The Realtime Ping Server

There’s going to be a moment in the near future where FriendFeed needs to deliver realtime search over IM. In a response to Robert Scoble yesterday, Paul Buchheit indicated track or something like it would be high on the list of things to commit to in the next round of improvements to the system. After a series of realtime services and API extensions late last year, FriendFeed has slowed its feature release pattern in favor of incremental improvements to its user interface.

The reasons for a realtime search mechanism can be debated, but what will be unavoidable is the implications of not moving forward for both FriendFeed and Twitter. The market for micromessaging may appear to be growing, what with major media outlets signing on and blogger firestorms about the value to individual brands dominating the holiday slowdown of so-called “real” news. But in fact, the Twitter bandwagon has propped up only the least significant part of the micromessaging wave.

The tip of the iceberg, Twitter’s marketing power or FriendFeed’s “conversation”, has not yet reached the kind of ubiquity micromessengers expect and not so secretly count on. The ROI on managing a Twitter Follow community produces reasonable economies of scale for only a very few. The Scobles must maintain their clouds regardless of the effort expended, while the up and comers are getting squeezed on FriendFeed by noisier folks willing to marginalize the threads with noise. People like Tim O’Reilly apparently avoid the comments and likes, and as a result the domain itself.

The disconnect and inequity of the current environment is substantially due to the lack of Track, which smooths out the signal to noise cycle by providing a realtime actionable opportunity that can be mined after the fact as “content.” Filtering built on top of Track will help tamp down spam and bot strategies, and it will also produce a set of favored filters that will both curate the flow and keep it open to new talent.

The small window of time when Track first emerged was not enough to evangelize its powers to the general audience, and obfuscation by vendors and entrenched interests kept the lid on for the second half of the year. But FriendFeed’s investments in realtime took the baton from Evan Prodromou and the open source Laconica project and made it clear that at least three business models are orbiting the realtime micromessaging opportunity.

In historical terms, we’re seeing a battle for control of the core realtime ping server. How long this post will take to reach RSS will illustrate how the major platform vendor in RSS is handling its responsibility as the major player. With its purchase of FeedBurner and recent rolloever of FeedBurner URLs to its own domain, Google has experienced some significant latency in syncing new posts to its dominant RSS feeds. This post, when published via WordPress, is made available to Feedburner within seconds. Yet most posts I’ve published in the last several weeks have averaged over two hours before they propagate to the RSS audience of Google Reader, FriendFeed, Twitter, and beyond.

TechCrunch engineers have implemented an auto ping service to try and fix the problem, but Google’s stated policy is to provide replication within 30 minutes. Whatever the case, and whether or not we’re correctly implementing a ping or not, the notion that blog posts are effectively removed from a realtime audience which is increasingly dominant is mindbogglingly stupid. Some even suggest there are competitive reasons for this lack of a strategy, but I can’t quite construct a convincing rationale for it to date. However, I will throw out the apparent fact that Google makes much more from Web pages than they do from RSS pages.

Inevitably, FriendFeed will roll out Track, and so will Twitter in short order, perhaps even sooner than FriendFeed’s smaller team can prioritize it. Until then, we will continue to model our Twitter cloud in FriendFeed constructs, make do with a lack of filtering tools to constrain the friend-of-a-friend overspill, and look to other players (Microsoft in particular) to compete directly with Feedburner at the RSS routing layer. There is no reason why RSS can’t be an effective protocol at the realtime layer, and FriendFeed’s growing arsenal of features is both a roadmap and a toolkit for the transition.

Note: I am publishing this post at 3:31PM Pacific time.

Update: 5:01PM No RSS.

Update: 5:52PM Still broken.

Update: Finally hit RSS at 6:17PM.