Guest post by Google has finally released Android, the opensource platform that will be used by the Open Handset Alliance. The platform is based in the Linux kernel, freetype, sqlite, webkit, a 2D/3D subsystem and other pieces, but the application framework is built in Java using a embedded-optimized VM called Dalvik. The SDK is available for Linux, Mac and Win and it includes an emulator. Video here. Update: The WebKit browser failed to render the desktop version of OSAlert, so now we feed it our mobile one.
Let me be the first to say: “Great news!”
From just playing with it right now, it looks like this might make a nice phone OS once it’s polished up and put on a real phone.
As for the development side, it looks like current JavaME developers should make the switch easily, but once they get in farther, it offers a richer, more smart phone like development experience since there’s more than just JavaME to play with. I could definitely see good things for this, if actual phones are ever released.
The browser appears to be webkit based, though it could still use some work
After playing like 45 min with it, and as a full time java developper I find it very promising and I’m quite confident it’ll be very successful.
It’s an amalgam of open-source technologies, with a basic Java API on top of it: it’s Hibernate/Struts for phones, except not as advanced as either. And a quick look would seem to indicate that the XML schema they’ve settled on is a bit ugly, e.g.
<TextView id=”@+id/text1″ xmlns:android=”http://schemas.android.com/apk/res/android“
android:layout_width=”wrap_content”
android:layout_height=”wrap_content”/>
It depends how well it’s dressed up, but it’s a disappointing effort for the amount of time they’ve ostensibly spent developing this software stack. The absence of an equivalent to iPhone’s LayerKit is particularly telling – Unix desktops are already using high-level abstractions like Cairo and Arthur and others and Apple has CoreImage, and LayerKit on the Mac: OpenGL and SDL simply doesn’t cut it.
It does look like they’ve invested time in performance, but the usefulness of that is open to question given how fast mobile hardware is developing (even bog-standard Nokias come with some basic 3D acceleration these days). Frankly, I can’t see many people flocking to use it, which is why loads of companies have “joined” the alliance, but none have made any public commitment to use the software in a future product.
I think you underestimate one crucial point : the ease of development. Currently we have three choices:
– Windows Mobile: can’t comment, my religion forbids it.
– J2ME (what I do everyday): Runs on pretty much every device but it just has too many limitations in every possible place. Even with a signed application there just is no way in hell you can make things easy for the end-user. Trust me, we try.
– Symbian: once you’ve read how to do a string comparison in Symbian, you know why you don’t go further unless forced.
Android (as it seems to me) is bringing fast development on the mobile. It is finally a platform for which you can program using a welcoming language, java, while still being able to reach most of the phone’s features (eg. acting upon a phone call).
Well, maybe I am wrong, but I for one, after having watched thre three introduction videos, am pretty excited.
Edited 2007-11-12 23:04
…
basically, you adore Apple and IPhone and you’re scared something else would pop out and be cool at the same time.
Face it, eclipse plugins (with RAD wrappers for development) around XML can pop out any time and don’t even need to be developed by Google. All that google needs is proof of concept.
It does look like they’ve invested time in performance, but the usefulness of that is open to question given how fast mobile hardware is developing (even bog-standard Nokias come with some basic 3D acceleration these days). Frankly, I can’t see many people flocking to use it, which is why loads of companies have “joined” the alliance, but none have made any public commitment to use the software in a future product.
This is open platform (or at least how I understand it, but I’m not so in mobile development). So, it can evolve. But evolving from open platform is much easier than evolving from closed one.
Lets take your favorite IPhone, closed hardware with closed platform. All evolving goes to Apple. Its not better than Symbian, heh even Symbian runs on different hardware,
I don’t adore the iPhone, I simply believe it has set the standard for user-interface design in the next generation of phones. This hasn’t addressed that, if anything, it’s a replacement for the current generation of Blackberry / SmartPhone software.
Nor did I say it was pointless or useless, I said it was disappointing and uninspiring. Creating a struts-style library isn’t particularly difficult, nor is creating a limited number of data-abstraction layers.
As a result, while it evidently simplifies mobile development, I don’t think it has the necessary hook to really grab mobile-phone makers attention, and so is unlikely to succeed on a large scale. Look at it this way, Google almost certainly contacted Nokia about this platform several months ago, but Nokia have chosen to use Symbian for the iPhone clone, and their own software stack for their own phones.
I don’t adore the iPhone, I simply believe it has set the standard for user-interface design in the next generation of phones. This hasn’t addressed that, if anything, it’s a replacement for the current generation of Blackberry / SmartPhone software.
If you take like that (replacement for blackberry), then I agree.
Nor did I say it was pointless or useless, I said it was disappointing and uninspiring. Creating a struts-style library isn’t particularly difficult, nor is creating a limited number of data-abstraction layers.
You’re wrong and right here. Creating library is simple, yes. Having it politically, religiously and API wise correct is something completely different, it is as hard as it gets. By watching those videos I got the impression that was their main goal.
And as every open platform, it is burden of community to drive it forward with extensions (development and feature wise).
btw, I can’t really understand what you expected. It is OS template that is providing background services to be integrated in phones. It is not compiz where you can easily make video that impresses masses. Did you expect to see some bling, gods revelation or what?
As a result, while it evidently simplifies mobile development, I don’t think it has the necessary hook to really grab mobile-phone makers attention, and so is unlikely to succeed on a large scale. Look at it this way, Google almost certainly contacted Nokia about this platform several months ago, but Nokia have chosen to use Symbian for the iPhone clone, and their own software stack for their own phones.
