HaXe, my new toy!

After Adobe MAX 2011 everything should not be the same for me and maybe for a lot of flash platform developers around the world, Adobe brings some “directions”  that didn’t find my consent mainly for the way that communicate these news and the impact that had in the market, but we know that Flash Platform is not dead and it will go ahead for many years.
Obviously nothing was the same after that, in fact many developers started to look around for new technologies and frameworks like Backbone.js, Sencha Touch, Ext JS and so on.
Personally I started to checked in last few months many Javascript frameworks because my aim was find something that could replace Flash Platform in the future and I have to spend time in next years to consolidate it and go ahead with Flash Platform too.
Last week a big friend of mine gave me this link: http://www.haxenme.org/ and when I started to read what you can do and how you can do it, I immediately started to go in deep with HaXe in my spare time and trust me that I had a lot of fun!

First of all what is HaXe?
HaXe is an open source multiplatform programming language, it allows to write once and deploy everywhere (in the right meaning of therm “everywhere”).
In fact with HaXe we can write in a programming language similar to Actionscript 3 (strictly typed, OOP, …) but more powerful (it has enum, generics, dynamic type, …), with HaXe we can target our projects for Flash, C++, Neko, HTML5, Node.JS, PHP, iOS, Android… if we work with multiplatform APIs we can write once and deploy our project for multiple targets.
So for many developers that come from Javascript, Actionscript, Java and so on, will be so easy to start deal with HaXe.
Another interesting thing of HaXe is that we can work with the library present in the SWF files and integrate movieclip in our project, we can create also SWF file without Flash Professional with SWFMill that is used for the generation of asset libraries containing images (PNG and JPEG), fonts (TTF) or other SWF movies.
That’s so interesting because it means that designers that usually prepare assets for developers don’t need to change own daily workflow!
If you need to extend your target platform we can add new features with external libraries, it’s so important because we can really cover everything with this feature; we can find a lot of ready to use libraries directly on the lib HaXe website.
With HaXe you can communicate between different languages like JS and Flash in both direction, you can easily find many frameworks and library porting in HaXe, for example javascript like JQuery, Sencha Touch, Node.JS and so on.

What about the IDE to work with HaXe (so important for a developer!)!?
On Mac you can use TextMate or FDT on Win FDT or FlashDevelop this one seems the best one but I didn’t try it. For more specs I suggest to take a look at HaXe site section, maybe you can find your favorite IDE in the list.

Finally I made an easy sample to understand better the powerful of HaXe NME, this sample loads an external XML file and an external SWF library with a movieclip inside exported for Actionscript, so I added a drag&drop feature to the list. Then I tried to compile it for iOS, Mac OS X Lion, C++ and SWF with the same basecode and everything work so well and smooth!


You can download source files here, to compile it take a look at HaXeNME section and you can find everything you need to try this sample and start to play with HaXeNME!

If you want to deal with HaXe, I suggest two books, the first one is really a good start to work with this fantastic language:
HaXe 2 beginner’s guide
Professional HaXe and Neko

Last but not least, next April in Paris there will be World Wide HaXe conference, I’ll be there to learn more about the future of this amazing platform if you are planning to be there it will be a pleasure for me catch up for a beer!

I hope soon to publish more experiments and informations about HaXe because it is a thrilling programming language!!!
So stay tuned!

InDesign and Flash: PageFlip controller

One of our last project was to create an easy but integrated workflow from InDesign to web.
In particular they request us to create a PageFlip that will be published automatically online every change they made in the offline version without spending a lot of time on this activity.
Like you can see for a Flash Platform developer nothing more obvious that a project like this, so you can take a PageFlip online, you export all pages in swf or jpg files and you popolate the XML behind the PageFlip… yes but we don’t bring this direction.
First of all we suggest to our client to add interaction elements directly in InDesign because from CS5 version (if I remember well) it adds the animator engine of Flash, so it is quite easy make animations directly in InDesign for a graphic designer that is more comfortable with this software instead of Flash Professional.
After that we study the InDesign exporting settings and we find that you can have your PageFlip exporting InDesign document directly in unique SWF file; the only problem is that you haven’t any control panel to add interaction in your PageFlip, for example if my catalogue is composed by 300 pages and I have to see the 250, from the beginning I have to click 249 times on the right page to see my page.
Another interesting thing that we see during this development is that there are many PageFlip on the web but it’s not easy for a graphic designer, more focused on the paper, to customize them because you have to know how Flash works and in detail how the PageFlip, that you chosen, works too.
Also the PageFlip made by InDesign has great performance instead of any other PageFlip tested with many pages in particular if pages are vector based.
Finally we decide to decompile the InDesign PageFlip to analyze if we will able to create a control panel for this SWF file and… WOW… we really bring an interesting direction!


