08/02/2012 By Dirk 0

9 months of Android Programming

About 9 months ago, I started my first endeavour into Android programming by porting the former Java MIDP platform NORS onto Android, now released as the AIRS platform. It was positively surprising how quickly I made progress. AIRS has progressed since with many sensors being added and stability having improved. While NORS suffered constant instability (on my latest Nokia phone, the software barely ran longer than 30 mins before needing a restart), AIRS can now record an entire day for the personal recordings that I am currently doing. I usually start in the morning and quit AIRS when coming back from work.

That says a lot about the stability provided by the Android platform. Considering the wide variety that I am usually recording (ranging from GPS over Bluetooth, sensors like pressure, light and proximity as well as system info such as RAM, running tasks, battery etc), this stability is indeed surprising and encouraging.

What has been even more encouraging is the usage of the Android market to disseminate the software. While AIRS is hosted as an open source project on Github, it was the release on the Android market that has led to now more than 2000 downloads of the software (with more than 300 people still having AIRS installed on their devices). Given that AIRS is only recording (Dana Pavel has recently agreed to include her parsing software into the open source release!), it is surprising that so many people tried out the software. With the Android Market providing a very good model for error tracking (these errors reported by end users who click the “Report” button on each force close of the software), the software can be constantly improved and stabilised. This shows the power of the “app model“, where ideas can be easily released and therefore tested by a wider audience.

This is even more important for software released as a result of research projects since it opens up a new and promising route for exploitation and dissemination. Based on the encouraging results from AIRS, we recently added this route for dissemination to a project proposal we submitted for funding. Furthermore, current efforts at Cambridge University aim at porting the PURSUIT project prototype Blackadder onto Android. Also for this effort, we plan to release the port to the Android market, provided that we solve certain usability and usage issues.

9 months into Android programming, my intermediary conclusion of this ‘experiment’ is very positive and I encourage other researchers to utilize the power of the “app model” for their research dissemination efforts. It is well worth!