Sorry, the blog title is borrowed from the TED website but I love this resource.

James Howard Kunstler: The tragedy of suburbia

Theo Jansen: The art of creating creatures

Deeper insights into the 7th Mass Media channel, mobile is to the internet, what TV is to radio by Tomi T Ahonen

On April 30, 2008 Sony Ericsson announced its new Project Capuchin technology which is a Java ME API that defines a bridge between the Java ME and Adobe Flash Lite programming environments. This API makes it possible to use Flash Lite as the front end and Java ME as the back end of applications, meaning that Flash tools can be used for UI design while still having access to all the phone services available to Java ME.

This is an interesting project and will certainly make mobile applications easier to prototype and build. Although it's a proprietary Sony Ericsson technology I wonder if the other device manufactures will do something similar?

Sony Ericsson's new Project Capuchin bridges Java ME and Flash Lite

We haven't had much luck finding any one up till now but we are still looking for someone to join us as a full time Java ME programmer. Originally we were looking for someone with Java ME experience but it seems those type of people are *very* hard to find thus we have changed our requirements somewhat and are now looking for a young London based enthusiastic Java programmer who wishes to get into the fascinating world of mobile technology. Ideal candidates don't necessarily need a degree but their Java skills need to be impeccable and, most of important of all, they need patience and a strong problem solving ethos.

Would you like to work on cutting edge projects like Palringo? Would you like to work in a small dynamic team who spend 100% of their time working on mobile applications and projects? Do you want to learn everything there is too know about Java ME? Do you have a mobile gadget fetish and want to have access to to a bucket full of devices? Would you like to be trained by a kick ass Java ME guru (my business partner Steve)? Is so please email me as we are looking. Thanks!

If your email to jason@paxmodept.com bounces (due to our mail hosts overly zealous spam protection) try jason.delport@gmail.com

I get plenty of searches arriving at this blog looking for information about the infamous "907 Invalid JAR" Java ME OTA provisioning error. After 6 months of solid data from our provisioning platform I feel I am finally starting to understand the causes of this frustrating error.

Update: After Enrique's comment I decided to make my post more comprehensive so I have included the causes of 907 errors that one would encounter during development / early deployment phase. I will add to this post over time so that it becomes an all inclusive help file. Let me know if you have anything to contribute.

(1) Your manifest file or JAD are incorrectly formatted. Ensure the end-of-line characters are consistent and that the files are properly formatted (no weird spaces/line breaks). My recent troubles downloading the CNN application were as a result of a badly formatted manifest file.

(2) Your manifest or JAD are missing attributes (MIDP/CLDC declaration for example). This is easy to do if you are manually creating or editing your descriptor files. Rely on your build tools (J2ME Polish, Antenna, Netbeans, etc) to do this for you.

(3) The JVM you are targeting doesn't like something about your JAR. Even though your application works perfectly on Nokia and Sony Ericsson devices be prepared for some JVMs to throw this error when trying to install the application. The only solution for this is painful manual debugging. Strip your application down to the basics and slowly add features until you find the offending bit of code then refactor mercilessly until it works.

(4) The download times out and the network operator proxy server sends a message which the Java Application Manager (JAM) doesn't understand and it throws a 907. Ask your user to try again, maybe at a different (quieter) time of day. If people are persistent on slow networks they can still get the application. This error happens fairly regularly on all devices on slower 2.5G networks.

(5) The certification validation fails. If the application is certified ensure the device has the correct root certificates and the device date/time are set correctly.

As you can probably tell from the lack of updates recently I have been really busy and I have felt guilty about blogging when there is so much work to do. Most of my development work in the last few weeks has been mobile web stuff and it has given me the opportunity to do a lot of thinking about the current and future direction of mobile web development.

Some thoughts from this time are below.

What is going on at Motorola? Their new Symphony browser is pretty bad. I left the London MotoDev Summit feeling optimistic about where Motorola were going but after having had to battle the PEBL2 U9 browser recently I am once again a sceptic. Although most people still don't use the mobile web there is a clear industry expectation that mobile browsers need to get better. This just isn't the case with the Symphony browser which is in fact a step backward from the version of Opera (8.5) they had in the RAZR2 V8.

