Re: [apps-discuss] The acct: scheme question

["Paul E. Jones" <[e-mail]>]

Melvin,

 

Just to the below point, it is the common discovery pattern that acct
addresses.  A service provider might identify users  in any number of ways,
but people need a well-defined and predictable way of using WebFinger.
Plus, they need an ID that has been proven to be understood by the average
user.  If I see a twitter ID like "paulej", I should be able to safely
assume that I can discover information via WF using acct:[e-mail].
Users will likely not enter the "acct:" part, but that's fine.  They often
do not enter "http://", either.

 

The reason I say that identifying Twitter, Facebook, Flickr, or other users
is in scope is that, regardless if those services publish via WebFinger or
not, the approach taken to discovery given a username and a domain name
should be consistent across the web.

 

Things like Open Graph are fine, but we need something consistent across the
web for simple, consistent discovery of information related to a URI.

 

Paul

 

Twitter, Facebook, Flikr and Google Plus all use HTTP URIs to describe their
user accounts.  This is best practice on the web, and recommended in Linked
Data Principles [1] which is the dominant discovery mechanism on the Web
today.  The gap in linked data, which I hope Webfinger can address, is a
common discovery pattern (using well-known locations) for email addresses.
If Twitter or Facebook and others have not asked for yet another way to
identify users, I am unsure why this use case is in scope.  Do note that
Facebook already have a first class discovery system in the Open Graph
Protocol.