Google Announces Android, Linux Based Mobile OS

Nope, I was wrong. It's worse than Symbian.

Hey, devs, have fun writing to 50 billion proprietary UIs.

I think Colligan is actually dancing a jig today! Go, Ed, go. (But, yes, you still must RESIGN!)

In the newsgroup there's a hot discussion about lack of support for native C/C++. While google hasn't ruled out such a possibility, the initial SDK only provides support for java applications. Many developers feel that this is a severely limiting factor, both for performance reasons and support of their current code base.

Performance is a complex issue. AFAIK no consumer device has yet been optimized to run a jvm fast: an android device may yield some surprises.
Legacy code on other hand will become less relevant if android gets a large enough market share, which would make porting profitable enough. However in the meantime that could be quite a big issue especially since the platform is a new one.

The sdk is available for windows and requires java to run. It includes a emulator running a browser, contacts and a couple of other applications. An eclipse plugin is included though I did not have the chance to try it out. Not even for an hello world thing, shame on me....

--------------------------
Hey Admin: Why do we have to keep two profiles?

"You keep using that word. I do not think it means what you think it means." (Inigo Montoya)


"twrock is infamous around these parts" (from my profile over at Brighthand due to my negative 62 rep points rating)

I don't believe Google could ship an Android device "right now": Android seems as though it's got a fair distance to go until it's really "ready for prime-time".

Check back March 6th.

We've also got our own applications, of course, including the browser, the full PIM suite, Garnet VM emulator, etc. Even the applications coming out of Google, as part of their own development efforts, are pretty sketchy in comparison...

How do you figure? For starts, Android uses its own (completely non-standard) graphics stack called "Surface Manager", which no one's ever seen previously (and which you won't be able to look at for a year, most likely); you'd need to either take that along in toto (or rewrite Dalvik to tie into the iPhone's presumably Quartz-like interaction stack) at the very least.

Once you've got that taken care of, you'll need to adapt Android's component brokerage architecture (of which only the driver portion's been released as source and would itself need to be rewritten to work with the Darwin kernel); I guess rewriting the entire user space (neither available as source, nor documented) for that would keep you occupied while you're waiting for Google to actually release some of their soi-disant "open source platform"...

That's just two examples of the difficulties such an effort would face. There are certainly others.

See, the bottom line is that there are no Android sources to be "ported", and there won't be for a year, most likely. Beyond that, the overall architecture is completely nonstandard, once you get outside of the Linux kernel.

It's unclear how much access to the graphics and interaction stack will be allowed from the iPhone SDK; it's certain that adapting Android to anything without the underlying sources won't be easy, straightforward or, ultimately, terribly useful.

Nah, ya won't. You'll whine about that, too.

=D

frankengarnet?

i wonder if he's talking about OS's like frankengarnet. a 1996 OS cobbled and hacked together to limp along in the modern world.