[foaf-dev] Re: Demo of social network portability while preserving
henry.story at bblfish.net
Tue Aug 28 16:07:26 BST 2007
I have built a little Java Applet to allow you to make and test Bloom
It also summarizes some of the discussions we have been having here,
showing how blooms, sha1sums, Access Control and other mechanisms
should be able to be used effectively to allow us to move to a much
more open world.
Thanks to Dan Brickley for helping me think through this. Btw. I am
not sure the Bloom algorithm I have is compatible with others. There
is talk about salts that I have not integrated yet.
On 24 Aug 2007, at 16:13, Dan Brickley wrote:
> On 24/08/07, Anthony Romano <dominatus at gmail.com> wrote:
>> Maybe I just don't understand this approach, but wouldn't it
>> require a
>> social network with 100 million users to query the filter against
>> *all* of them to see if there are any matches?
> I like the idea. I've been playing with bloom filters a bit today too,
> storing property/value pairs, where property is a uniquely identifying
> property such as homepage, openid, mbox, mbox_sha1sum,
> isPrimaryTopicOf, aimChatID etc.
> This could be useful without matching against each and every users.
> What I'll often do on eg. Flickr or Dopplr, Facebook etc is
> re-discover my contacts by finding them in the contact pages (or photo
> comments, walls etc) of contacts that I already have identified. So,
> given a few bootstrap contacts it should be possible to recover
> others. And the check could be recomputed eg. whenever one of your
> buddies adds a friend. I think we have to allow that there are
> multiple ways of identifying each buddy...
> (btw handy Ruby code I stumbled across:
> http://snippets.dzone.com/posts/show/4235 BloominSimple and BitField,
> though I'm having trouble rehydrating a bitfield from a stored
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 2429 bytes
Desc: not available
Url : http://lists.usefulinc.com/pipermail/foaf-dev/attachments/20070828/349c75b0/smime.bin
More information about the foaf-dev