Federation Functions and Tabulator (was Re: Federation protocols)

On 2 June 2013 17:05, Nick Jennings <nick@silverbucket.net> wrote:

>
>
>
> On Sun, Jun 2, 2013 at 3:06 PM, Michał 'rysiek' Woźniak <rysiek@fwioo.pl>wrote:
>
>> Dnia sobota, 1 czerwca 2013 o 21:35:23 Nick Jennings napisał(a):
>>
>> > On Fri, May 31, 2013 at 4:52 PM, Michał 'rysiek' Woźniak
>> <rysiek@fwioo.pl>wrote:
>> > > > > > This could be things like federated media sharing or quick ways
>> to
>> > >
>> > > add
>> > >
>> > > > > > a social layer to their mobile app or game.
>> > > > >
>> > > > > Great. Let's promote a single, well-defined protocol and this
>> will be
>> > > > > possible.
>> > > >
>> > > > Where do existing protocols like pump and buddycloud fail? What
>> would
>> > > > the single unified protocol do differently?
>> > >
>> > > Network effect:
>> > > http://en.wikipedia.org/wiki/Network_effect
>> > >
>> > > Usefulness of a social network grows exponentially with the number of
>> > > users.
>> > > If BuddyCloud, Pump.io, Diaspora and Friendica could all seamlessly
>> > > interop,
>> > > the "power" of such a network would be several orders of magnitude
>> larger
>> > > than
>> > > the simple sum of all parts.
>> >
>> > I think it's been touched on, but maybe we need to be more specific
>> here.
>> > What would these sites federating look like, specifically? If you could
>> > dream it up. Are we talking a "kitchen sink" of a social network? It's
>> > twitter and facebook and chat and google+ all combined into one
>> monolithic
>> > feed that can be interpreted by any of these sites?
>>
>> Correct me if I am wrong, but all of these have a similar datastructure of
>> entries and of user profile data. Content, author, date, optional title,
>> optional tags.
>>
>>
> As has been pointed out by others in this thread, the devil is in the
> details.
> There is a basic common-denominator, but the more you look into a social
> network the more little things make huge differences, and that's why I
> asked that we get more specific about what we mean by federation.
>
>
> Why are you putting "chat" into this? Chat is more real-time, instant
>> communication. It does not make sense here.
>>
>
> Why doesn't it make sense? It would make sense that two people who mutual
> follow each other (and allow it) share their presence information. In which
> case some form of instant messaging would make sense. Google & Facebook are
> doing it, so why not consider it in some form?
> .
>
> I can dream it up, it's quite easy. I would like to be able to have an
>> account
>> on any of these and be able to communicate (poke, friend, comment, etc)
>> with
>> any other account on any of them. Just like I can across Friendica and
>> Diaspora to a large extent.
>>
>> I don't see any concrete reason why this should not be possible.
>> Calling that a "monolithic feed", "kitchen sink" is disingenuous. It
>> would be
>> interoperable, yes -- that's the whole point. But "monolithic" is not the
>> right word, just as "kitchen sink" is not at all appropriate, as each user
>> would create their own environment by "friending"/"following" people and
>> putting them into "aspects".
>>
>
>
> You just described the ability to poke, friend, comment, etc. across any
> number of "federated" social networks. Which would effectively mean that
> each of those networks would have to implement each of those features. A
> twitter clone would need to implement poking and friending? commenting and
> tweeting are different concepts too. Messaging, posting, commenting,
> tweeting, activity updates (Jim played Texas Hold'em on Facebook, on your
> Twitter stream?)... If this is what federation is, then that's exactly what
> I see as one giant monolithic feed that social networks are all expected to
> support.
>
> Take XMPP. There are a huge number of things you can do with XMPP, various
> extensions that describe certain features and provide the XMPP stanzas to
> implement them. However, most XMPP chat clients don't support even half of
> the extensions described [1].
>
> In the end, when you talk about XMPP federation & interoperability, you're
> only really talking about *the lowest common denominator* which is sending
> a message from one place to another (OK, and presence & buddy lists). It's
> more common than not, that all the nifty extensions your XMPP client
> supports will not show up in someone elses chat (encryption, alert/buzzing,
> even 'user is typing' /'user has stopped typing' is an extension not every
> chat client supports).
>
> So, if we follow that same principle, in the end what we have is "
> http://twitter.com/user1 sends a messages to http://facebook.com/user2"
> ... but hell, on twitter you can't even send messages to someone who isn't
> following you.
>
> That's what I mean about what do we mean about federation :)
> -Nick
>

I think it may be worth pointing out Tim Berners-Lee's paper from 2008
which was one of the motivations for the creation of this group.

http://dig.csail.mit.edu/2008/Papers/MSNWS/

Although this is now very old by today's standards, he does capture many
important functions of federated social network.

It seems evident that friending between two networks (either uni or bi
directional) is a pre requisite.

Another aspect is that Tim mentions is a "personal wall" which you can
write to.  I would also add that it will be useful to send messages, as you
can with the facebook API.  This is perhaps fundamental.

News Feed and Photos are also listed as possible services that social
networks can offer, though of course each network will have their own set
of features.

Id actually say we've learnt one more thing in the last 5 years, and that
is that apps have become increasingly important to social networks.  The
app world seems to be balkanized so that is also perhaps a good areas for
standardization.

Tim also worked on a reference implementation in his own "Tabulator"
project:

http://tabulator.org/

But due to lack of resource the people at MIT have not been able to
modernize it instead working on ( http://myprofile-project.org/ ) and (
data.fm ) in my view two first class standards compliant projects.

I'm going to spend some time over the next year trying to moderize the
social pane in tabulator and hopefully being able to interact with any
others that wish to federate.  The nice thing about tabulator is that it's
all javascript so can largely run in the browser or use a shim to make the
web into a read / write social space.


>
>
>
>
>
> [1] http://xmpp.org/xmpp-protocols/xmpp-extensions/
>
>