Good news!

A *real* version of skype has been released for Android. The only qualm is that it restricts calls in the US to WiFi (carrier restriction, no doubt):
http://www.engadget.com/2010/10/05/s...ill-wifi-only/

... but that's not all! It seems that Fring has been released as well with FringOut:
http://www.engadget.com/2010/10/05/f...android-calls/

Android VoIP just got a lot more compelling for the everyman literally overnight.

In unrelated and off-topic news... OnLive has ditched their monthly fees!
http://blog.onlive.com/2010/10/04/on...play-for-free/

Now, if *only* there was an Android client, then we could be playing Halo at ultra-settings on our Tabs! However, there are rumors that this will exist:
http://onlivefans.com/showthread.php...id-Development

Even if Onlive was limited to WiFi (with the huge amount of data transferred for this service, WiFi would be ideal anyway), it would be a winner on the Tab. The only thing left to solve is the Keyboard and mouse issue. Android may require a BT mouse driver for game compatibility.

Here's a nice HQ video showing Flash running in the browser. In fact it shows the open Papervision3D flash API for 3D in flash, and it runs quite well on the Tab!

The video also shows Angry Birds running, and it seems to run perfectly smoothly on the device.

In my direct experience flash made the stock browser perform real slow (online newspaper pages with several flash advertisement boxes). So much I had to deactivate it.

I've noticed this in videos as well. When there are a lot of flash ads on the screen, scrolling/zooming becomes very choppy. Even with a single element on-screen the effects are noticeable. It's interesting because there's a clear distinction between the browser redraw and the redraw of the flash areas on screen.

I don't think that a device with hardware like the Tab is ready for multiple flash windows just yet, but I do believe that the browser is mainly at fault for this. With some much needed optimizations, Google's browser could perform quite well despite the heavy and slow flash.

Here are my thoughts.

--------

I think performance of flash in the browser could be greatly improved with a few changes. Of course, these are based on speculation of the underlying browser architecture, so don't take too seriously:

1) GPU accelerated scrolling in the browser. This can be accomplished a number of ways, I'm sure (eg. mapping screen elements onto quads -- should work well with site transparencies)

The iPad and iPhone feature near flawless scrolling which many speculate to be because of this feature. Additionally, there is a clear separation between movement and redraw. The CPU in this instance would only be required to draw new portions of the screen and only after scrolling has taken place. What you get is an ultra-smooth window scroll, and a checkerboard board pattern that quickly disappears when the CPU has time to redraw. Similarly when zooming, it's clear that a scaling of the current window contents happens and the the CPU re-draws at the new zoom level when it is able. Again, the result is a very smooth operation.

Also less CPU activity may positively benefit battery life.

FYI anybody (with a google account) can vote for this issue here.

2) The flash window should give its content to the browser to render, and not be responsible for rendering to the screen directly.

I'm speculating that this is indeed the case, but when I see flash elements scrolling at different rates than browser scrolling, it seems as though this may be what's happening. With a single point of rendering, this would not occur, and flash would not need to be called when a scroll event fires -- an almost certainly expensive operation, especially with multiple objects on scree. It also leaves the door open

3) De-prioritization of flash in the OS scheduler. I don't know if this is the case, but flash should be given a very low priority compared to the host browser. In other words, even for a flash heavy site, flash should slow down before the browser does.

-----

Certainly the latest Chrome Dev includes GPU acceleration which the chrome team has mentioned has many speed benefits:
http://www.tomsguide.com/us/GPU-Acce...news-7895.html

I notice that even the iPad flash implementation features FLAWLESS scrolling, which leads me to believe that at least some (if not all) of these optimizations have been implemented.

I think that the Android browser is leagues behind where it should be. Though processors are becoming faster and faster, I have yet to see an Android superphone scroll a web page as proficiently as a 1st generation iPhone/iPod touch despite the technically superior hardware. And android *is* capable of it. If it can display a complex game like Need For Speed Shift, or Nova, at smooth rates, it can handle smooth scrolling. Of course the code is the bottleneck.

It remains to be seen weather or not Gingerbread will address these concerns. Here's hoping that it does, as a browser is a MAJOR application for mobile devices and arguably one of Android's largest failings to date.

Here's a lower price in the UK for the Tab £551.08 including VAT:
http://www.moresamsung.co.uk/extra.a...0CWAXEU&nmaa=1

Released in November. 750 EUR for 32 GB version. 850 EUR for 64 GB version. Source is nu.nl (Dutch) who have it from Samsung.

Will contain Android 2.2 (Froyo) and Flash 10.1. It is more expensive than an iPad. TBH, a bit too much for me for a large phone with a large screen

I can appreciate this perspective. However, loading a full linux distro on this surely must be enticing! Ubuntu Unity.. Yum..

I just wish it had the specs of the PlayBook! However, I'd be happy with a better performing browser, and I'm not about to jump onto a closed distro (QNX) posix-compliance or not.

The next round of Samsung Mobile Galaxy Tab ads are up!

http://www.youtube.com/watch?v=t41sicKJ8mk
http://www.youtube.com/watch?v=JeknEnrlrYc
http://www.youtube.com/watch?v=TomP7MDhurY

I don't know if this has been posted, but here's a very thorough tour of the Galaxy Tab:
http://www.youtube.com/watch?v=QH2QPr-D_JU&feature=fvwk