Report from ReadWriteWeb’s Real-Time Web Summit | FirstRain

Report from ReadWriteWeb’s Real-Time Web Summit

Guest post: Marty Betz, VP Technology

Today, YY and I attended ReadWriteWeb’s Real-Time Web Summit. It was held at the Computer History Museum in Mountain View, hosted by Marshall Kirkpatrick of

Like many conferences recently, they’d adopted a structure based on an open, spontaneous agenda which gets filled in by the participants during the first hour. It turns the meeting into a focused networking event that fosters exchange of technical questions and information between the attendees, and minimizes the time spent having one person speak at a large audience. This worked surprisingly well. It seems this is a natural evolution of conferences. Now that much of the information that was once delivered from a lectern can now be found online, meeting new people and discussing specific problems is where unique value is found. I’m always impressed by how many attendees are willing to put themselves out there and deliver a semi-impromptu presentation, or run a discussion.

The talks ranged widely, from protocols underlying real-time communication, to human factors associated with information processing, to data extraction and natural language filtering techniques. Of course, there was also a good share of “how do we monetize this” chitchat.

My initial observations are that, even though the crowd was self-selected for real-time web interest, there was healthy cynicism about it’s value. One talk was even titled “Why do you hate real-time…” On the other hand, there was a palpable sense that something important will emerge out of the web’s new, faster and denser information flows; we just can’t predict what it is yet.

The host, Marshall, started by proposing a spectrum on which ideas were organized based on whether they focused on person-person communication, person-machine communication, machine-to-person, etc.

Every hour had several different talks to choose from. I first dropped in on a talk about the service PubSubHubbub from Google, and then another about linguistic techniques for filtering twitter feeds. My next hour included discussion of structured information and tools for collecting and deriving structured entity relationships.

YY ran a great session on how to make use of the way the web changes, over the course of the day — and longer time periods. It raised questions about how can you distinguish the significant change patterns from the incidental ones? It evolved into discussions of when and how long “old” information remains valuable, and it gives valuable context to recent data. Others in the group talked about filtering highly noisy sources, the practical requirements and limits of human attention, questions of push vs. pull information delivery, and the frequency with which people can really ingest and make use of content streams.

I finished the conference attending two light afternoon talks. The first was on the importance of “emotion” in the value we attach to real-time content. There was animated conversation about how data you get from another specific individual can come with the baggage and the benefits of your immediate relationship and that person’s mood. Finally, I enjoyed a round table discussion of how soon “augmented reality” will become a reality.

Overall, I thought the conference was a great experience. As usual, the web continues to change year after year, and new problems arise, raising new opportunities. I am confident that the increase in real time data will, like other changes in the past, inspire FirstRain to find unique loads of derived value in the web.