January 03, 2004

Plink

Via John Batelle and Anil Dash comes notice of Plink, a browser for the FOAF universe. I think it is a nice start, sporting a less technical interface than other similar engines. I wish however that he would support profiles without mbox properties so that all the meinbild.ch profiles could be added, too.

Posted by seefeld at January 3, 2004 20:52
Trackback
TrackBack URL for this entry:
http://www.bernhardseefeld.ch/mt-tb.cgi/80

Listed below are links to weblogs that continue the discussion on 'Plink'
friend of a friend
Excerpt: i have to admit, that i slowly start to get the whole thing behind FOAF, mostly because i started my journey through the foaf-space againg after reading B.'s post. and i can say that plink (here's my plink-page) is a...
Weblog: /blog
Tracked: January 6, 2004 04:33 PM
Comments

The point about PLINK is that it shows relationships between people, and currently the only unique identifier for a person is the mbox hash.

One of the problems with what you're suggesting is that the moment a person outside the meinbild community links to someone inside the community, they are no longer anonymous.

I'm happy to look at other ways to uniquely identify a person, however I really think that the way to do it is to let the individual decide whether they want their mbox hash published in their FOAF file, so they can make up their own minds just how anonymous they want to be.

Posted by: Dom Ramsey at January 6, 2004 02:56 PM

Actually, you can use any property that is inverse functional, i.e. whose value should be unique to a person. foaf:homepage as used by meinbild.ch is such a property.

In this way, you can even link meinbild users without exposing them in a verifiable way. You can still claim anything you want about them (within the FOAF architecture), including their realname and real email address, but nobody can verify it.

I suggest you support all inverse functional properties as unique identifier.

Posted by: Bernhard Seefeld at January 6, 2004 03:45 PM

Could do, but I would prefer to use something that is more likely to be unique.

foaf:homepage is very often not unique (a family or group of friends might share the same home page).

Anyway, simply 'supporting all inverse functional properties' is not an easy task and would involve re-writing a lot of the current code behind PLINK!

Also, this doesn't change the problem that the moment a person is linked to by someone else, they are no longer anonymous.

Posted by: Dom Ramsey at January 6, 2004 04:00 PM

There are differing degrees of anonymity and their exposition. E.g. if I claim on a webpage that Marylin Manson is in fact Paul Pfeiffer from the Wonder Years - and I do that in FOAF -, this doesn't really mean, that Mr. Mansons's anonymity is actually compromised in a serious way (not as much as my reputation, anyway). OTOH, if Mr. Manson publishes his email addresses' sha1sum and I can then publish a FOAF file with an unencrypted mbox property of paul.pfeiffer@wonderyears.tv, which would have that same sha1sum, Mr. Manson would be more definitely exposed.

Anyhow, if foaf:homepage is too often used wrong it would exactly be the popularity of tools like Plink that would expose this soon enough. Otherwise the specs should be updated, and Plink would still have many other inverse functional properties left to use.

And BTW about that claim:
http://www.snopes.com/music/artists/marilyn.htm

Posted by: Bernhard Seefeld at January 6, 2004 04:15 PM

pretty site you run Thanks

Posted by: Doriana at July 31, 2005 09:07 PM
Post a comment












Remember personal info?