Android 2.3/3.0 Releases

Posted: September 3, 2010 in Android
Tags: , , , , ,

There has been some talk about being two Android releases upcoming as in Android 2.3 and Android 3.0. What follows is speculation and best guess.

OpenGL ES 2.0 implementation was slightly goofed up namely the java bindings. The Android 2.3 release cold be a fix for those OpenGL ES 2.0 java bindings. Mario’s OpenGL 2.0 wrapper is here. You can see how the fix was too late for Froyo, in this message.What other fixes could be included is the revamp of the UI.

I think Android 3.0 is different in that it could include the usblib work that has been occurring to develop a  public USB API for the application level. That would probably mean no longer only one module usb ports but true AB USB ports from OEMs for android devices and true USB data synching-out-of-box with no hacking to enable. Other possible items that might be included OpenGL ES 3.0(I meant to say that one could see with heavy use of GPUs and dual core cpus in android devices a need to include the OpenGL 3.0 core since its openGL 2.0 compatible.).

At this point its pure speculation until we get other information, however Google/OHA could do what they did earlier this year in that there would be soft release for 2.2 to fix the openGL ES 2.0 java bindings and than a Android 3.0 release before the end of he year with most of the change sin the Android 3.0 release.

Enhanced by Zemanta
  1. Mike Leahy says:

    Well… I highly doubt an interim 2.3 release is coming out at least in any meaningful way to the larger ecosystem at least. It’s going to be enough trouble getting OEMs/carriers to update to 2.2 in any decent time frame or wide enough device list let alone think they would upgrade to an interim 2.3 release months before 3.0 which I understand may be Q1 ’11 or earlier. What would be more useful is a series of release candidates of 3.0 that can are released early so that those of us that do do cutting edge development can proof the release and we won’t get a repeat of the debacle of the untested OpenGL ES 2.x Java bindings released in Froyo amongst other issues. Of course this begs the question that shouldn’t Google test their own work, but I’d rather a release candidate be available with enough of a window so corrections can be made to the most egregious and broken parts before they become “official”.

    Also Mario offered his OpenGL ES 2.x binding to Google in March and that was plenty of time to compare and find the bugs. His work was just declined with a loosely not created here / we already finished this line of discussion in March.. Well what was finally released in FroYo was badly broken and clearly not even tested and there was still a 2 month window before final official release as the leaked FroYo “RC” after I/O gave away the problem. One could have found the problems easily comparing both bindings back in March.. well I won’t continue on the other fundamental ways to have caught this error let alone say having a test suite of basic demo apps to test against.

    The major problem with the ES Froyo Java binding issue is that it isn’t enough for it to be fixed in future releases. There are far too many devices that may be stuck on FroYo and abandoned by OEMs/carriers. Such that if you use Mario’s or another 3rd party binding there is never a reason to use the official Java bindings again or is not really recommended by me at least. Until FroYo is off the ecosystem map it’s highly inadvisable to use the “official” Java bindings for OpenGL ES.

    Web apps can be additively created with bug fixes and improvements continuously released in a transparent fashion, but low level libraries and OS level SDKs simply need to be extremely well tested before release. So far Google has a lukewarm at best (feels frigid to me) track record with providing stable low level audio & graphic SDK / APIs.

    Also I’d like you to send me the specification for OpenGL ES 3.0 considering it’s not even posted on the Khronos site ( It’s fud to say suggest Android 3.0 will support OpenGL ES 3.x. Of course 3.x is likely being worked on by Khronos and manufacturers and will likely be released in the coming 1-2 years, but I’ve not heard anything substantial at all on any timing or even specification data and generally would think there would be more publicly available info considering the release date of Android 3.0 is relatively coming up in less than 6 months. Also something as major as OpenGL ES 3.x would be announced at SIGGRAPH and such which it wasn’t this year

    GL ES 3.x will likely coincide with a whole new mobile GPU line from PowerVR and the like moving beyond the SGX series, etc, etc. I don’t believe we will be seeing ES 3.x until 2012-13; maybe tail end of ’11.

    I know you labeled your thoughts as pure speculation, but….

  2. Fred Grott says:

    First, the soft release in May for 22 did make it OEMs hence that suggestion..

  3. Mike Leahy says:

    Aww.. Yes.. I see I celebrated the angry birds beta release with angry dev day.. ;P OpenGL & OpenGL ES became slightly divergent, but OpenGL 4.1 on the desktop now is fully compatible (superset) of OpenGL ES 2.x; That doesn’t mean the desktop equivalent will run on mobile GPUs. Quite likely we are going to see OpenGL ES 2.x milked for all it’s worth and in ’11 see multicore GPUs that will scale current GPU tech massively increasing the specs (vertex & fill rate) around the board. Regardless I expect a spec announcement at SIGGRAPH and it could be as late as ’12 announcing OpenGL ES 3.x then likely up to ~6 months after that new devices will appear. IE I would expect it to be in “5th generation” Android devices. 4th gen being the multicore PowerVR SGX GPUs and similar tech from other manufacturers introduced in ’11/’12. We are just at the cusp of the 3rd gen with the Galaxy S being “2.5 gen” (2nd gen CPU / 3rd gen GPU).

    This seems to be the way things will shape up from my perspective considering that desktop OpenGL (4.1) is just now fully compatible with OpenGL ES 2.x I don’t see a major jump for up to 2 years on the ES side as the horsepower must increase via the multicore GPU path before some of the more recent desktop GPU tech can be viable in the mobile space.

    I know you’ve been in this Android game for a little while, so feel bad on stomping on some enthusiasm with any “angstish” thoughts as really at the end of the day I want to see Android succeed, but big G and the OEMs need to up the quality control for low level audio/graphics over the initial formative years (up to now!). It’d also behoove Android/Google management to focus on major point releases and getting as many OEMs/carriers to upgrade existing devices, so putting out an interim 2.3 release would be counter to getting a concerted push to 3.0 in next 6 months. I think more OEMs/carriers would EOL devices on a potential 2.3 release if that became a somewhat wide release target. Of course dev / early RC releases of 3.0 seems the best possible situation. Let the wider community truly help Android out in making sure 3.0 is the best it can be because it’s obvious big G QA is missing the target for low level audio/graphic support (IE always at least one major problem per OS release & yes granted the ES 2.x finagle is Java API oriented), etc.

    Keep up the good work and blogging about Android and such though.. 🙂

  4. Seems that a major goal for Android 3.0 will be more support for Android tablet devices and more grace in the CDD for hardware that would not necessarily be required for the tablet “form factor” (such as rear-facing cam, GPS, or SIM card slots).

    Though Google says Android 2.1 and 2.2 don’t support the tablet form factor, we’ve begun selling cheap Android tablets at for anyone who wants to start messing with a larger UI interface than most phones.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s