Sun and ARM are teaming up to make Java faster on cell phones with Jazelle (Java on a chip) technology.
This will allow games to execute much faster. It gives a needed boost for enterprise apps too, but keep the client thin since it's not enough boost for wireless provisioning or high-end (XML/SOAP) processing yet.
Stay simple with your wireless apps now. The cell phone is slowing evolving and will get to much needed J2ME CDC technology with better performance in about 2-5 years.
[Java ME and J2ME] ( September 07, 2004 09:09 AM )
Permalink
The right way to design current and near-future J2ME apps for cell phones is to first start out with a design that can run disconnected. For now, a J2ME programmer should not rely on the network connection giving access to a back-end server.
Here's a good example of a recipe database and viewer. Nice and simple. No need for a network connection. It refreshes data on demand (Pull Technology) and does not rely on being always-connected.
Keep hooks in your design for future back-end server connections over the network that will be always-on, but for the first version have it be standalone and refresh data on-demand (when you know you have a good network connection). This will give better customer satisfaction and will work with current and near-future technology instead of having to wait until the 2.5G and 3G networks catch-up to handle true always-connected Web services.
[Java ME and J2ME] ( September 03, 2004 09:34 AM )
Permalink
Sometimes the cluelessness of the Wireless industry makes me sigh. :-|
There's the very common mistake technologists make of envisioning something that is AOIT (ahead of its time). That's the case with Web services on cell phones. Companies are trying to stuff SOAP, XML, OSGi, downloadable/provisioned Java apps to cell phones way before the technology is ready for that.
Wrong move. The flakiness of 3G networks, the CPUs on phones, the OS on phones, and the J2ME functionality (especially as it relates to speed of processing and size) indicate that trying to use the typical Web service client stack will still have to wait until the technology can catch up.
You have to learn to crawl, then to walk, before you can run. You can't go from simple MIDP 1.0 games to Web services. That's like learning to crawl, then attempting to run the marathon. There are a bunch of steps needed in between to get from the beginning to the end goal.