Original URL: http://www.theregister.co.uk/2009/05/30/google_waviness/

Google Wave - interwebs idealism in real-time

It's not a product. It's a user happiness pill

By Cade Metz

Posted in Applications, 30th May 2009 05:40 GMT

Google I/O When Lars Rasmussen first floated the idea, Google co-founder Sergey Brin wasn't impressed. "He came to me and he said 'This may sound kinda crazy, but we're going to reinvent communication and we just need a bunch of engineers to go of to Australia for a while and we'll get back to you after a couple of years,'" Brin remembers. "It was not a very compelling proposal."

But Brin greenlighted the project anyway. After Google acquired their Where 2 Tech startup in 2004, Lars Rasmussen and his brother Jens had spearheaded the Google Maps project, and in an extreme case of Google's much-lauded commitment to creative freedom, Brin gave the pair just what they asked for.

"Lars and Jens had previously redefined what mapping was like - they already had a success under their belt - and communications was one of those trigger topics," Brin told reporters yesterday afternoon at Google's I/O developer conference in downtown San Francisco. "We decided to give them the benefit of the doubt. It was also an interesting experiment. It was one of the most autonomous development groups we've had at Google."

The way Brin tells it, the decision allowed the Chocolate Factory to "innovate how we run things." But by all accounts, this amounted to letting the Rasmussens do whatever they wanted. The result - after two years of development - is Google Wave, the new-age communication and collaboration tool the company unveiled on Thursday to a standing ovation from hundreds of gathered developers.

Still under construction - and available only to the developers who attended Google I/O - Google Wave is a browser-based app that crossbreeds email with IM and document sharing, allowing you to respond to email-like messages with IM-like chatter - and vice versa. And inside and among these cascading discussion threads - aka "waves" - you can drag-and-drop photos, videos, maps, documents, and more.

Wave promotes a kind of "real-time interaction" that's closer to real-time than most. When you send an IM, each character appears on your buddy's screen as you're typing. And when the two of you share a document across the wire, you can edit adjacent text at the same time, eyeballing each other's changes as they happen.

Google Wave

A Google developer preview

Some have painted Wave as a slap in the face of Microsoft's Ray Ozzie-inspired collaboration app, SharePoint Workspace. And others see it as some sort of Twitter antidote. But during a press briefing Thursday morning, Google insisted that Wave was built without competing products in mind - much less a concrete business model.

As Vic Gundotra, Google's VP of engineering put it, "One of the luxuries about working at Google is that we get to focus on building the technology and making users happy, and once we've achieved a certain amount of success in terms of user happiness, only then do we start working about how to make money from it."

With engineering projects such as Wave, Gundotra later added, "We don't think about what competitors are doing... We believe that you build for the user and the rest will follow. Part of the excitement is rethinking the problem and coming up with a fresh approach."

When one reporter questioned whether he was telling the whole truth, Gundotra quickly repeated himself. And judging from Google's track record, we're inclined to believe him. At least for the moment, the company's top-secret search money machine is pulling in more than enough dough to fund such idealism.

Clearly, Google wants Wave to be an open means of communication along the lines of email, and not just another addition to its very own online Apps suite. Gundotra and company are prepping a Wave comms protocol so that any other communication tool can talk to its app. They're opening up "the lion's share" of the app's client and server code, so that outside developers can build their own Wave apps. And they're publishing various Wave APIs so that world+dog can slap the app's (near) real-time conversations onto blogs, wikis, and other webpages.

"We think of email as a successful open protocol," Lars Rasmussen explained. "But we think that with the advances in computers and all the experiences that have been made with different types of communication, [the world] can do better. Google Wave is our suggestion."

How very Googly. Two guys asked if they could replace email. Sergey like them, so he said yes. They went off and built something. And now, the company has tossed it against the interwebs to see if it sticks. As the company rolled out the developer preview at Google I/O, you almost got the impression it was a last-minute decision.

Gundotra pitched Google Wave as the spearhead of the company's efforts to revolutionize browser apps using HTML 5. Coded in Google Web Toolkit, the app leans heavily on the still-gestating standard. "This is an unbelievable demonstration of what is possible in the browser," Gundotra said just before Lars Rasmussen's demo. "You will forget you are looking at a browser. I want you to repeat after me 'I am looking at an HTML 5 app. I am looking at what is possible in the browser itself.'"

