Open Source and the Federated Social Web

The Structure of Communications Revolutions

Evan Prodromou

StatusNet Inc.

Agenda

1. Federation
2. Protocols
3. Projects

Communications revolutions

• Email
• Documents
• Personal publishing

Email: 1993

• large consumer systems: AOL, CompuServe, GEnie
• University networks on Internet
• Government systems (Minitel)
• Proprietary systems in corporations, government (VINES, Notes, Microsoft Mail)
• X.400 - ITU recommendation, government-mandated
• ad-hoc bridges
• BBSes (FidoNet) for customers or communities

Email: 1995

• Unified around Internet email
• Hierarchical addressing: user@domain
• BBS → ISP
• Telecom companies
• AOL opens up (1995)
• BBSes almost disappeared
• Open Source: Sendmail

Documents: 1992

• Proprietary, complex, internal in enterprise
• File sharing
• BBSes for customer download
• AOL, CompuServe, etc.
• Some FTP systems

Documents: 1997

• Web almost ubiquitous
• Hierarchical addressing: http://domain/document
• Intranet/Internet/Extranet
• HTTP + HTML
• Open Source: Apache

Personal publishing: 2001

• "Home page" on ISP
• Geocities, Tripod
• Frontier, Blogger
• RSS confusion
• Personal vs. Business

Personal publishing: 2005

• Hosted blogs ubiquitous
• URLs for identity
• Feed readers
• Podcasting (enclosures)
• RSS 1.0, 2.0, Atom: more than we need!
• Personal and business
• Open Source: Moveable Type, WordPress

What's my point?

• Revolutions happen quickly.
• Unconnected islands → federated networks
• Commercial adoption drives need for control
• Open Source implementation is key

Why Open Source?

• Low- or no-cost to install
• Bottom-up adoption (requires skill not capital)
• Rapid innovation as things scale

What is federation?

• Network of networks
• Open protocols (usually a lot)
• Uniform namespace
• Hierarchical addressing
• Anyone can play

Other federated networks

• Postal system (country + postal code + local addressing)
• Telephones (country + area/city code)
• SMS (ditto)
• TCP/IP (IP addresses)
• DNS

What drives federation on the Internet?

• TCP/IP
• DNS
• Scale
• Globalism

What drives federation?

• Control
• Distrust
• Greed
• Innovation

Metcalfe's Law

• Value is proportional to the square of the number of nodes.
• V = cN^2
• "The more, the merrier."
• Much, much merrier.

Metcalfe's Law and federation

• Makes big networks stronger, easier to resist
• Makes federation important for small networks
• Once network of networks > big networks, makes them join too

Social software: 2010

• Facebook: 500M users
• Twitter: 100M+ users
• Application-specific networks threatened: Flickr, Digg, YouTube
• National networks threatened: Orkut, Friendster, Bebo, Hi-5
• Niche networks threatened: LinkedIn, Ning
• Social gaming

One vision

• Some networks become defacto substrate for Internet.
  • Facebook: "Open" Social Graph
  • Twitter: Social messaging, "real-time"
• "Open" means "Use our API"
• Compare: Skype for voice
• Compare: Google for search

Another vision

• Commercial adoption of social messaging
• Need to connect
• Threatened networks fight to survive
• Open government requires 100% engagement
• Business needs of providers

Social software: 2012?

• Email-like identity evan@status.net or URL-like http://evan.status.net/
• Distributed real-time follow
• Combination of small and large networks
• Application-specific networks → general representations

Federation

• Adds players
• Builds value
• Stimulates innovation

Changing the game

• Winner take all
• Winner take some

The monolithic model

1. Get venture capital
2. Make social network software
3. Get everyone in the world to use it
4. WIN!

The federated model

1. Have people
2. Make or buy or install social network software
3. Connect to the rest of the world
4. WIN!

New players

• Big portals
• Universities
• Enterprise
• Mobile carriers
• Governments

Why should hackers care?

• Most important parts of our life: family, friends, romance
• Politics requires open discourse
• Making software that matters

Protocol suites

• "Email" = SMTP, RFC 1822 messages, addresses, MIME, ...
• "Web" = HTTP, HTML, CSS, JavaScript, favicon.ico, robots.txt ...
• "Blogging" = Web, RSS or Atom, Atom API, enclosures, pingback/trackback, ...
• "Social Web" = ???

