Google Wave dev preview hands-on and impressions

After an impressive debut at Google I/O, the company's newest experiment and collaborative chat client has been making its way into the hands of developers in the lead-up to a torrent of new testers on September 30th. We had a chance to stop by Google's San Francisco office last week for a guided tour of the latest build of Wave with creators Lars and Jens Rasmussen, and have since then spent the better part of our free time working through the ins and outs of the new communication platform. Does it live up to the hype, even in this bug-infested interim build? Read on to find out.

"Everything's shiny, Capn'. Not to fret!"

Firefly fans may instantly recognize that quote, but participants of the Google Wave dev preview tend have an almost Pavlovian aversion to the phrase. For them, it's a signal that you've managed to find a bug that's crashed the program, grinding your experiences to an abrupt halt. "We're working hard on three things right now," said Lars, "stability, speed, and there's a stack of usability problems that recent users have uncovered for us." On its September 30th launch, there won't be any surprise features from what we've already seen shown -- "nothing new, but less of the 'shiny,'" he said, referencing its fail whale error screen. Our first ten minutes with the web app were apparently pretty typical for new users, tinkering around with every feature as we write and edit each other's incoherent babble, watching in a stupor as characters materialize on screen in real-time with the other person's typing.

The interface is easy to adjust to, especially for those familiar with Gmail. From start up, you've got your navigation and contacts on the left rail, your inbox in the middle, and your current open waves on the right. Any window can be minimized, and doing so will send it to a tab along the top row adjacent to Wave's logo, and can be fully accessed from there as drop down windows. Honestly, we found this preferable to having the non-wave windows propagate the screen, as it cleared up valuable screen real estate and gave us more room to spread out multiple wave conversations. Unfortunately, its system for organizing multiple open windows was puzzling: with five waves open, four vied for space in the left column while one particularly empty wave hogged the right all to itself. It was pretty illogical, and we've got no idea at this point how to drag them around and fix, but we're willing to go on faith for now that this will be remedied before too long.

In any given wave, you have the option to start a new entry, edit an older one, or even edit someone else's message. Replies can be situated after a thread or even embedded in the middle of another message altogether, and each one can be either public or private / shared with only a select number of participants. Pasting in YouTube and Google Docs links give you the option to turn the URL into a working embed, and with Google Gears, you can drag and drop images to create a gallery -- in theory at least, we're still having trouble testing that last bit ourselves. Any time you hop into someone else's message to edit, the byline changes to reflect your contributions, but at that point, there's no way to tell who wrote what without going step-by-step through the playback history. In this sense, Wave's best suited for use as a collaborative tool, and Lars' real-world example here is contract negotiation, whereby multiple participants go through a legal document, create in-line discussions (public and private) of sections, and copy over sections into new waves for more fine-tuning. Lars noted that the underlying algorithms link the original and copied wave together, and further down the road, there'll be an option to synchronize the changes you made in the copied wave with the original. At this stage, there's no export to Google Docs function, but that could change in the future and, either way, someone could write an extension to get the job done.

One of Google's initiatives to attract business / enterprise customers is the ability to create your own Wave server that doesn't live in the cloud, and as it was explained, any part of a wave that's privy only to people on the enterprise server, including private in-line replies to public threads, will exist only on the local server, while portions shared publicly or with a member outside of the server will co-exist in the cloud.


One of the coolest tricks we saw at Google I/O -- certainly the one that earned the loudest applause -- was real-time text translation. We haven't done any language conversions ourselves, "Swedish Chef translator" notwithstanding, but we did get a glimpse at what else Wave's extensions would be capable of. Some of the more interesting (albeit not necessarily exciting) examples were kasyntaxy and i-cron, two bots that you can add as users to a wave that would format and execute your code, respectively -- again, not the most thrilling demonstration, but it does an apt job of showing off the capabilities and potential use beyond novelty. Unsurprisingly, there's also a Twitter extension in development, for making a wave dedicated to just reading your feeds or for sending out updates through the platform.

Other than an ad hoc wave listing extensions, there's no integrated database for developers to show off their wares. To our surprise, though, Lars said the team is toying with the idea of an app store with revenue sharing. He was quick to point out this was just one of many monetization strategies being floated around at this point, but we gotta say, it's definitely an intriguing idea.


Back at Google I/O during a Q&A session on Wave, no one would say for sure if or how the live updates would work on the mobile platform. Fortunately, someone on the team found a way to make it work as recently as a few weeks back. It's confirmed to work on your iPhone or Android as a web app -- it's even got a home screen-friendly icon to boot -- but at this point, and by Lars' own admission, it's too buggy to really work at all. At our meeting last week, Jens was able to go into a wave via iPhone, watch the text update live, and even go through a gallery of pictures. As you can see in the video, however, we've had nowhere near the same stroke of luck, instead getting stuck with the "shiny" error message with every attempt to peer further into our inbox.

The web version's built around the HTML5 standard, and something Lars is excited about is that it's the same code base with only five to ten percent effort to make it work on mobile, by his estimates. A native app has been discussed, and while it would allow access to the camera, address book, and compass (as Latitude has shown, GPS is already web app-friendly), it'd take an "extraordinary amount of work" to build native software vs. just sticking with a web-based solution. We gotta imagine bypassing the app store would be nice at this point, too, but going native is still an option at this point. For the most part, there's expected to be a feature parity between the desktop and mobile versions, with one key exception: editing. It's something that's not quite a part of the HTML5, but the Rasmussens said they work close with the standards body and are hopeful they can fix that.


Put plainly, Wave is a message board in a constant state of self-editing flux. To put it another way, it's a collaborative text editor on steroids -- as our mobile editor Chris Ziegler quipped, "I think Wave's capabilities actually exceed human capacity to interact." There's a lot of great ideas here, many we'd love to see find their way into other Google apps, but we're still trying to figure out just how this will fit into an online world already dominated by numerous social networks, messaging clients, forums, and microblogging services -- something that'll proper render itself more clear as it moves from dev build to public beta. It's not that we don't think Wave will be successful -- and make no mistake, we are enthused -- but calling it "revolutionary" might be setting the bar a bit too high. Those 100,000 spaces opening up September 30th got filled just hours after the I/O conference, so if you signed up any time in the aftermath, you're gonna have to hold off a bit longer. In the meantime, it's probably best to pare down those exorbitant expectations while you wait.