Federated Social Web Summit Wrapup

Last Sunday July 18th, in Portland, we met with 45 people from 30 different projects and companies working on the federated social web. It went great; I wanted to share some of my personal experience from the event, as well as my thoughts about future collaboration with the participants.

The event was organized at arm's length by StatusNet people in Montreal and San Francisco; it's not easy to put together a big event in another city. But Jon Phillips (StatusNet community manager) and Florica Vlad (Jon's collaborator at Fabricatorz) did a great job, and the space at the Left Bank Project was open, spacious, and usable. Florica greeted attendees at the door with nametags and free #FSWS2010 T-shirts. We had coffee and bagels, and early-comers had time to get fueled up.

We opened at 9AM with self-introductions, then I did a "state of the federation" talk that seemed to be well-received.

We then had 3 hours of presentations by different groups using different strategies and technologies to solve the problem of federation. JSON, OStatus, RSSCloud, XMPP, RDF -- there were half-a-dozen strategies with different fine points around solutions. It was a grueling death march of tiny presentations -- hard for speakers and for listeners -- but by the end of the time everyone in the room was aware of what everyone else was doing. (Most of the presentation slides are at federatedsocialweb.net/wiki/Schedule .) Jon was pretty brutal -- he had a 5-minute timer with a train whistle that put an abrupt end to any presos that went over.

Free form

We broke for lunch (catered by the Left Bank -- really good sandwiches and salad) and conversation surged. It was awesome to see so many of the people who are doing work in the federation space talking to each other about their projects, issues they're facing.

As lunch broke, I huddled with Jon Phillips and Kaliya Hamlin. We quickly designed a half-day Open Space event with 6 spaces and 3 1-hour sessions. Joseph Smarr had started a Google Moderator process for discussion, so we kept that on the overhead projector. When people came back in, we collaboratively designed a schedule and set down to it.

There was a large contingent that focused on the very important issue of building "privacy" into federated social web protocols. It's a difficult subject, and most of the protocols we're using for syndication of social activities, like PubSubHubbub, don't easily map to our expectations for private communications.

Since that seemed well-covered by very smart people, I took the opportunity to discuss some other issues. First, I pulled together a session on branding for the federated social web. Do we need a name for the pattern of connecting social networks, or is it better to use names for protocols and suites, like OStatus? We didn't come out with a clear result on this; I think the best observation was that important federated technologies like "The Web" or "email" took a long time to settle down to easily-understood terms.

I also sat in a session on data portability. What formats should we use for backing up and restoring social network data? Although there's not a clear consensus, the group though using ActivityStreams here would make sense. Text could be encoded in events, and uploaded images and other files could be stored directly in the Atom elements. Additionally, the social graph could be stored as "Follow" events, and other data like favorites and tags could be stored similarly. When restoring data, a server can simply "replay" these events.

I got a few minutes to read Dave Recordon's blog post about developing social standards. I think it was a good focus -- starting with a technology and trying to match it to a problem is never a good way to solve an issue. However, I don't think it often makes sense to build new technologies up from scratch for each new problem set. Sometimes you have to use the tools at hand to get things done. I think what we've done for the OStatus suite is a nice balance of re-use of (sometimes barely) existing protocols.

What next

Finally, in our last session, I asked what we'd do next. Our original intention for the meeting was to make sure that people working the problem space knew each other's names, projects, and approaches. I had expected some conflict -- there were some long-time rivalries represented in the room. Instead, there'd been a remarkable effort towards cooperation.

We made a few important decisions here. First, we decided to open up our mailing list and wiki, originally created just for the one event, for further discussion. The intention here is to enable more "light" collaboration and keep people in touch. There were a few people who were invited who couldn't come, regrettably. I'm glad we'll have a way to keep working with them.

Second, we decided not to schedule a future meeting. My experience with this kind of meeting is that sometimes people drag it out unnecessarily. We agreed that if there was a need for this same group of people to get together in the same way, it would come up in our discussions. It may make more sense for people working in the same protocol clusters (XMPP, OStatus, etc.) to have more focused meetings.

Third, we discussed the very specific issue of standards. From the side of OStatus, we'd like to move the project along to a standards body, but we also know that the level of change right now is too important for us to fossilize the suite at its current level. I think there may be some options of giving OStatus a "foster parent" for now, and as the innovation continues, we share and document that process.

Finally, we decided to announce a challenge. Dave Recordon recommended a simple social federation scenario and following the examples of the WaSP Acid tests, we called it the Social Web Acid Test, level 0 -- SWAT0 for short. The challenge is to have as many discrete pairs of implementations pass the test by September 30, 2010. It's ambitious, but I think it will accelerate interoperability and make us take advantage of current momentum in federation.

At closing, I announced our decisions. We took a group picture and people started to head out to the Douglas Fir for drinks and dinner. I was pleasantly surprised when Ilya from Diaspora offered to demo cross-site subscription between Diaspora and StatusNet using OStatus. It was impressive! Not to be outdone, Tuomas of Buddycloud showed even better integration with Buddycloud. Hard to believe how quickly this is moving along.

My conclusions

In general, response from the attendees has been that the FSWS was a resounding success. I know that it's energized the implementers and given us impetus to work together. I hope that our immediate next steps will encourage others in the social networking space to consider federation as an option.

Finally: there's been some criticism of our decision to hold an invite-only meeting. I'm very sorry that we overlooked some people working in the space -- Dave Winer, for example. We made an exceptional effort to reach out to many social web developers, and we did make an announcement a few weeks ago specifically to make sure we didn't miss people. Everyone working in the space who contacted us got an invitation. Some people got missed; I'll take the blame for that.

I don't apologize, though, for choosing to have an invite-only meeting (that was my decision, by the way). That's what having a "summit" is about. From StatusNet's perspective, we simply needed to get together the people cutting code for federation; anyone else was going to get in the way.

Inviting participants also gave a sense of urgency; we had a few respondents who said, "We'll come to the next one." I told them, "There's not going to be a next one; this is your chance." Although we had some regrettable missed opportunities (BuddyPress, DreamWidth and Elgg are the main ones I see), our response rate was excellent. We had attendees from Asia, Europe, and North America -- I just don't think you get that kind of participation if people don't have a good idea of who else they're going to see.

If there's another general-purpose federated social web event, it probably won't be organized by StatusNet. I don't know if it'll be private or public; I can offer our support for whoever is willing to put work into making federation happen.

My heartfelt thanks goes out to everyone who came to the summit. I know that many of you made great personal and financial sacrifice to be there. I'm personally energized to continue my company's work on this issue, and I'm looking forward to working with you all in the future.

If you weren't there, and you want to be part of this: join the mailing list, and work on the wiki. If you've got code on the ground, help us pass SWAT0 by September 30, 2010.

• federation
• fsws
• fsws2010
• interop
• swat0

• Evan Prodromou's blog

Recent blog posts

Trackbacks for this post