Who cares about mobile developers? Avoid Eclipse bundle loading thread deadlock (hanging/freezing)
Jul 31

One of the most crucial entry-level barriers to mobile software development is the wide variety of platforms. To date there are thousands of Java-enabled handsets. Large companies attempt to purchase as many devices as possible and thus cover their bases but what about the small and medium-size players?

Is buying every single device on the market an option? Clearly not.

One idea being tossed back and forth in the mobile industry is the concept of an universal emulator. That is, an emulator which would be able to emulate every single device out there. At present, device vendors such as Nokia, Motorola, and Sony Ericsson have been very good at providing free SDKs and emulators. The problems are still manifold. First of all, not every single device is supported, and the ones that are supported differ in the level of emulation accuracy compared to the actual device. In addition there is no real standard in terms of emulation, other than the Unified Emulator Interface (UEI) which is starting to pick up pace (using UEI emulators). Even if this does become the standard, we’d still have to support non-UEI emulators. Furthermore, managing all emulators from every single vendor would be a nightmare.

Here’s a proposed idea: test every single MIDP and CLDC api call on every single device

This is an almost impossible task of course. But if we did have this kind of data, and if it were always fresh and up-to-date, then we could more easily reproduce API bugs. We would essentially be aware of all KVM discrepancies in the market. We could even build a UI tool to emulate a real KVM which would thus reproducing every API bug on every device.

That’s as close as I can think of getting to the Universal Emulator idea.

Leave a Reply

You must be logged in to post a comment.