Adventures in Realtime

realtimeEver since Leo Laporte enabled a foldback loop on the video feed coming from his TwiT studios, the Gillmor Gang has hit a new sweet spot in the Adventures of Realtime. Prior to the foldback loop, we were still back in the Nightline days of staring blankly into the camera and pretending to see what Ted Koppel’s expression was. Harry Shearer of the Credibility Gap and more recently the Simpsons immortalized these Warholian head shots in a ground-breaking show we co-produced at Caroline’s in the Eighties.

When we started recording the Gang live on Twit.TV, the only way to get a sense of the continuity was to monitor the outbound BitGravity client and its delay of somewhere between 5 and 8 seconds. It’s a lot like spacewalking in slow motion; you see expressions float by that sometimes are the opposite of what you expected.

Of course, this is the same reason why Twitter’s Track feature was so valuable, and also why it was withdrawn and remains unavailable: the conversations you have in realtime have a feedback loop that produces a synthesis of collaborating minds unequaled in today’s virtual reality. Latency is the enemy of closure, the march through countervailing positions to the discovery of new people and ideas that add something to the accumulated context of the information stream.

We all know the problem, manifested earliest in school classrooms where the darting hands of “pick me” reveal the next question as coming from some earlier context far removed from where the subject has moved. Twitter realtime is one way, one to many in each direction. You can point the conversation at a subject domain via hashtags, or at several users with @replies (now @mentions). But hashtags form swarms that are vulnerable to noise, with management issues slowing down realtime to the point where it’s not worth it.

Track, unlike search, requires no management. Hits are pushed to you as alerts, or integrated into realtime flows if that functionality is available. Realtime critics accurately portray constant monitoring as unscalable, but we don’t sit waiting for the phone to ring in order to utilize its realtime technology. The real issue is that anyone can potentially interrupt you with a call, most commonly at dinner or at the moment when you finally negotiate a single show the whole family can watch and are running out of time before the youngest’s bedtime even if you fast forward through all the commercials.

How you control the interrupts is a function of filtering — determining the balance between discovery and ROI. Most solutions today involve topic filtering, but as we move into the age of gestural permissions, it’s people filtering that presents the highest return. Anonymity of content generation destroys the value of conversations, but harnessing the gestures of the anonymous produces much of the value proposition minus the privacy tax imposed by weak social contracts.

In the short time that FriendFeed realtime conversations have been enabled, the relative credibility of aggregated identity has fostered much greater signal to noise. Given that most users use their real names or a common pseudonym to sign onto multiple services, FriendFeed’s aggregation of these feeds encourages the identity owner to take more care overall to avoid polluting a blog feed with random hostile tweets or Likes of material that might offend users of other media.

Best of the Day threads are morphing from Scoble-centric to swarm-centric, as the concentric overlapping circles of realtime comments are expanding the discussions about realtime and in the process expanding the swarm to branch out into using realtime on emerging subject matter. That’s not to say that Scoble is losing traction, but rather that his advanced usage models are now becoming more understood and therefore merged into the mainstream.

Twitter backlash is not having much of an effect on FriendFeed dynamics, nor is the Facebook baiting working all that well either. While critics of Facebook’s opening of its stream to the beginnings of realtime complain that caching limitations prevent the kinds of public streams that Twitter provides, users are really waiting for what feels like one single sign-on across the clouds they are using. That’s what Plaxo’s McCrea and Smarr demonstrated at the Facebook event, and it comes amazingly close to what the enterprise calls a single throat to choke.

That;s why the argument that Facebook can’t provide comparable value to the Twitter public stream is only valid if you accept the notion that that’s what we want. In fact, what we want is the ability to be notified when something of interest has occurred, and then translate that notification into the ability to communicate with the resultant swarm around the affinity group that has similar interests. Does Twitter play in this? Certainly. As a trigger for the conversation, often.

Does Facebook? If it is the identity provider at the head of the chain, yes. Does Facebook slow down the discovery and acquisition of people and ideas? Currently, a little, as when comments replying to comments in the Facebook stream are not available outside (in other words, not released with an Everyone classification). In realtime clients such as Seesmic Desktop. you can see both Twitter and Facebook messages and comments, and respond in kind in what feels like realtime. The limitations: Twitter rate limiting of roundtripping and Track services, and no approved Everyone stream data beyond the 24 hour caching limit.

It appears that Facebook is more concerned with FriendFeed than they are with Twitter. The caching limits make it impossible to aggregate Facebook data on FriendFeed servers past 24 hours. But remember: it’s the user that wants the realtime services, and the user is in one of two modes – monitoring or catching up. Seesmic already provides the first mode. If Facebook stands at the head of the identity chain, it solves the second mode. If you want to go back more than 24 hours in a search, hand the request back off to Facebook and reintegrate it in realtime back into the FriendFeed results.

Here’s where Track wins again. If a message shows up in realtime (i.e. in the caching phase), it’s an opportunity to switch the conversation over to a stream that has no caching limitation. The Track triggers a process where the conversation flow is reversed, through the conversation hub (FriendFeed) and then back out to Twitter and Facebook for completeness. It’s not the Full Monty of realtime, but it’s the high value discovery that captures the lion’s share of the ROI. Inevitably, the message embargo will evaporate, most likely with the Everyone designation freeing those messages from the cache.

Like the old Abbott and Costello routine, slowly but surely, step by step, inch by inch, we are escaping the bounds of latency and having fun in realtime.