Tuesday, December 08, 2009

Google Apject Takeover

Greeting from the Sydney Wave User Group meeting. The first topic is the takeover of Apject by Google. There was some concern by the user community that EtherPad's non-Google implementation would be discontinued. In response at short notice a decision was made to provide it as open source.

Pamela Fox from Google gave a quick introduction to Wave, explained as three components as a product, protocol and a platform. The product is Google's web based implementation. The protocol includes a way of having transactions and a data model, this allows companies other than Google to implement Wave. The Platforms include Robots, Extensions and Gadgets. Robots are computer applications which can participate in the conversation, assisting human participants. Gadgets are small applications which extend the wave interface. Extensions allow Wave content to be integrated with other systems.

Wave uses XML for storing and communicating. Robots operate on the XML "Wavelets. Each Wavelet has participants, a title and one or more "blips", which represent the atomic components of a conversation. The blips can be manipulated by Robots.

The Blips are text, annotations, and elements. Annotations are similar to HTML markup. Elements are added items such as a map. As these components are kept separate this allows easier manipulation of the information. In a way the Wave way of doing markup is analogous to Wiki text and designed to make changes, particularly by multiple users, easier. This allows, for example, chnages in real time to the mark-up of a document. This is one of the features which Google find "cool", but I find confusing.

Robots react to events and perform operations. Each robot has a URL which is sent events and performs operations on them. The state of the gadget is stored in the XML sent to the gadget and is therefore not secure. Also there are all the usual problems of real time and multiple updates of the same data. An example of a robot is one which goes through waves converting web addresses to links.

ps: This is about the third Google Wave event I have attended over the last six months. The first glimmerings of understanding of what Google Wave is are now starting to form in my mind. This is three times as long as it took me to get "The Web". I remain to be convinced that Wave is the next big thing or is significant at all. The Web turned out to be based on a few simple ideas and lots of reuse of existing technology. I am not sure Wave could be similarly decomposed.

No comments: