Hello, Bertrand, Felix Oghină and I worked together at Adobe trying to implement a prototype for a client-side device detection library, based on browser capabilities. Even before we started to actually write the code for the project we had decided on open sourcing our project so that the community could use our code and contribute to it.
Everything is written in JavaScript and it's meant to be flexible and easy to extend. A small demo using only dynamically modified device-specific URLs is available at http://raducotescu.github.com/browsermap/index.html The code can be checked-out from [email protected]:raducotescu/browsermap.git and you can find some basic documentation in the README from https://github.com/raducotescu/browsermap. We've already integrated it into the main product that the three of us are working on and the tests we've done so far were successful in correctly identifying all the mobile devices we had access to. There's a strong debate in the community over which client-detection mechanism is better. Both server-side and client-side detection mechanisms have their advantages and disadvantages. However we think that a pure client-side detection approach is easier to maintain and use in applications with moderately to high amounts of traffic since the server will actually only serve meaningful content instead of processing client info to determine the device type. Furthermore, relying on the *actual* browser capabilities, the results are more accurate compared to server-side compiled results (e.g. the WURFL database). If you think that BrowserMap fulfils the main goal of the DeviceMap project, which is to clearly identify devices and their capabilities, we'd be more than happy to contribute our project to ASF. Regards, Radu [cid:[email protected]] Radu Cotescu Computer Scientist Adobe Systems Romania +4(0) 31 413 36 93 (tel) +4(0) 744 406 353 (cell) [email protected]<mailto:[email protected]> Anchor Plaza, 9th Floor 26Z Timisoara Blvd 061331, Bucharest, Romania www.adobe.com<http://www.adobe.com/>
