Lars Rasmussen and his brother Jens, the creators of Google Maps, trotted out a first look at their latest development, Google Wave, at the I/O developer conference last week. The audience gave them a pass on a few mishaps, but overall the new application was quite impressive. Wave introduces new concepts in communication by thinking of conversations as container objects where you can drag and drop people, threads, links, documents, images and even robots that enable the content to instantly appear on connected clients over the Internet. Wave is written entirely in the Google Web Toolkit (GWT "Gwit"). The developer codes in Java and the tool converts to HTML5 & Ajax automatically. For developers, Wave stores updates to UI state in the local XML of your gadget. Then Google transmits that state over the network where the other instance of your gadget updates in real-time. Google estimates only 5% of the code needs to be adapted for mobile device browsers; most of which involve just a layout change. Wave removes the structure found in email replies by creating a hub of conversation trees where users can chime in at any level, playback what they've missed and leave replies for others to see. There is even support for Twitter using Twave to merge wave posts to and from the popular micro-blogging platform. My main concern is the amount of network traffic generated by hundreds or thousands of users in a real-time, collaborative web application. Another open issue is federated identity; Google requires users to have an account to access any of their applications. OpenSocial gadgets will be supported natively in Wave and I expect to see more consolidation in the social web authentication space this year. No wonder Google was quiet about Twitter acquisition rumors, they've got bigger ideas.