Why are Nokia putting so much effort into the WebRuntime when it would probably be better spent on the actual S60 WebKit browser itself? The beauty of the web is that you don't have to install anything (most of the time at least). I would prefer if Nokia spent more time ensuring the features available in the WebRuntime were implemented in the normal S60 WebKit browser and developers could just build sites like they would normally do but have access to some low level device features using Javascript.

On the subject of the S60 WebKit browser, why don't Nokia provide more detailed information about creating really great sites specifically for this browser? The new beta Yahoo! mobile portal, which uses more advanced browser features, looks great and works really well. Surely the number of WebKit browsers out there make this feasible for developers? There are fewer iPhones than S60.3 devices but developers seem happy to create sites specifically for them. The Javascript and DOM support in the S60 WebKit browser is actually pretty decent and I could imagine creating some interesting web applications if so inclined.

Device Atlas is looking more and more interesting to me. A $100 annual charge is trivial for most companies and if the data is as accurate as they say it is then it is worth the fee. Also, the size of the data (I seem to recall someone claiming it was in the region of 350Kb) is appealing. I will wait for the Python API to appear and then I am going to migrate our mobile site to the new Google App Engine as a proof of concept for both technologies.

As an aside, I have a music mix up over at Whomp! FM. It's under the name JfMaMjJaSoNd and it's a slight deviation from the norm for the good folks at Whomp! FM as I am more inclined toward rock than electronic music but they insisted. Enjoy!

We have been collecting JVM names (as displayed in the UAProf document) for a few months. The list makes for some interesting reading.

Aplix_JBlend
aromaMVM/2.0.1
aromaNG/1.0.0
Aromasoft VM
CLDC-HI
ESMERTEC JBED6
HotSpot
JbedME4.0
JBlend
JBLEND/2.3.8
JMM 10.1.2.76
LGVM/1.1.2
microJBlend/1.0.6
microJBlend/2.0
Monty VM
Monty VM, J9 VM-2.3
Motorola VM, Symbian 9.2 UIQ 3.1
Philips/JWare-v3.1
QVM 4.2.2
QVM/4.0
QVM/4.1
SonyEricsson Enhanced VM
SUNCLDC-HI/1.1.1
SunJRE/1.0.3
SUNKVM/1.0.3
SunKVM/1.0.4
SunKVM/1.1
SunKVM/1.1.3
SunKVM/2.0.0
Tao MIDP 2.0: v10.1.7.571

I don't know what Motorola are doing. We own the RAZR2 and the PEBL2, both are Linux feature phone devices, both are *very* average. The RAZR2 was okay... the UI still needed work but it was on the right track. The PEBL2 is a good looking small phone and the UI is certainly an improvement over the RAZR2 but the new Motorola Symphony browser and JVM on the PEBL2 are driving me nuts! This new Symphony browser seems like a massive step backwards compared to the version of Opera (8.5) that was on the RAZR2. As you can see in the screenshot below the Symphony browser doesn't fair very well in the new W3C Web Compatibility Test for Mobile Browsers.

Even more annoying is the fact that we are unable to install *some* midlets onto the PEBL2 via our provisioning platform. Now I spend a lot of time debugging these types of issues and have a rigorous problem identification process in place but I am completely stumped as to what is causing these particular errors. Arrrgh, damn you Motorola!

The W3C has released a web compatibility test for mobile browsers. You should also read the original blog post.

The Tests:

CSS2 min-width
Transparent PNG
GZIP support
HTTPS
iframe inclusing of XHTML-served-as-XML content
Static SVG
XMLHTTPRequest
CSS Media Queries
Dynamic SVG
The canvas element
contenteditable
CSS3 selectors

Here is a Wiki entry from the Open Ajax Alliance discussing Mobile Device APIs. I am fascinated with the idea of the mobile browser becoming a miniature application execution environment and this page is a good over view of where we currently stand on the matter.

Last but certainly not least Luca Passani's Rules for Responsible Reformatting document has been signed by two transcoding proxy vendors (InfoGin and OpenWave). This is great news. The W3C is a vital institution doing an indispensable and difficult job but like all bureaucratic organisations in can get bogged down in it's own processes. Sometimes you need an outside voice to shake things up and get things done. This was the case with the WHATWG and it's certainly the case with what Luca has been doing. While I don't always agree with Luca's tone and manner I really appreciate the way he has represented the small guy in this matter. Luca has ensured that our real world concerns didn't simply get made a foot note in the minutes of a W3C meeting; he got results and that's what counts.