Re: [Social] Fwd: GNU/social legacy

[Melvin Carvalho]

On 10 December 2012 17:08, Ben Laurie <address@hidden> wrote:

On 7 December 2012 16:38, Melvin Carvalho <address@hidden> wrote:
>
>
> On 7 December 2012 16:30, Evan Prodromou <address@hidden> wrote:
>>
>> That's an interesting position, considering that there are 2-4 OOM more
>> people on the OStatus network than on every other federated social network
>> put together.
>>
>> What's your position?
>
>
> Facebook has it right, tent is almost there, OStatus has it wrong.  Many of
> the constituent technologies of OStatus have been described as 'fails' by
> the authors, tho this is I think is too strong.
>
> Build a federated network based on http identifiers, rather than, trying to
> breathe life into webmail style identifiers.  Tolerance in the identity
> space is what allows scaling.
>
> This was the advice given to GNU Social when Matt Lee met with Tim
> Berners-Lee.  It's as true today as it was then.

Really? Almost everyone else in the identity world disagrees! In
particular, the main reason it is thought OpenID hasn't caught on is
because it uses URLs instead of email addresses as identifiers.

This is largely untrue.  Fist of all, Facebook have a quite successful system based on HTTP identifiers and the open graph protocol. 

It's not rocket science HTTP GET to read, and HTTP POST to write.  HTTP was designed for for follow your nose, and to link, to sever data, and has identifiers that are federated from the start.

Follow your nose on email is much harder, and apparently involves having to invent a new URI scheme, still nobody knows how to do it and we're in the 4th year and counting.

You may speculate as to the failure of OpenID, but in truth there were many factors involved, one of which was trying to reinvent large parts of the web stack.

None of the above is even that important.  The important point is that exclusion is the key to making walled gardens and silos.  Tolerance is the key to federation and interop.