13 Mar Can the Real-Time Web Be Realized? Notes from SxSWi
One of the more interesting panels I was able to attend at this year’s South by Southwest Interactive Festival was called “Can the Real-Time Web Be Realized?” As status updates (via Twitter, via Facebook, via Foursquare) become increasingly important, the ability to figure out what’s happening increases. As long as we can organize that data.
But beyond that searching for products to buy, and being able to see if it’s in stock near you, up to the minute, would be equally important (particularly around Christmas, when this year’s version of “Tickle Me Elmo” is being fought over.)
The panelists all have a stake in the game. They include:
- Scott Raymond, Gowalla
- Brett Slatkin, Google
- Dare Obasanjo, Microsoft
- Marshall Kilpatrick, ReadWriteWeb (moderator)
- Jack Moffett, Collecta, a real-time search engine
Data Formats
To get this to work to maximum effect, you need to have uniform data streams. Moffett pointed out that some formats, like Atom, RSS and even PubSubHubbub (Push) are good starts, but they have limitations. We need to push toward uniform data sources that can deal with scale.
Interoperability
Brett Slatkin pointed out the issues today with cross platform compatibility. Remember when you had to pay extra to call someone on another cell phone network? Remember when you could only text message or instant message with people on your same network? We’re seeing that now, where Buzz doesn’t communicate (directly) with Twitter or Identica. Gowalla and Foursquare are different platforms that don’t talk. Ideally, you could use the platform that worked best for you and the data would work together.
But Which Specifications?
Obasanjo pointed out that developing 6-7 different specifications sounds like “a lot of work for a whole lot of people.” What we need instead is a series of protocols that are not proprietary, but allow the data to flow. The end user doesn’t care what platform they are in, they only care if it took a Twitter update took 2 hours to get into their system.
How Does Business Match Common Ground?
Raymond of Gowalla noted that “there’s a whole lot of work that needs to be done” on the balance between the individual incentive of the company (Gowalla has an interest in attracting users at the expense of Foursquare right now, for example), but the community has an interest in open platforms. When a company has only so much effort they can apply to growth, how much do they apply to their network and standards, or how much do they put into building their closed gardens? I’m paraphrasing him, but that was the main point.
What If You Want to Delete Something that’s Been Shared?
Moffitt wondered about the issue of deleting something if it’s already been shared all around the web. For example, if you upload a Flickr photo which then goes out to Google, Bing, and Collecta search engines, but you then decide to delete it, what happens? Currently the streams picked up by these other platforms don’t include a “delete this if you have a copy of it.” So once it’s out there, it’s out there. Slatkin of Google felt that dealing with deleting is technically simple, while dealing with much larger issues of access control need a lot of work. In other words, who gets to see your data, and can you segment that (easily) by each piece of content?
With 400m on Facebook, Are We Already There?
Kilpatrick asked if the public, at least 400m of them, have already voted by becoming Facebook users? But others, including Obasanjo said we’re doing “a horrible job” balancing Facebook’s goal (pushing out widely shared content for an active network) and the end user’s goal of throttling their data down to only those they want to see it. Moffitt noted that he doesn’t want to share his Amazon purchases online, but they use it for recommendations anyway. And he pointed out that Netflix shared a very small, “anonymized” dataset of users, but smart people realized that if they could decode the person who gave a review, they could follow that data to figure out who many of the users were. Once you know that, you know the movies they like to watch, ostensibly private data. That led the FTC to step in and pressure Netflix to cancel the second round of their effort to improve their recommendation engine with crowdsourcing.
Away from Data Silos, Focus on People
Obasanjo made an excellent point that anyone at SxSW can relate to: To figure out what your friends are doing here, you need to check Foursquare, and Gowalla and Twitter. The ideal situation would be for a person to be able to look in one place about their friend, and get a picture of all their doing. A people hub. Slatkin, however, wondered how you monetize that data if you can’t serve ads up next to it (he’s been taught well at Google). Obasanjo noted the conflict, but figured the person who does the best focusing on the person first will do well for their brands. Good point. One of the challenges, however, is getting the end user to think about who they want to share each piece of content with. Today, nobody wants to do that.
As you can see, we may be able to realize the real-time web, but there’s a lot of work yet to do.
