Twitter Jitters

I for one am not buying the Twitter jitters I see emerging in the technorati.

Duncan Riley is the newest voice suggesting the bloom is off the rose, quoting Dave Winer, one of the earliest of the twitterati to suggest the end is nigh:

[from Twitter Has Jumped The Shark]

Perhaps the saddest thing with Twitter jumping the shark in [is?] the loss of the Twitter community, as Dave Winer writes “People wondered what would replace it (Twitter). It’s becoming clear the answer to that is the worst possible one — nothing. The energy of Twitter is evaporating. Which is terrible.” I’m still hopeful that enough people will decide to migrate to one service that it becomes a Twitter replacement, however to what? FriendFeed could build a Twitter killer but they haven’t, Jaiku is a great platform sitting on Google’s servers…but it’s closed for signups, Pownce is…well…Pownce, and Plurk has no API.

The obvious business opportunity: somebody comes out with a Twitter competitor with API and scalability. If you’ve got one ping me so I can take a look and give it a plug.

Likewise, Rafe Needleman recently suggested closing Twitter down until the service stabilizes, like a restaurant undergoing renovation (see Rafe Needleman Wants Twitter To Close For Renovations).

I for one think that these guys are looking at this in a very short-sighted way.

There is a growing ecosystem of micro businesses (or micro applications, perhaps) that are building on top of the Twitter APIs, and exploring new ways of delivery novel user experiences based on Twitter. (I am working with Greg Narain on one of those, the 'presencetation' micro application called Front Channel.)

There is growing adoption of Twitter outside the technorati, and looking at the Alexa numbers show a bumpy, but relatively consistent -- and very fast -- rise in pageviews and reach.

Twitter v The Gnats, originally uploaded by Stowe Boyd.

One concern: It's a bit worrisome that the tech scene can take a tone of a witch hunt, and begin to call for a company's product to be reverse engineered as a public utility, like Michael Arrington and others seem to be trying to do to Twitter. The argument in a nutshell is something like this:

1. Twitter plays an increasingly interesting role in people's lives, and as we have come to rely on it -- in lieu of other communication systems, like IM, email, and so on -- we need to have it operate dependably.
2. If we wait too long, Twitter will become the monopolistic center of its own service, like AIM, or Microsoft Windows.
3. However, like Jabber or Linux, we (or someone) could start by defining a protocol for a distributed, open service based on reverse engineering the elements that make twitter interesting, like the 140 character limit on messages, the Twitter APIs, and so on.
4. Then, various parties in the ecosystem, like the various Twitter clients and services, could generalize away from reliance on Twitter, Inc., and instead operate around the twitter protocol.
5. Other companies could create a network of communicating twitter servers that could cross authenticate, and then users could sign up to have twitter accounts wherever, not just with Twitter, Inc.

One snag: Twitter, Inc would have to play along by cross-authenticating and cross-communicating with these other twitter servers.

What does this all remind me of? Well, obviously, the instant messaging world. The Jabber protocol (or jabber
to stick with the upper and lower case distinction) is a distributed open model based on the XMPP standard. Jabber, Inc. (a former client, who actually rejected a recommendation I made a few years back about my Nerdvana idea, something very like Friendfeed, by the way) is a for profit company that has sought to market itself and the XMPP jabber standard against AIM, Yahoo, MSN, and other proprietary IM solutions for years. And what has happened? Individuals have continued in general to use the free services offered by media giants -- now including Google, Skype, and Apple -- and have avoided the hacker mystique of jabber. Corporations have opted for various proprietary solutions -- like gussied up versions of MSN Messenger and IBM Sametime -- or have allowed their employees to use consumer solutions, like Gtalk. No real convergence has emerged, although AOL and Microsoft don't spend as much time these days fighting against the Trillians and Meebos of the world.

So, in essence, the technoids are suggesting that someone should develop a jabberesque standard, and that we could all opt to use solutions that talk that standard, below the hood, which would provide some degree of freedom and scalability, in principle.

