A Google Wave reality check

MOUNTAIN VIEW, Calif.--Lars Rasmussen sighed, half an hour into a demonstration of Google Wave, the company's audacious attempt to reinvent Internet communication: we'd found another bug.

Rasmussen had patiently worked around other minor bugs during the demo Tuesday at Google's headquarters, but when images dragged into a wave wouldn't load properly, he asked his brother Jens, seated at the conference room table, to get an engineer on the issue right away. It's about two months before Google opens up Wave access to a larger audience, and there is a ton of work to be done.

Google Wave was unveiled in May at the Google I/O Developer conference, and dazzled attendees with its goal: a combination of real-time communication with social-networking and search capabilities built into a familiar interface. Wave is more than just an in-box on steroids, however. It's also a communications platform that developers can use to build their own applications, something that many were excited about in the early hours of Wave's life on the public stage.

Behind the scenes, the reality is sobering for the Rasmussens and the 6,000 or so people actively using Wave. Job No. 1 for the brothers Rasmussen--who are managing the Google Wave project--is making sure Wave is stable enough to accommodate 100,000 new users that will start doing the Wave after September 30, when Google opens up the limited preview to a wider audience.

At the moment, around 25 percent of all Wave sessions end in a crash, Lars said. That's obviously not acceptable and, in an ironic twist, the highest priority bug on Google Wave at the moment involves search.

"I would imagine in six months this will be fast, slick, stable and usable," Lars said. "Right now, you have to be a super early adopter (to use Wave). By September 30, an early adopter."

Wave has been in the works for about two and a half years. The original prototype--constructed in nine months to pitch the concept to CEO Eric Schmidt and co-founders Sergey Brin and Larry Page--was actually discarded in favor of a system that provided better scale, Lars said.

Much of that time has been spent simply designing the workflow of Wave: how to add people to a wave, reply to a wave, add pictures, and create rules. Wave shares some basic infrastructure with Gmail, but is essentially a completely separate undertaking and has been a bit of an "organizational experiment" for Google in terms of giving an important project a great deal of autonomy, Lars said.

So why go public now, with so much yet to be accomplished? The brothers Rasmussen have heard the shouts of "vaporware," and actually chose the opposite launch strategy for the product that launched their Google careers: Google Maps wasn't unveiled until it was complete.

The difference with Wave is that Google believes developer feedback is crucial to its evolution as a product. "We wanted to get people thinking about how we're going to use it and what people are going to use it for," Lars said.

For now, however, Wave is carefully labeled as a "developer preview," a status that doesn't even rise to the level of one of Google's ubiquitous beta projects. While Google still has no formal process for determining what projects are previews as opposed to betas as opposed to full-blown products, the goal for Wave is reduce the number of crashes to less than 1 percent of all session starts, at which point the "beta" tag can be more confidently applied.

When introducing Wave in May, Google said it hoped to open the service up to the general public some time in 2009. That seems unlikely when viewing Wave in late July, but launching a product that has been hyped as much as Wave with anything even close to the number of bugs currently present would be a disaster.

Lars knows this. "Google can be a cushy place to work; we're not going to run out of payroll anytime soon. But we're putting a lot of pressure on ourselves."