Java Development

[superdirt]

On another note, I ported most of one of my BBOS applications to the the C/C++ NDK for the PlayBook in under a week of downloading the NDK. It's been years since I've previously done any big C/C++ projects, so I spent a little time refreshing those skills. I made a few bonehead moves in the process, but I used C/C++ for a few years and within minutes I remembered most of everything I knew.

QNX is a robust, mature platform. I prefer power over ease, so the NDK is a great choice. The speed rivals that of a desktop computer. Threads can accurately sleep for microseconds, not inaccurately in milliseconds (one of my biggest pet peeves about BBOS). However, the first thing you may notice about using the NDK is that UI support is not rich in the current API. The main modality for graphics manipulation is OpenGL ES, a graphical system designed mainly for use in gaming, not for basic UI. I knew zero about OpenGL ES before jumping into the NDK, so that was my biggest road block.

GL ES is an entire language within itself, but really this is not any different than UI for BBOS. I've said this before: about 50% of programming with BBOS is spent on developing the UI - maybe more. In BBOS I was overriding so many fields to customize them, that in the end I was feeling held back a little bit and was seeking a little more power from BBOS. Now using GL ES, I am probably doing more programming to create the custom UIs than I was in BBOS, but I have considerably more power over the UI. The rich UI API is being made available at a later date. I know lots of programmers dislike having to write their own UI features, so I am warning you now, you might not like the current version of the NDK. Really though, creating basic buttons and field managers is not a whole lot of work and you can reuse that work in other projects. At the risk of tooting my own horn, I must say that I have created some pretty cool looking UI features for the PB.

Honestly, PlayBook OS is really growing on me. I have never learned how to program another mobile operating system faster than I did with the PlayBook. I'm excited about the changes with BB (and a little scared). If PB OS is a sign of what BBX will be like, BBX is going to be great to work with.

My business strategy for the short term: In the mean time while I wait for the release of BBX, I will be moving my core applications to the PB. Working under the assumption that programming for BBX will have many similarities to PlayBook OS, my core applications will only need tweeking when BBX is released. I'm confident in the PlayBook. In a few months when I show a large entertainment company the demo for an application they want produced, I plan on including a PlayBook version along side versions for Android and iOS.

Scott

[smiley]

---

superdirt wrote:

My business strategy for the short term: In the mean time while I wait for the release of BBX, I will be moving my core applications to the PB. Working under the assumption that programming for BBX will have many similarities to PlayBook OS, my core applications will only need tweeking when BBX is released. I'm confident in the PlayBook. In a few months when I show a large entertainment company the demo for an application they want produced, I plan on including a PlayBook version along side versions for Android and iOS.

Scott

---

Glad to hear that you are moving along and thanks for sharing.

I assumed most BB JAVA Devs are not game developers but writing utilities that integrates into the BBOS.  So the current NDK is kinda useless unless people want to get familar with NDK and study the sample apps.

[silver20]

So, BBOS was a diminished experience, but between Android, Flash, and the NDK itself, there's no UI consistency on the platform? Another "is coming Later", like the N.Y. Times PlayBook review said.

[superdirt]

I assumed most BB JAVA Devs are not game developers but writing utilities that integrates into the BBOS.  So the current NDK is kinda useless unless people want to get familar with NDK and study the sample apps.

Agreed, I think most Java devs will not like the lack of UI support in the NDK. The NDK is definitely geared towards gaming. I don't create games but I do enjoy developing custom UIs, so I like the NDK. If you want UI support, the AIR SDK is the better option.

So, BBOS was a diminished experience, but between Android, Flash, and the NDK itself, there's no UI consistency on the platform? Another "is coming Later", like the N.Y. Times PlayBook review said.

Well RIM has come out and said if you want to use stock UI features, the AIR SDK is what you should be using. There is UI consistency in the AIR SDK. The NDK really is more for games. Mark Sohm says stay tuned for the documentation on how to access the NDK from AIR.

Scott

[MSohm]

---

superdirt wrote:

 Mark Sohm says stay tuned for the documentation on how to access the NDK from AIR.

Scott

---

The wait is over.  

Creating Air Native Extensions for BBX

[sblantipodi]

Is there someone considering to support their apps via android player?

[br14]

"Is there someone considering to support their apps via android player?"

Not me. Switching to NDK or WebWorks depending on the app.

From what I've read Android apps will be second class citizens without full API access. I'd love to utilise my 13 years of Java experience, but apart from support I doubt I'll write another line of Java.

Given Adobes announcement today, I don't think RIM could have got their QNX development platforms more wrong if they'd tried.

While I can see the underlying motivation, all that time and energy invested in Flash and Android looks to be the wrong strategy.

To quote Alec Saunders from todays blog post:

"We believe that HTML5 is the future of mobile, and are delighted to be aligned with Adobe on this"

[simon_hain]

I attended the workshop for Air native extensions at DevCon.
While my C-skills are quite rusty it seemed like a very powerful tool to combine native processing with Air UI.
One of the speakers mentioned that most of the upcoming native clients (like mail etc) will be written with this method.

[sblantipodi]

---

br14 wrote:. I'd love to utilise my 13 years of Java experience, but apart from support I doubt I'll write another line of Java.

---

In this case you should write native android apps

[br14]

---

sblantipodi wrote:

In this case you should write native android apps

---

I guess that depends on whether Google reaches a settlement with Oracle.

I read that Oracle wants $15 for every Android device. That's almost $3 billion. Will Samsung and other manufacturers be as happy to produce Android devices by the time all the IP cases are settled? They're already having to pay Microsoft a license fee for Android.

In September Ellison himself sat down with Larry Page for hours trying to resolve the dispute. From what I've read there's little doubt Oracle will win. It's just a question of how much.

Oracle is playing a game of chicken with mobile device manufacturers over Java. RIM have already dropped out.