In fact, in the document class of the SWF file called IDSWFFile, you can find many methods that could help you to create, for example, a navigation panel to jump from a page to another one or to create an index to navigate trough chapters.
The main thing that you have to remember is that InDesign PageFlip works putting each page in a frame so if you have 20 pages you’ll have 20 frames inside the SWF file generated.

Below you can find some useful methods that could help you to develop your personal control panel:

  • getFrameCount() return the number of frames (so the number of pages inside the SWF file)
  • getCurrentFrame() return the actual frame (so the page that user is reading)
  • getThumbnailForFrame(frame:int, width=32, height=32) return a bitmapdata of a frame in the size that you prefer, the default values are 32×32 px
  • goToFirstFrame() goes to the first page of the PageFlip
  • goToLastFrame() goes to the last page of the PageFlip
  • goToPreviousFrame() goes to the previous page of the PageFlip
  • goToNextFrame() goes to the next page of the PageFlip
  • goToFrame(frame:int) goes to a particular page
  • stopAllAnimations()
  • stopAllSounds()
  • stopAllVideos() 

There are also other public methods but for me those are the most interesting to create a PageFlip controller.
To work with those methods you have to cast the content of Loader like a generic Object and then you can call all those methods like you can see here:

// here we create the generic object to call PageFlip methods
private var _pf:Object; 
// we load our file generated by InDesign
_loader = new Loader();
_loader.contentLoaderInfo.addEventListener(Event.COMPLETE, setPosition);
_loader.load(new URLRequest("pageflip.swf"));
//when PageFlip is loaded I add on the DisplayList
private function setPosition(e:Event):void{
_contPF = new Sprite();
_contPF.addChild(_loader.content);
addChild(_contPF);
_pf = _loader.content; 
}

// below I make functions to navigate the PageFlip when user click a button
private function prevPage(e:MouseEvent):void{
 _pf.goToPreviousFrame();
 }
 private function nextPage(e:MouseEvent):void{
 _pf.goToNextFrame();
}
 private function lastPage(e:MouseEvent):void{
_pf.goToLastFrame();
}
private function firstPage(e:MouseEvent):void{
_pf.goToFirstFrame();
}

If you’d like to create a more automatic workflow that allow your user to focus only on the content of PageFlip without waste his time with exporting issues.
In our case we use InDesign Server to solve this problem and create some scripts that allow user to create thumbnails and zoom pages without spend a minute on Photoshop or InDesign but easily upload own InDesign file on a server and via scripts we make everything user needs.
But I know that not everybody could have InDesign server in house, so another solution that I’d like to suggest could be create an InDesign panel with Flex that make the dirty job for the user, preparing all images for thumbs and exporting the PageFlip with right settings.

I know very well that it’s not rocket science for a Flash developer but I think that the workflow behind could be interesting and could be helpful in many situations and also I guess that InDesign users could find good stuff in this post.
Finally we have to remember that technology has to help people in their daily work accelerating process and maybe substitute the human interaction with a computer interaction giving more time to what people should be better like think to new stuff.

Introducing Starling: book review

Hi All,

first of all I apologize with people that usually read this blog if I didn’t insert any new post since last year but I’m working a lots in these few months to open a new market opporunity for my company out of Italy and I’m totally absorbed in this new activity, but in the meanwhile I’m studying during my spare time and I’d like to share with you my thoughts about “Introducing Starling”.

This book is for any Flash Platform developer that is looking to create next generation of mobile and desktop apps (or games).
It’s a book so practical that introduce you to the Starling framework, explaining how it works with simple examples of code that you can put in practice in a while.
Thibault guides you showing each object presents in this framework, that is an abstraction of Stage3D API introduced with Flash Player 11 and AIR 3.
With Starling you can aim better performance in your 2D applications thankfully the GPU acceleration added on Stage3D, with this book you can discover what there is behind and starting to develop with it.

Flash Platform Galaxy: why choose Flash Platform

In those days I’m reading a lots of mailing lists, forums, blogs and so on where Flash Platform supporters are so disappointed about the latest marketing movement of Adobe.

In fact yesterday Adobe announced that they stop the development of Flash Player on Mobile devices (on Desktop they are going ahead).
The road is clear HTML 5 inside the browser and Flash Platform for RIAs, Games and out of browser in combination with Adobe AIR.

