian.org
> Sent: Wednesday, September 19, 2012 5:12 AM
> Subject: Re: [Freedombox-discuss] Friendica Red
> On 19.09.2012 04:38, Russell Edwards wrote:
>> You keep the logs of what happens on /your/ site. But every time you
>> comment in a context hosted on another node (wall-to-wall, community,
>> "private" messages etc.), you must consider, where will this be
> Obviously, I can control what happens on my own server but I have no control
> over what other people do with theirs.? That's always going to be
> potentially an issue with any system, and the only solution is to choose your
> friends carefully, and remove any who seem to have been inactive or who have
> moved their accounts.
>> We can't necessarily trust personal
>> home-hosted nodes, either, but at least they a) generally have greater
>> legal privacy protection and b) individually represent low-value
>> targets for snooping.
> I don't know about the Facebook connector, since I don't use it myself,
> but the main point is that use of home servers spoils the dragnet surveillance
> business model by rendering it uneconomical.? An adversary intent on the bulk
> collection of dossiers/timelines for commercial and/or political gain would have
> to compromise each box separately, whereas with a centralized Facebook-like
> system it's just a matter of someone at the company running an SQL query.
That is false.? It is not prohibitively expensive.
On a different note:
I'm quite pessimistic about the value of any Facebook alternative that doesn't come
standard with user-friendly data-mining utilities.? After all, those utilities are the whole
purpose of the software in the first place-- they are the very reason the UI and
network are designed to encourage users to leak certain types of data to the network
and inform on-- um, I mean "tag"-- their friends.? Any alternative that uses a similar
UI to Facebook without supplying the data-mining tools is naive at best.? After all,
how can users judge the efficacy of the system's privacy features if they don't even
know what kind of inferences can be made with that (meta)data?