The snag is that users don't care about architectural niceties: they want functionality, low cost, and reliability. That's why millions of people everyday use Skype and AIM, and very few use jabber, unless it is a jabber-based product backed by a media giant, like Google.

But performance is the issue, the tech pundits might reply. Yes, but that is an architectural defect of the current Twitter implementation, and one that can obviously be fixed. After all, if AOL can support billions of messages and presence updates everyday, why can't Twitter? It's just a complex problem, but one that is well-understood. (By the way, buying Twitter and integrating it into AIM should be the number one priority of AOL's messaging group -- another former client of mine, by the way.)

So, individuals are more than willing to use a proprietary -- even monopolistic product -- to get the benefits that really matter to them.

My bet is that the following is the most likely scenario for Twitter, and it is one where the squawking of the pundits won't matter much:

1. Twitter -- newly recapitalized -- rearchitects the core of Twitter to actaully scale up to millions of users, and beyond. By late 2008, uptime reaches industry norms.
2. During 2008, hundreds of other Twitter API-based applications are developed. Twitter moves aggressively to create a strong and open developer community.
3. In late 2008, Twitter announces an open source-ish developer framework, to help members of the development community more quickly create Twitter-supported applications.
4. The combination of the growing stability and growing userbase leads an industry giant -- Google, Yahoo, MIcrosoft, Skype, or AOL -- to acquire Twitter and integrate it as the core of a redesigned and reconceptualized instant messaging and micropublishing hybrid.

Meanwhile, I need to go see if anyone pinged me on Twitter. Later!

[PS The guys at Obvious have taken precautions to avoid service outages tomorrow because of the Apple conference.]

[PPS Euan Semple points out

Facebook, Twitter and all the other tools that people like to disparage as being silly are probably the latest attempt to express collective joy in the face of "grown up" resistance!

]

Comments

[Mary Wallace]

I personally didn't like Twitter's 'good luck with that' responses to a victim asking for help against cyber-bullying recently, it smacked of "we want your good experiences, not your bad". But that's my line in the sand for whether or not someone has integrity. Not whether or not someone deserves to still stay in business. The Twitter team has been working non-stop to undo a faulty business beginning that wouldn't scale to the unexpected online success it created. They're not going out of business, just because they falter. Instead, they are staying in business by doing the gutwrenching work needed to reimagine their platform. When twitter is down, get a life. Then check back into twitter and bring your new richness to your twittermates. Its just a short period of time, a few months perhaps, until Twitter is much more stable. Calling for its death, or wholesale abandonment of it, is egotistical. Twitter has new people signing up all the time, not simply because Dave Winer or Arrington said its 'cool', but because it represents a new form of IMing that's more global.

That being said, I'm having much more fun over at Plurk!

[Dennis Howlett]

I agree there is something of a witch hunt quality to the recent conversations which never bodes well but then the Twitter people have been incredibly dopey when it comes to understanding the problems both from a communications perpsective and from an architectural one.

The best clue I got to all of this was when I asked Jack Dorsey whether a CEP approach might work and it was obvious he had no idea what I was talking about. Whether that approach *would* work is another topic but to be wholly unaware seemed odd to say the least. Similarly, it is now clear the core architecture is wrong. Therefore any hopes of resolving the issues anytime soon seem remote.

They've tried a lot of solutions (some of which I am privy to by other means) which demonstrate a lack of maturity in understanding scalable architectures. I'm not sure they're communicating anything that gives confidence.

It would only require another team to come up with a clone with an API for the Dave Winer's and Robert Scoble's to defect, with the attendant numbers held in their respective gravity fields. My sense is Twitter is running out of time. I hope it isn't too late.

[Neil Vineberg]

A year ago when I was promoting Jaiku the reliability of Twitter's platform was a question mark that enabled us to achieve some parity for Jaiku. Reliability is still an issue. Since then we landed an unmanned science platform on Mars. I'm not a tech expert, but one would think that in the course of 13 months that issue would have been laid to rest using the first infusion of several million dollars.