Jonathan Van Gieson has a list of Friendster Power Games that is just wonderful:
Top Friendster Power Games:
The Pre-Rejection
You just signed up for Friendster, and you notice that I’ve been using it for a month, and didn’t invite you. Perhaps we’re just not as close as you thought we were.The Delayed Approval
You can see by my profile that I was active yesterday. You sent me a “new friend request” three days ago. I haven’t approved it. Maybe it’s because I’m waiting to see if anyone worthwhile signs up to be your friend before I commit to having you on my friends list.The Unreciprocated Testimonial
You wrote me a very nice testimonial three weeks ago, yet your page still displays the pathetic notice: “No testimonials yet. You can add the first!” Gosh, it looks like you’re more interested in me than I am in you, doesn’t it?The Mexican Standoff
You’re one of John Smith’s friends. I’m one of John Smith’s friends. We know each other, we can clearly see each other in the “John Smith’s Friends” page, and yet neither of us has attempted to add the other as a friend. It’s a battle for status, and the first person to send the new friend request will forever be the loser.
This is really interesting, thanks.
These reasons and related are why FOAF (http://rdfweb.org/foaf/) doesn’t actually have a ‘foaf:friend’ relationship type. We started out with friend, knows and knowsWell, but that was just icky, and created situations of expected reciprocation like those above. So instead, we flattened things down to the more enigmatic and less controversial ‘knows’, and opted instead to decorate that with additional things like photo metadata (eg. who is in same photo as whom?), bibliography/co-authorship, workplaceHomepage and other less emotive ways of indicating various kinds of likely connection.
Some more notes on this at
http://rdfweb.org/pipermail/rdfweb-dev/2003-July/011358.html
Since FOAF, unlike Friendster, Ryze etc., is just a distributed set of documents, there is little that can be done to stop people saying hurtful things (ie. people control their own data), all we can really do is design the vocabulary/schema in a way that avoids some of the more obvious pitfalls…