Re: FYI: Blog post about Web of Identities

Enjoyed reading the post.  In answer to your questions:

Do you share my vision of the trend illustrated?

Yes, I agree that the "Web of Identities" is starting to get underway,
and will probably form the fabric of the next iteration of the
(social) web.  As you mention, there's lots of work being done in this
area.  Personally I find FOAF and the technologies built on FOAF to be
the most promising.  FOAF was largely behind the original proposal for
OpenID by Brad Fitzpatrick [1], and more recently behind FOAF+SSL [2],
but I think that's just the tip of the iceberg, and over the next
years we'll (hopefully) see much more innovation in this area,
particularly as W3C has recently grown its stack to include SPARQL and
RDFa which will be invaluable in dealing with linked (social) data.  A
simple example of browsing FOAF profiles can be seen here:

I'd like to see more possibilities of seamlessly jumping from server
to (eg facebook to myspace) without needing registration or
synchronisation.  However, given that FOAF is just linked data, you
can very easily move into the LOD world or start using rule based
inferences.  The neat thing about this framework, is that you can go
away and write your own vocabulary to deal with any pain points that
arise, and if other people like it, they will start using also, giving
you a kind of democracy of ideas, without necessarily having to go
through a formal specification process.

How could the named privacy issues be addressed?

We're in early days on the privacy front.  I think facebook (though
often maligned) have found a solution that seems to have gained
acceptability.  That is, by default only allow friends to see your
activitiy stream.  In the linked profile world, you need to build
privacy on top of identity.  Identity can be authenticated by any of
the standard patterns (SSL/OpenID/username,password) and will yield a
machine readable profile, as some refer to it, a "Web ID" [3].  The
advantage of SSL is that you can authenticate using your browser,
rather than redirecting to an IdP.  It doesnt really matter how you do
it, so long as you can authenticate and get a machine readable profile
at the end of the process.  Once you have a standardised machine
readable profile pulling out a friends list is not too difficult (for
example, SPARQL for foaf:knows) and use WebAccessControl [4] to
provide privacy options.  Similarly, portability of your friends list
becomes reasonably easy once you've built in the auth, as FOAF is by
definition a portable format.

There's a few solutions in this space, but I would suggest that all
are at early stages.  We've come a long way since MS Passport was the
dominant identity technology, and we probably still have a long way to
go, but the direction seems pretty positive, particularly from the
point of view of the end user.  The answers above are probably quite
"W3C-Centric", and I acknowledge that there are several more
technologies out there, and that interoperability will be important,
however your post was leaning in the direction of LOD, so I've tried
to present LOD in a wider perspective.

I'm pretty positive on the work being done in this area, starting with
LOD, but also moving to Good Relations which is quietly becoming the
"FOAF or e-commerce" and also SPARUL/WebDAV which combined with other
techonolgies will hopefully take us right to the frontiers of a
"ReadWrite" web!


On Sun, Jul 12, 2009 at 10:53 PM, Alex Korth<> wrote:
> Hello everybody,
> yesterday I posted a new article [1] on ReadWriteWeb and would love to
> discuss it with you. Do you share my vision of the trend illustrated? How
> could the named privacy issues be addressed? Thanks for your comment, Henry!
> Cheers,
> Alex
> [1]
> --
> Alexander Korth

Received on Monday, 13 July 2009 10:37:34 UTC