Most of my time these days is spent crossing the blood-brain barrier between Twitter and the rest of the cloud. Twitter stands on one side, a coursing stream of social data emanating from an ad-hoc framework of asynchronous follows and vanity track filtering. On the other side, the legacy blogosphere, RSS items floated via Google Reader shared items and planted in the Twitter stream via TinyURLs.
Managing the transfer of data across the barrier are two applications. One (FriendFeed) is disguised as a social media aggregator, and the other (Twhirl) is disguised as a rich internet application extension of Twitter that allows multiple users, point-and-click UI enhancements of the vanilla Twitter feed, and, common to all third-party apps, a licensing limitation on polling the Twitter API.
On Tuesday, Twitter suffered its first substantial test since the 3-or-so day outage several weeks ago, the Indiana and North Carolina primaries where Barack Obama essentially sealed the nomination of the Democratic Party. As the polls closed and traffic spiked, the Twitter real time gateway through IM and SMS collapsed, leaving those of us who live on that transport high and dry. within minutes, we switched over to Twhirl, which slowly but more quickly came back online than the gateway through, in my case, Gmail’s Gchat.
For the next several hours, I ping-ponged back and forth between the two services, Gchat arrayed on the left of the screen in a vertical browser window, and Twhirl in its AIR container hovering above the right of the screen and notifications rolling up from the bottom of my MacBook AIR as they were received from API requests. The Gchat gateway went up and down, alternating between no service and old tweets paging in as the database of outstanding tweets was flushed, until sometime after 7PM Pacific they synchronized just about the time Obama gave his victory speech.
The outage illustrated one more time (as if it were not obvious already) the need for a scalable and reliable Twitter, or at least one third party service that also provides the gateway functionality: Real time conversations between discoverable endpoints not necessarily aware of each other until the swarming characteristics of an event, an idea, a personality, an affinity group, or any combination of these elements are enabled. Twhirl’s Loic Le Meur announced such features on the May 2nd edition of The Gillmor Gang.
Friendfeed will likely follow suit, but it raises more questions than it answers with its expanded comment infrastructure and extended harvesting of non-Twitter streams such as delicious and blogs. Robert Scoble has used Friendfeed and its Hide function as a refuge from too much noise on Twitter answering his 20k followers, but only when Twitter implements track filtering will mass following cease to be a feature driver.
Less solvable are the tactical feints by startups that masquerade as standards-based solutions to the so-called centralization problem. Gillmor Gangs on Thursday and Friday delved into the mysteries of decentralization, but I remain unconvinced that these strategies do little more than shift the controlling authority for the Twitter namespace to other potential landlords. First, it won’t happen as long as Twitter executives maintain open XMPP access to third parties, and provide timely and responsive solutions to track spam and predictive scalability for event thresholds during the next few months.
Second, a careful reading of tech politics suggests the takeover of Twitter is an unlikely occurrence given the weakness of second tier players like Yahoo and Sun and the strengths of Microsoft and Google. Yahoo looks like Hillary’s shadow campaign as it walks through the motions of building out a social media personalization strategy while Microsoft’s Mesh infrastructure obsoletes the portal logic it’s based on. Sun is courting social media superdelegates while IBM is piling up the popular vote with customers in the midmarket. In both cases, the numbers are brutal in their inevitability. Scott McNealy should engineer a merger of the two weaklings and give Jonathan Schwartz some tools to survive, matching Yahoo users with Sun/Amazon clusters.
But even that unlikely mating would be swift meat for Microsoft, who is all over why Twitter is fundamental to the next phase of the enterprise network. No matter who owns the pipes, the real struggle is to deliver the drugs across the blood brain barrier. Mesh abstracts out the hardware layer at a deeper level than Amazon or Solaris with its virtualization layer — down at the social layer where the users live and control the domain. It’s the users, stupid, as Carville famously put it. Once switching costs are controllable, the user can band together in affinity groups and mandate the price vendors will need to pay to be listened to.
At its simplest (its true power) Twitter is a phone switch for routing information flow. Those who control the flow control the price for the information. In a virtualized platform, the hardware is the razor and the software switch is the blades. The software switch is an affinity-based construct that manages the signal-to-noise ratio of the information flow based on the contouring signals (gestures) of the members of the group. In the language of Twitter, it’s who you follow times what you track divided by how you filter.
The trick is squeezing the firehose down into multiplexed channels across the blood brain barrier and then expanding them as they flood the brain and its synaptic map. The architecture of swarms has unique characteristics that we are seeing modeled in the contortions of Friendfeed, Facebook Connect, Ustream chatrooms, Google Reader Notes, Disqus, and the rest of what Marc Canter calls the open mesh. It goes beyond bootstrapping, harnessing the brain’s ability to add the gut instinct of survivability to the equation of what choices can be made about information triage.
Simply put, you have to have the ability to broadcast an acuity for successful guesses. We’re at the doorway of gesture farming, where individual gesturers go beyond implicit behavior harvesting and aggregation and overtly share not just what they like but what they ignore. We’re seeing this in the political realm, where people are tuning out repetitive and shrill networks built on track spamming (Reverend Wright, Day One, electability) and tuning in to credible authentic sources regardless of media affiliation. They’re going direct via TinyUrl and their social graph (follow/track/filter) ontology.
Those who laugh at Twitter and trivialize it are insulting the very users they want to engage with. In elections, that is a fatal mistake. In technology acquisition and adoption, it is similarly Darwinian. Ballmer’s buh-bye is still being discounted as posturing, but in a real-time conversation, once you’ve met the mettle of the (wo)man, you know what you need to know. I think Ballmer and Gates and Ozzie had already made the calculation before they made the offer, namely that they were looking for a partnership with Yahoo’s users and developers, not with its executives. That is not to say they were not valuable, just that they would have to prove their value in the conversation. They didn’t. The rest is still in play.
Decentralizing Twitter is unnecessary, if not impractical. Dave Winer was right the first time, when he intuitively grasped the power of Twitter was not in what it was designed to be but in what it could be used for. By building on top of it, Winer signaled that instinct that he marshaled into RSS, the gesture of respect, the idea that in Steve Stills’ words, “Somethings happening here, What it is ain’t exactly clear…” Twitter ain’t broke, and we don’t need to fix it.