Protocols: OpenID

• http://openid.net/
• Authentication (AuthC)
• URL for identity http://evan.prodromou.name/
• Devolving to a few identity providers (Google, Yahoo)
• Whitelist oriented

Protocols: OAuth

• http://oauth.net/
• Authorization
• Widely implemented
• Whitelist oriented (consumer keys)
• Not a lot of social parts
• Authentication (Sign in with Twitter)

Protocols: PubSubHubbub

• http://code.google.com/p/pubsubhubbub/
• Real-time publishing ("PuSH")
• Atom- or RSS-based
• Web hooks
• Great support: Google Buzz, Posterous, Tumblr, WordPress, LiveJournal, StatusNet, Cliqset, ...

Protocols: ActivityStreams

• http://activitystrea.ms/
• Represent social actions in Atom with XML namespace extension
• Subject, Verb, Object
• "Evan published a photo"
• Powerful when combined with PubSubHubbub

Protocols: Salmon

• http://www.salmon-protocol.org/
• Subscriber to publisher ("upstream")
• Comments and replies
• Activities

Protocols: WebFinger

• http://code.google.com/p/webfinger/
• Email-like identity for the Web user@domain
• XML document format matches identity to URLs (My photo service is X, my social messaging service is Y, my profile is Z, my OpenID is W...)
• LRDD uses URLs instead of Webfingers

Protocols: Portable Contacts

• http://portablecontacts.net/
• Define social relationships
• Static social graph
• User-controlled sharing of contact data
• Compare: XFN, FOAF

Protocol Suite: OStatus

• http://ostatus.org/
• Webfinger + LRDD = discovery
• PuSH + ActivityStreams = follow
• Salmon + ActivityStreams = reply
• ActivityStreams + PoCo = profile

Protocol Suite: XMPP

• http://xmpp.org/
• Originally developed for IM ("Jabber")
• Distributed system with email-like identifiers (evanprodromou@jabber.org)
• Social relationships == buddy list
• Profile == vcard
• Supports publish-subscribe
• Not widely implemented, not HTTP-based

What's missing?

• Privacy
• Client API
• Microapps

The Open Source enabler

• Who will be the Apache of the Open Social Web?
• Not sure yet
• Many contenders
• Trying to work together

Projects: Diaspora

• http://joindiaspora.com/
• 4 students in NYC, 1 summer, $200K!
• OStatus-like stack
• Ruby on Rails
• AGPLv3
• No working version... yet

Projects: DiSo

• http://www.diso-project.org/
• Based on WordPress
• Chris Messina, Steve Ivy
• XFN
• Leading ActivityStreams

Projects: Elgg

• http://elgg.org/
• General-purpose social network
• LAMP
• Commercial version http://elgg.com/
• Some federation (PuSH), more coming
• Lorea fork/branch/whatever
• GPLv2

Projects: GNU Social

• http://groups.fsf.org/wiki/Group:GNU_Social
• General-purpose social network
• LAMP
• AGPLv3
• No working code yet
• Will be based on StatusNet!

Projects: BuddyPress

• http://buddypress.org/
• General-purpose social network
• Automattic project, very nicely done
• LAMP
• GPLv2
• Very little federation

Projects: StatusNet

• http://status.net/
• Microblogging server
• LAMP
• AGPLv3
• OStatus for federation
• Twitter-like API
• Plugin architecture
• Identi.ca + 25K other sites on the Web

Projects: OneSocialWeb

• http://onesocialweb.org/
• Vodaphone project
• Use XMPP as core protocol
• Java plugin for OpenFire server

Projects: Appleseed

• http://opensource.appleseedproject.org/
• Indie
• LAMP
• JSON

Others

• AroundMe
• Appleseed
• Crabgrass
• NoseRub

Federated social web summit

• July 18 2010
• 45 people
• 30 projects

Outcomes

• http://federatedsocialweb.net/
• http://groups.google.com/group/federated-social-web
• SWAT0

SWAT0

• Social Web Acid Test, Level 0
• Simple use-case for federated social web
• Requires at least 2 implementations
• http://federatedsocialweb.net/wiki/SWAT0

What next?

• Projects working together
• Integration testing
• Real-life usage
• Innovation
• Growth

How to help

• Hack!
• Translate!
• Theme!
• Implement!
• Use!
• Spread!

Thank you!

• Follow me on the Internet!
• evan@status.net