Mobile Java Platform

Terrence Barr - Evangelist, Java Mobile & Embedded Terrence.Barr at Sun.COM
Wed Jan 31 11:17:38 CET 2007


Chris,

There is quite a lot of interest in Java on OpenMoko, both
from the Java SE as well as the ME perspective. I think Java
will be critical to the success of OpenMoko given the huge
software base that would be instantly available once the
Java stack(s) are ported to OpenMoko.

Based on the specs of Neo 1973 I believe Java SE is possible
but it would probably tax the system quite a bit so I think
it is not a slam dunk.

There are two important parts to a Java SE port:

- The VM itself
- The class libraries

As for the VM there are several open source options such as
JamVM, Kaffee, Cacao, and Apache Harmony. Of course, now
there is also Sun's Java SE implementation that is being open
sourced under GPLv2

    http://www.sun.com/software/opensource/java/index.jsp

As for the class libraries there is really not much point in
reinventing the wheel here multiple times. The GNU Classpath
project is seriously considering moving to Sun's class library
implementation now that it is under GLPv2. I expect that
eventually the market will rally around Sun's implementation.

As for the effort of a complete and functional Java SE port
I can't really make a statement ... I haven't worked on such
a port before. Since compatibility with other Java SE
implementations is key there will have to be a lot of focus
on compatibility and testing (including running TCKs).

Java ME (both CDC as well as CLDC configurations) are both
a good match for the Neo 1973 specs. CDC is based on Java 1.3
so that gives you a high level of funcationality. CLDC is
a subset of CDC and has the advantage of running the MIDP
profile which is present on nearly a billion phones today.
There are a lot of interesting things going on in the Java ME
space, including very cool UIs and Web 2.0-type applications.

As for open source Java ME stacks the options are fairly
narrow. There are some Java ME VMs in open source but I think
they are all somewhat lacking resources and are incomplete
(correct me if I am overlooking something). But since Sun
has also released it's entire implementation of both of
it's Java ME stacks I suggest this is the way to go:

    www.mobileandembedded.org
    https://phoneme.dev.java.net/

phoneME Feature is a full-featured, product-quality CLDC/MIDP
stack with additional JSRs. phoneME Advanced is a product-
quality CDC/FP/PP/PBP stack for advanced mobile and embedded
devices.

Both are developed in the open and present the cutting edge
of the Java ME platform market. They are designed for
portability and have been ported to a number of platforms
already, amongst them Linux on ARM.

So I think phoneME is an ideal candidate for OpenMoko.

Where to go from here:

I am working with Sean on how we can proceed. I believe it would
be best to create an official phoneME sub-project for the
OpenMoko port so that it becomes a part of the standard phoneME
repository. As a side effect the OpenMoko effort can then
leverage the expertise and infrastructure of the phoneME
(Mobile & Embedded) Community.

Let me know if that answers some of your questions.

PS: SuperWaba is *not* Java. I know SuperWaba has been around
for a number of years and is based on the open source Waba
project but I am not sure how much traction it has. There is
a lot of innovation and market penetration of Java ME that
SuperWaba is not participating in so it appears like a niche
solution.

-- Terrence

Christopher Heiny wrote:
> [I've cc'ed some folks who expressed interest in Java on the community list 
> (at least in the messages I've kept around).  I don't know if you're on 
> -devel - if you are, I apologize for the double posting.  If you're not 
> actually interested in Java on Openmoko, I'mm sorry to have bugged you.]
> 
> On Tuesday 30 January 2007 11:38, Jonathon Suggs scribbled in crayon on the 
> back of a kid's menu:
>> I had looked into developing some applications based on this platform a
>> while back and just got an email about it latest version.  Anyway, I
>> don't know how much work it would be to get its VM ported to our
>> platform, but if we did then we could instantly get more applications
>> that were compatible with our device.
>>
>> >From their website: SuperWaba is an open-source software development
>>
>> platform for PDAs (Personal Digital Assistants) and Smartphones.
>> http://www.superwaba.com.br/en/default.asp
>>
>> Let me know what you guys think...
> 
> Hmmm.  It's hard to tell on a 5-minute examination whether the Professional 
> version is fully opensource or not.  Their comparision page indicates that 
> you need the $495 SuperWaba professional to get image handling, email, 
> html, ftp, and printing.
> 
> Of the options that folks have mentioned on the community list, I think 
> JamVM might be best.  I believe that it's already checked into 
> OpenEmbedded, so it should be easy for it to be part of OpenMoko.  An email 
> here
>     
> http://lists.linuxtogo.org/pipermail/openembedded-devel/2006-December/001036.html
> Koen describes some first steps on getting Java installed and going (note 
> that the PDF linked in that mail is broken).
> 
> I noticed that the projects.openmoko.org appears to be up and running now.  
> Maybe we should set up project with the goal of getting Java going on 
> OpenMoko (probably via OE) and maintaining it?
> 
> 					Chris
> 
> [Disclaimer: Though I've been messing around with J2SE for about 8 years 
> now, is is my first venture in the ME world of Java.  It's kind of an 
> unfamiliar & bewildering territory, so please correct any egregiously 
> boneheaded statements I might make.  Thanks.]
> 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: terrence.barr.vcf
Type: text/x-vcard
Size: 289 bytes
Desc: not available
Url : http://lists.openmoko.org/pipermail/openmoko-devel/attachments/20070131/6629effd/terrence.barr.vcf


More information about the openmoko-devel mailing list