Again wrong here. Google has THE name. And their brand is the one that will push this to vendors.
If Nokia would do as you suggest it should have already done, Nokia would not exist anymore (sorry, but corporations without brains cease to exist). You can’t base your present on something that doesn’t exist yet. But you can base your future on that (that is, if that something already exists and has proved it self).
Main feature that will drive Android is Google it self. If we look at GMail. It is a f***ing mail. The fact that people know its Google mail is driving it, not the features.
thumbs up here. i guess they eliminted some deficencies of the java platform with a fixed OS-library-DB layer. and created a rather nice API for making applications fast. this is rather now, develope everywhere with multiple IDE’s ,debug once system. sun should have come up with this idea long time ago..
hoping that there will be an intellij idea plug-in, i will be more comfortable with it.
in some years, this may as well be a death blow to j2me tough. But hey, thank God it is Java.
sun should have come up with this idea long time ago
Well Java is already everywhere on phones, so I guess Sun did do their job.
well OS-hardware layer was always the weak point when it comes to the implementation. in android, it is a full stack, much easier to control and enforce standards for the API.
Hey, hasn’t j2me received its “semi-official” death blow yet? I kind of remember reading something from Jonathan, that it has become superfluous due to the phones getting smarter and smarter. I still believe that the desktop JRE is overkill for a phone, though.
So far this looks very nice. The emulator is really nice and debugging the device seems to run as expected.
I am not sure that its wise to do a whole new platform, and then not support javame (the android.jar does contain microedition classes though).
I’d like a device please
As a J2ME developer owning a mac I can say I am glad to see google take this multiplatform approach while most phone vendors only released their SDKs as windows only binaries (nokia has some linux ones, but they were very limited).
So now we’re having OpenMoko, Qtopia and Android, all pushing for (more) open phones, and each with quite some weight behind them.
I’d say the future for more open phones is looking good! Combined with mobile internet prices steadily dropping, this introduces exciting possibilities.
I can’t wait to use a SDK for a proposed device that will never actually make it into consumer hands because of lack of manufacturer interest.
lol YOU ARE WRONG
Oh, right. If you say so… pfffffffffffft….
Q: Consumers won’t see devices until next year, right?
Rubin: That’s right.
http://www.news.com/From-Dangers-realm-come-Androids-makers/2008-10…
Translation: We have hardware partners who are experimenting with the software, but we don’t have commitment from any carriers (AT&T/Cingular, T-Mobile, Verizon, etc) yet, though.
Motorola doesn’t make contracts with consumers. The carriers do. Until the carriers sign up, it’s vaporware.
Have you actually watched the videos on youtube about the project. It has alot of interest from T-Mobile, and other large players. I have no doubt this will be available
Clue phone: T-Mobile and others are not going to allow an open platform on their closed network infrastructure. Ain’t gonna happen. It would completely cut them out of the services market, and that’s how they make a tremendous amount of cash. Sorry, but they’re not gonna hand you the keys to their metaphorical car.
You are right that they don’t want it to happen, you are wrong that they can prevent it forever. It is happening whether they let it or not, so apparently they are working with the good guys on this rather than against them.
Google is basically strong-arming freedom into the handheld internet device world, and we must love them for it. All they want in return is more people using google.com.
Google is only strong-arming freedom for a device manufacturer to produce their specs to run on a future network of Google’s ownership.
Telcos will continue to lock down phones to their demands so long as cell phone developers don’t have high volume alternatives to sell to who won’t lock them down.
I’ve happy with any technology or innovation that shakes up the telco companies a bit. There’s no other industry that’s so tightly regulated and restrained. With Palm struggling it’s heartening to see another platform born that might fill the same shoes.
Hey Ballmer. What was all the crap about a product that didnt exist except on paper? Well the SDK is out at least. We’ll see whose laughing in a couple years….
this just seems like another java-on-mobiles implementation to me, ok so its not winCE underneath, but its hardly the first linux mobile either.
how do we get a console up – can we make kernel calls, can we program in python or shell, do we have dmesg, a tcp stack that operates outside of this java gui? otherwise i’m afraid its of no interest to me.
and to someone above, wasn’t qtopia (green phone) recently cancelled by trolltech? so we’re back to openmoko vs. android, of which i prefer the former (open hardware too).
it doesn’t seem like it has no interrest for you already. if it had, you would have already searched for the information and found it…
Qtopia was not cancelled, the greenphone simplystopped production. According to TrollTech the greenphone accomplished its job of getting qtopia out and in the hands of developers. Their intended product (the qtopia platform) is still actively supported and developed.
And while you may prefer the OpenMoko platform, the threshold for development on the Android platform is lower. Further the Android platforms API more suited to writing apps targeting the convergence device market IMO. Finally the amount of resources Google alone is putting on the Android platform pauper the resources working on OpenMoko. This does not even count non-Google resources.
Edited 2007-11-13 14:22
They claim it’s an open source telefon platform. Where is the source for the GSM/WCDMA stack? Giving out a API to use the phone is not open source and to use an OS under that is open source doesn’t make it open source either.
Nothing to see here, move along.
Bye,
There are many people interested in programming for the platform who have no interest in the actual coding of the GSM/WCDMA stack.
It remains to be seen whether android will be successful, or even attractive to serious developers. However giving out an API under an Open Source license is open source. Similarly building upon an open source OS means the underlying OS is open source. The fact of whether or not the radio stack is open source or not makes no difference to the two previous facts.