16 March 2006 · About 2 minutes read

Stop Press

Yesterday’s Apollo event dry-run of toupix has to be pulled at the last minute due to the ever occuring technical hiccups. It seems there is a problem running a fresh installation of Toupix which meant that the camera could not be initialised. A long test seemed to indicate a bug in displaying a temporary Alert during the camera’s initialisation. This alert has now been removed and the camera works fine.

However, in the process of that test, a couple more errors were found. Most worryingly, the legacy Bluetooth code (one of the first parts of the project to be written several months ago) seemed to cause an endless thread loop that meant taking more than 3 or 4 images was impossible. After polling, the repeating Thread would prevent any file data access. As a result, I took the decision to completely rewrite the Bluetooth code from the ground up, and I am glad to say that the new implementation is far more robust and extensible than the previous set of classes.

Finally, I have added a signup form to the MIDlet, so that new users can register a Toupix account directly from their phone. This will help to lessen the numerous installation/preparation steps that current testers have to go through in order to use Toupix. Based on some feedback so far, the process needs streamlining. The indended outcome of forthcoming developments to achieve Install..Run…Use…

Digital Signature

Another problem I currently have little control over is the interruption of permissions. They are an obvious neccessity for the phones’ security, particularly when accessing user files. However, the persistent request to use the camera, read and write user data whenever a photo is taken does somewhat break the continuity and UI of the MIDlet. I am looking into the option of getting the MIDlet digitally signed (through something such as Javaverified, although further research is needed into the various providers and the cost of a signature.

So, with rewritten code and still plenty to do, the beta has slowed to a crawl. However, with the latest additions and more robust codebase, I’m hoping that the next Apollo in 2 weeks time will be able to see the launch of the (fully) working beta.