Skip to main content

My views after the Mobile Developers Alliance BOF

No replies
Joined: 2003-06-15

So the Mobile Developer Alliance BOF on Tuesday was lackluster to say the least.
From Sean's verbal count there were 2 or 3 independent developer, and 1 or 2 corporate developers.
I being one of each that takes the count of developers able/willing to take time to show up to max 4.
Sad, truly sad. There were a couple manufacturers and 1 operator represented, if I remember correctly.

If we the developers don't care enough to show up to our own lobby group meeting then why should the big companies bother listening to us?!?

I think Terrence and Sean put it well as have many Sun execs during Keynotes of the past....
Developers have the ideas and can build the cool apps that will drive consumers to buy there services and sign those 2yr contracts. As a friend of mine that works for an Operator has told me, they are looking at the users that buy the big device/services like the iPhone as the customers they want, ones that buy the phone (no discounts), and pay a nice monthly plan for services. But it would seem they want the consumer to buy the high prices services in hopes they will not use them, thus putting the CON in Consumers. :)

Sean put up a chart showing quadrants for Independent thinkers & Puppets of Industry on the Y axis, and Ignored & Supported on the X, and saying what is the optimum goal would be an Independent Thinker that was supported, but typically there is a tendency for there to only be Independents that are Ignored, or Puppets that are Supported... I would tend to agree.

However the issue that I feel has been overlooked is the fact that cost of entry into the mobile developer space is too prohibitive. If you can't afford to build apps then you will move into other markets. You can see plenty of interesting apps in the BlackBerry market, I expect we will see cool apps when the iPhone SDK has it's storefronts activated, and WindowsMobile has lots of apps available.

So what is there secret? They control the signing certificate, and it is reasonably priced, in most cases. WindowsMobile while the most expensive using the Verisign ACS portal is still cheaper than Java, and your signed .CAB file will work on operators networks. In the case of Java just because your MIDlet is signed does not mean operators will allow your app on their devices/network. How silly, WinMo a known hackable, virus compatible environment is allowed to run free, while JavaME based on a secure design is punished.

So how was this thought taken at the event? It was immediately shot down, it seems the focus that Sean and Terrence had was for more influence be placed to independent developers on the JCP's Exec Committee and Exec Groups. Why? The specs are good, and that process tends to work. Everyone is concerned with Fragmentation... What Fragmentation?!?! I've built MIDlets for BlackBerry, Nextel Falcon series, Samsung, Nokia, Motorola, IBM J9 vm, SonyEricsson, doing things such as custom UI, 3d, Video Capture, Content Posting, Data Persistence, PushRegistry, SMS, WebServices and Encryption, deploying a single MIDlet Jad/Jar to multiple devices with no custom device code.

Here maybe the biggest concern to the operators, bad publicity, and mis-behaving MIDlets on their network.
This is NOT the developers sole fault, the tools for building and testing are a little crude, and as everyone knows if you didn't test it on that phone you don't know that it will work. One such case I ran into was with the K800i, Z650i & the K790. While the K800i and K790 are almost identical (800-uk/790-us) they behaved completely different when capturing video. The Z650 is supposed to have the same support as the k800, but did not playback video while recording. But a little coding and the same MIDlet was deployed without a bunch of if( K800i) then... else if(K790) then... else if(Z650).....

In the end I feel it is a MUST to get Operators to allow 3rd party apps on their network. If we can't push our apps to devices then it doesn't matter if the specs are detailed, complete, and fully compatible between devices and operators. If it is too costly to develop apps then no one will become a developer and learn the right way to program JavaME. But we need a company (like SUN) to take the lead and build an Identity system and service that operators can feel comfortable in allowing a signed MIDlet through that service onto their networks without years of paperwork, costly physical device testing, and network analysis that not even the operators understand.

At this point I am not feeling overly confident that the mobile developer's voice will be heard, I hope Sean and Terrence prove me wrong, but since the first part of February when this group was started there have been no posts on the project page, and the Wiki is silent.

Maybe I'm wrong, maybe I'm right, maybe there is a gray area that addresses everything. I can only hope we find this area and soon.


Thanks for listening.

PS. Sadly I missed the Post BOF ThirstyBear session so maybe more was discussed, but food and sleep was more important at that point. :)