Yesterday evening’s MoMo London gave an upbeat view of the current state of mobile Java. It was panel event with Simon Nicholson Sun Market Development Manager (and on the board of Java Verified), Sergio Falletti of Future Platforms, Andy Milloy from SurfKitchen, Kieran Gutteridge from Intohand and Adam Cohan-Rose from Kizoom.
Here are the main points I took away from the event…
- Mobile Java is still growing.
- Consider just one platform (Java, Symbian, iPhone whatever) if you think you can get decent return.
- One of the biggest problems of mobile Java is end user discovery.
- Java ME is getting easier because larger phone memory sizes mean apps can include everything and query the phone for capabilities and use just those.
- Java ME is getting more viable because there has been a big enough churn of devices since MIDP 2 (3 years ago) so that now 70% to 80% of phones are MIDP 2.
- The latest features (low down in phone api) will always be the most fragmented.
- LWUIT is large and sometimes needs parts stripped out to provide space for the main app logic.
- For testing, use the Java Verified list of lead handsets to get a representative group of phones on which to test.
- It’s now possible for 3rd party developers to contribute tests to Java ME TCKs.
- Consumer awareness of being able to download applications (to iPhone) has had a knock on effect of increasing awareness of being able to download Java applications.
- The Java community-driven process for enabling new features is slow and will never be able to compete with a closed process such as Apple’s.
- Features that might make MIDP 3 popular include access to telephony events and Java components (much like DLLs)
- Java still provides the greatest reach and can be ported to Android and BlackBerry with relatively little effort.
- For developing countries, Java is the only option. (Interesting news today on Microsoft’s OneApp Java application is an example of this)