Innovation on the
Social Web

Sandro Hawke, W3C
Slides at http://hawke.org/2018/webconf

sandro@w3.org
@sandro@w3c.social

27 Apr 2018
The Web Conference

Credits

Overview

  1. "Social Web"?
  2. W3C gets to work
  3. The helpful elephant in the room
  4. Innovation land

Conclusion

  • The Social Web is usable today
  • It's a great place to build new things

Part 1: Social Web?

2008

Social Software

1981

Social = Multi-User

  • Value comes from other users
  • Not too hard as client-server / DBMS
  • Extremely common today

At odds with Web's decentralization

Social features on www.site1.example

disconnected from

social features on www.site2.example

So the Web ends up centralized

Centralized Social:

airbnb craigslist doodle ebay eventbrite facebook flickr imdb instagram kickstarter linkin lyft match meetup pinterest quora reddit snapchat tinder tumblr twitter uber webex whatsapp yelp youtube

Even: wikipedia, github, easychair !

Decentralized Social

Rare!

  • SMTP email
  • IRC
  • XMPP

plus many that never caught on

Why?

  1. requires consensus
    • from everyone building the software
    • for nearly every change
  2. only modest financial rewards
  3. bootstrapping is very hard
    • value is other users
    • each new user sees a ghost town

Still want it!

  • Freedom to innovate
  • Autonomy for users
  • Security options
  • Privacy options
  • ...

Maybe...?

  • Blogging is decentralized + social

Add to blogging?

(one perspective)

  • data-like blog posts to allow nicer UI
  • immediate notification (pub/sub)
  • like & comment via your own site
  • access control (friends-only posts)
  • other applications beyond newsfeed

Part 2: W3C

Federated Social Web Summit (2010)

Incubator Group

Community Group

Working Group (July 2014 - Feb 2018)

Perspectives in WG

  • IndieWebCamp
    • blogs with metadata in microformats
  • OpenSocial
    • syndicate XML ActivityStreams
  • Solid
    • RDF Linked Data on Personal data stores

Oct 2014

Mar 2015

May 2015

Dec 2015

Mar 2016

Jun 2016

Sep 2016

Nov 2016

W3C Recommendations

  • IndieWebCamp style
    • WebSub - pub/sub for the web
    • Webmention - standardized back-links
    • Micropub - API for posting to your site
  • OpenSocial style
    • ActivityStreams - JSON-LD, Actors + vocab for "social"
    • ActivityPub - vocab/API for managing streams (s2s, c2s)
  • Solid style
    • Linked Data Notifications (LDN)

(some other parts didn't get to REC)

Multiple Standards?

Human Scale

More Machine-Friendly

Small Problem

Or opportunity

Right before we finalized ActivityPub...

Part 3: Mastodon

Adoption Curve 1yr Ago

From mnm.social

Using older protocol

  • We Decided to delay the WG 6 months
  • Added needed features to ActivityPub
  • Mastodon adopted ActivityPub (for server-to-server)
  • Win!

Other adoption

Now or soon:

GNU Social, pump.io, Pleroma, Peer Tube, FunkWhale, Nextcloud, Friendica, Socialhome, ...

Ongoing adoption

How?

  • 2500 Instances (Servers)
    • Whole communities signed up at once, on an instance
    • Local moderation rules
    • Home for folks not comfortable (or allowed) on Twitter
    • Independently run (or commercially)
    • Try https://w3c.social
  • Top-notch open-source development lead (Eugen Rochko)
    • Community supported via Patreon and Liberapay
  • JSON-LD with standard @context, viable RDF middle ground

Gap to Mainstream

Some items one might want to provide:

  • Full Search
  • Recommended Users
  • Algorithmic Feed
  • Advertising
  • Groups/Lists
  • Some kinds of website embedding
  • Trending topics
  • Reliable Metrics

Or not?

Part 4: Innovation Welcome

Privacy

TOR onion routing reportedly works

an interesting start

Make existing websites more social

Activity feeds out (news to folks who care)

Activity feeds in (aggregating news)

Embedding of social actions (like "Like" button without a commericial 3rd party)

Games with social features

Like when Facebook was mostly games

Multi-platform integration

Folks are so spread out today

brid.gy is a start

more connections needed

Land Rush?

It's a good time for first-mover advantage

Like early days on a platform, maybe

For product or personality

Monetization

Explore new forms of monetization

  • Pay to follow is doable
  • Disclosed native advertising?

UI innovation

Modify one of the open-source UIs

Or start from scractch

(without needing to run servers or hold user data)

Social Readers

RSS Reader + Like/Comment + Access Control

See more on indieweb readers

and IndieAuth for WordPress

Handling bad behavior

Welcome to the modern Web

But community instances may be a more scalable model

(study this!)

Tunable feeds

Instead of a single opaque algorithmic feed

let users make multiple configurable feeds

and switch between them.

Machine learning under user control

Surfacing more credible content

Explore ways to signal increased/decreased credibility in your content and social graph

See https://credweb.org

Extensibility

ActivityStreams/ActivityPub is JSON-LD (extensible as RDF)

but Mastodon's db has a fixed schema

and used a microblogging client-server protocol (not AP)

Nudge the community?

Beyond Microblogging

group chat, photo sharing, ride sharing, meeting scheduling, commerce, job search, ...

Identity Migration

Now you control your on-line identity, but ...

what if you want to move it?

or back it up?

(many interesting issues)

Conclusion

We have a thriving Social Web MVP. Let's build up from here.

Consider joining https://www.w3.org/community/SocialCG/

Join the fediverse, eg at https://w3c.social

Read https://www.w3.org/TR/social-web-protocols/

Slides at http://hawke.org/2018/webconf