Apparently, Sergey Brin needed much the same talking to. According to Adam Schuck, one of the Wave's early developers, when the team first gave Brin a demo, the Google co-founder said "It's cool you're coding in Flash."

Let's send an emailim

The project began in 2007. Google Maps had launched, and the Rasmussens were looking for what the world insists on calling a new challenge. It was Jens who suggested that they reinvent online communication, and Lars said "Yes" right away - or almost right away.

"The thing that really caught me was that he said instant messaging-type conversations and email-type conversations shouldn’t really be two separate things," Lars remembers. "I remember being sold immediately – though Jens says it took several hours."

But the brothers agree that the project was sparked by this fairly simple idea. They weren't trying to mimic or compete with existing systems, they say. They were trying to build something new.

Of course, they then looked to existing systems for help. "We tried to take inspiration from absolutely everything that we could get our hands on and read about in terms of communication and collaboration," Lars said, though he avoided pointing to specific tools.

Later in the day, Wave developer David Wang pointed out that the app's concurrent editing engine - one of its more impressive tools - is based on the client-server "operational transformation" (OT) architecture introduced by the Jupiter Collaboration System developed at Xerox PARC. Basically, the server holds all the documents, and the client can't edit without sending an operation to the server. "The server takes each received operation and transforms it against the operations you've already applied," Wang said.

But Google gave OT a new twist. A Wave client can't send an operation unless it first gets a response from the server. That way, the server needn't mirror the document for each client. It can maintain a single copy, Wang says, rather than many, thus improving speed.

Wang bills Wave as the only concurrent editor that handles rich text. But Google's OT twist also allowed for something it calls "playback" - another piece of demo eye candy that had developers ooing and ahhing at Moscone Center. If you visit a document that's already been edited by others, you can playback all their edits, reviewing the changes as if they'd been screen-captured.

Wave is brimming with such small pleasures - tools guaranteed to wow an audience of Google-worshiping developers during a surprise Google demo. Some new. Some old. Some both. There's an automatic spell checker, built with Wave's extension API. There's a button that lets you respond to a particular sentence within an email, not just the email as a whole. And as Lars Rasmussen pointed out, Google has gone back to the future with its character-by-character IM chats.

"It occurred to us that by doing this live submission of characters – which by the way was how the original instant messaging clients worked – you could speed things up tremendously," Lars said. "It’s not just because you don’t have to sit and wait for the person typing, but in actual practice, you lose your attention on the conversation when you don’t see what the other person is doing."

Rasmussen and team acknowledged that you might not want the world to see every character you type - and mistype. But they haven't gotten around to adding an off switch.

'What good for the web...'

And that's Wave for you. It makes a great demo. But as Google said time and again yesterday, it's just a developer prototype. Yesterday, the company indicated that a public version will arrive by the end of the year. Where things might go after that is anyone's guess - including Google's.

Vic Gundotra did indicate that pieces of Wave will be rolled into existing Google online apps, including Gmail and Google Docs. "What you saw was an early developer release of some fantastic capabilities. Those capabilities will be packaged with Wave, but there may be other places where you'll see them," he said. "Over the next couple of months, we'll figure out how to bring these things together so that our existing hundreds of millions of users of Gmail and Docs can also get access to these things."

But the thing to remember is that Google Wave isn't just an app. It's a protocol. Or at least it will be. The idea was to create a new communication paradigm that everyone will use.

Googlers like to say: "What good for the web is good for Google." And at yesterday's press briefing, Gundotra said much the same thing. He even acknowledged that the web brings Google lots of money.

"Google has a very deep desire and interest in moving the web forward," he said. "We think everyone benefits from that, including Google. If the web becomes more powerful, as richer applications emerge, that all creates greater usage. And greater usage of the internet eventually means more Google searches. So we're not only interested in moving the web forward because it's [generally] a good thing. Economically, it's good for us as well."

Of course, there's always the possibility that Wave will fail to move the web anywhere. And even if it does, there's a limit to the cash generated by mere web usage. Otherwise, Google wouldn't be scrambling for ad dollars over at YouTube. But for the time being, Sergey Brin can afford to let the Rasmussens do whatever they want.

We asked the brothers what's next. But they wouldn't say. ®

Bootnote

You can view the full Google Wave demo here: