Hands-On with Google Wave

Ed. - To warm us up for the forthcoming Telco 2.0 Exec Brainstorms on new business models (this week in London and 9-10 Dec in Orlando, Florida), Telco 2.0 reports from last week's eComm.

As the above ought to make obvious, we've had an opportunity to use Google Wave during eComm. So what the hell is it? What's it for?

Well, the notion that it's more in competition with e-mail than anything else is roughly true. Rather, it's more like e-mail than it is like voice or video or the Web, but it's not very much like e-mail.

The simplest description of the experience is that it's a hybrid of instant messaging, webmail, and a collaborative text editor like Google Docs, delivered in a user interface that is made up of elements borrowed from several other Google services. You post text - mostly - to the thematic waves, and others can then reply or edit your text. Changes are distributed in real time. There's a contacts window, borrowed from GMail, and an inbox of things you've been using recently. As well as text, images, files, and YouTube videos will go in there too, as will small applications (known as robots).

The upshot, with a hundred-odd caffeinated geeks frantically typing stuff into it, is a fairly stimulating way to discuss and take notes in real time. However, it's also quite distracting, and the user experience is...itchy. Your typing is mirrored in real time as well, so as you try to take in information, the user interface focus shifts annoyingly every time someone else presses a key. A lot of screen space is taken up by noisy bits-and-pieces, so you have to peer at the messaging and content you're meant to be collaborating on. It is very easy to accidentally stop editing before you're finished, and have to start again. And there are some horrible bugs; waves frequently hang and refuse to open, and the application is crashy.

Of course, Google will no doubt improve these things with time and kaizen. Indeed, it wouldn't be at all surprising if a lot of Firefox/Greasemonkey developers are already working on user scripts intended to improve the look and feel. At the moment, there's little chance of full alternative clients emerging, because although Google has a draft standard for communication between Wave servers, they haven't standardised the client-server element yet.

The good news is that the Wave Federation Protocol seems reasonable - it's intended to let non-Google Wave servers communicate with each other and with Google, like e-mail servers do. It works on the principle that the originating Wave server keeps the canonical version of data users give it, and that changes are communicated between servers using the well-understood XMPP protocol's ability to transport chunks of XML as payloads in its instant messages. At eComm, Google provided that rare thing, a live demonstration that actually worked, of a Google Wave user communicating with a non-Google Wave server via server-to-server communication.

It's well worth remembering that Google's Chat/Talk instant messaging service, and its Talk voice service, are all XMPP-based; Wave can be seen as an effort growing out of GMail to integrate a variety of XMPP-based realtime communication applications.

At the moment, the only voice element is that there are a couple of third-party "robots" that control Web-telephony applications from within Wave. Martin Geddes, reliably too cool for school, turned up with one that implements a Ribbit plugin; Tim Panton of PhoneFromHere demonstrated one that used an Asterisk server's AMI Web interface and the Skype for Asterisk plugin to interact with Skype.

Similarly, Wave is a long way from even considering a business model; it's worth pointing out, as Stuart Henshall memorably twittered, that in a sense Wave is the most un-Google product ever. Google's core product is search and its core business is advertising; Wave doesn't carry ads at the moment and the search function deeply disappointed most people at eComm. In fact, Google Wave engineers present admitted openly that it left a lot to be desired. As an example, if you searched for "eComm" during the conference, you would get no results; you had to provide the rather odd search string mentioned at the top of this post.

However, we can certainly imagine that advertisers could well be interested in Wave as a conversational marketing tool. It's been said that for some reason, you can't charge for access to a Web application, even if you optimise it for mobile, but you can sell an iPhone app that wraps around the same site's APIs - the app has a sort of "thinginess" that people attach value to. Wave's "robots" could well play a similar role.