Re: [Social-discuss] P2P or server approach?

[Melvin Carvalho]

2010/3/25 Laurent Eschenauer <address@hidden>

On Wed, Mar 24, 2010 at 10:17 PM, Blaine Cook <address@hidden> wrote:
> On 24 March 2010 17:06, Sylvan Heuser <address@hidden> wrote:
>> As I see it, we have two approaches between we must decide.
>>
>> The pure P2P approach:
>> *snip*
>>
>> The network of independent servers with small user groups approach:
>> *snip*
>
> I'd second the idea that there are hybrid approaches that are readily
> possible. I think even in the hybrid case, you need a shared
> addressing space. The reason you need a simple, shared addressing
> space is so that people can add each-other on contact lists. Once you
> have that, then two people who meet at a bar or on a bus can exchange
> contact information.

I second this. You can mix hub&spoke with P2P. In fact, this is what
we aim to do in onesocialweb and is straightforwad using XMPP.

My identity could either be:

Hub&spoke: address@hidden
In this case I delegate to the server the job of managing my profile, etc...

P2P: address@hidden/me
In this case, the work is delegated to a resource (could be a bot, my
laptop, a mobile phone..). The server only acts as a router. The good
thing with this last point is that you can use any existing XMPP
account tomorrow with OSW. And yes, you could even drop the /me part
and have XMPP Disco take care of telling the other end that your
social networking stuff is handled by a resource called /me. So it is
transparent to the user.

Not sure how this would translate in a Webfinger/WebID world...

+1 for shared addressing space (global identifiers)

One issue with email style identifiers is that you cant natively dereference them (ie with HTTP).  FingerPoint/WebFinger was invented as a work around for this limitation ( fingerpoint is the one I would personally use ).  I think Jabber/XMPP have their own technique to get more data from a JID. 

A big advantage of FOAF / WebID, is that dereferncing the identifier is straightforward, and aligned with the Web i.e. use HTTP.  This enables you to find out more information about a user with ease, and without inventing a new protocol, and you can even get back to the XMPP ID, email or even psyc ID, as well as the list of friends (which are again dereferencable) etc. 

For a centralised system for something with a large data center and high performance webfinger server, the email style identifier, with translation, seems reasonable.  However, I think that dereferencable (HTTP) global identifiers and a decentralized systems are a great (and probably necessary) match.