Ben Hammersley finally wakes up to the fact that his FOAF file was broken.
foaf stuff.
I had an email the other day from Marc Canter, asking where my FOAF file was. I've recently moved servers, and it had been left behind. It's back now but I'm still not happy with it.
My problem is that I'm not entirely sure who to include in it. Although there has been a lot of work on delineating relationships within FOAF - knows, isAcquaintenceOf, worksWith, isMarriedTo, hasComfortShagsWith, and so on - my problem is more that although I have a large and varied social network and want to share it with the world, I don't want to introduce people into the Foafsphere against their will.
I could, without this worry, easily automate my AddressBook.app to produce my FOAF file, and list everyone I know and their details. A bit of AppleScript would, I'm sure, do the business. But doing so would be a breach of each persons' trust in me when they gave me their contact details - no matter how sparse those details may be - and to remove such people from my FOAF file presents a level of user interaction I don't want.
There is a solution, however, and I'm putting it out here for comments: What if I was to produce a FOAF file where the list of <foaf:know>s contained only the mailbox hashes and no other details. I'd be perfectly happy to share my entire address book in this way, and if the person featured didn't publish some form of public online identity wherein his or her email address was available to all, then there would be no breach of trust on my part. If they did do so, any one of the FOAF aggregation sites would be able to fill in the missing details given the hash.
The value of the social network would remain the same whether you know the firstName or not - at least in my way of thinking about it. The key to me of these things is not that you might know someone I don't who may be useful to me, which seems a bit too much like a tacky singles bar than an application, but rather that we both might know the same person without realising it. By only publishing the hashes of the people I know, we might still be able to see that mutual relationship without either of us disclosing the private details of the man in the middle to the rest of the world.[Ben Hammersley's Dangerous Precendent]
Ben is one of those RDF guys who taught me how namespaces can work with RSS 2.0 and 1.0 - at the same time.
This issue of maintaining anonymity is obviously important, but how do you do that and have a universal digital ID system - that enables people to discover other people's info? Ben's suggesting that the hash system of foaf:mbox_shas1um would be a fine way to both hide people's real emails, but also enable them to be 'decoded (assuming they have a compatible system to do that.)
I told Ben - we'll have a solution to this challenge, just as soon as Eric Sigler get back from getting married.