For me Adobe for the second time (the first one was during Adobe MAX) has totally mistaken how to communicate this news and obviously tech blogs bring this announcement like the end of Flash…
Personally I don’t think that is the end of Flash but I think that Flash is moving on a new position in multimedia world probably out of browser.
I’m an Adobe addicted, like you know, and in particular I’m a Flash Platform supporter, so I think that we have to move on and make something to spread the word about this foggy situation, guys, Flash Platform is ALIVE!
To do this, I start making a pdf file called Flash Platform Galaxy that could help people to have an idea of what is Flash Platform and why choose it ( I know, I’m not a graphic designer but I think it could be useful), if you want to add more informations or change something feel free to leave a comment at this post or drop me a line via email.
Let’s go guys, we have a platform to save ;)

It’s time for a new web design era… probably without Flash

I’m here in L.A. for Adobe MAX, before the beginning of the conference we were very excited about the 2 keynotes because we were waiting for amazing news about our favorite platform… but it didn’t happen, instead we find 2 days very focused to designers (and it’s normal I guess) and to HTML 5, JQuery and CSS 3.

This is an important signal from Adobe, in 2 keynotes they mentioned Flex 2/3 times maybe, all the new stuff for Flash was focused on games and 3D and AIR had only 5 mins to show the new release and its new features.

So the message from Adobe is quite clear: “Use HTML 5 for the web and Flash for the game and out of browser”.
This recommendation from Adobe would change online web design classes significantly in the future.

It seems crazy but it’s the truth… the same company that in the beginning of 2000 was scared from the little Macromedia and its best technology Flash, today substitued the player with a new one.

In latest years I took part of different Flash Platform projects, mainly desktop and mobile applications and I hoped to embrace the right direction with the Flash Platform, with this MAX I’m pretty sure  that I made the right decision.
I’m disappointed that my favorite platform will be out of the browser but I’m more disappointed that Adobe is following the market instead of make it.

Today I can say that it’s time to move on and start again to study new “trendy” technology instead of a good and solid technology like Flash Platform.
Could Flash be the new mobile and desktop technology? I really don’t know but I’m sure that this year a new milestone of the web was put by Adobe.

Playing with Google+ API and Actionscript 3

Yesterday I saw that Google has released Google+ API, so I started to played with in my favorite part of the day: the night!

For now Google has published only public APIs that allow you to retrieve user’s informations and his activities list, all API are made in RESTful and JSON that are so easy to add in your project.
When Google will release next APIs, I hope soon, you can consume them with an oAuth 2.0 authentication, like Facebook.
I made a simple example to retrieve data from my Google+ profile in Actionscript for Android, you can download the source files directly from here.

If you want to play with Google+ API, first of all you have to request your Google+ API key; to do that go to Google API console and activate Google+ service, then in the details page you can find your key.

Google+ RESTful services get us a JSON response, to read it in your Flash/Flex application remember that you need as3corelib that have the JSON deserializer, for example you can push in a generic Object all data retrieved from Google+ service or if you prefer you could create a Value Object that it could be more useful.
In this case I decode directly in a generic Object, like you can see in this code snippet:

var data:Object= JSON.decode(dataToRead, true);

then now you can easily access to informations:

var icon:String = data.image.url;
var name:String = data.displayName;
var tagline:String = data.tagline;
var description:String = data.aboutMe;

Another easy feature that you can add in your Google+ application is a static image of the map, like in your web page, using Google Maps Static API.
You can easily make a query to Google Maps passing in GET params like: the dimension of image, the location and the zoom; you can also add more params that you find in the docs of Google Maps Static API.
Here a code sample to add this feature in your Flash application, in those line I request for an image with width 480px, height 200px, with roadmap skin and with a zoom of 15x:

var mapLoader:Loader = new Loader();
//city is a variable with the name of the place that you have to retrieve.
mapLoader.load(new URLRequest("http://maps.googleapis.com/maps/api/staticmap?center="
+city+"&zoom=15&size=480x200&maptype=roadmap&sensor=true"));
mapLoader.y = this.stage.stageHeight - 200;
addChild(mapLoader);

Like you can see work with those APIs are pretty easy, so now we have to wait for final release and then we can start to create our Google+ integration with the Flash Platform.

Tricks for tween on mobile devices with Flash Platform

In this quick post I’d like to share with you my experience about Tween on Flash Platform projects delivered on a mobile devices.
I started work on mobile since Flash Lite 1.1 so I grew up with mobile, I lived all the Flash mobile evolution and now, on tablet and smartphone, I had some good tricks to share with you, I really hope that those tips could help you during your developer life.
OK, let’s start:

  • Use quality property of stage
    This is a really good technique to use when you have to improve performance of your project, when you need to make a fluid tween before launch it, set stage quality to low and when tween will finish set stage quality to high or best.
    Avoid to use this technique when you have vectors (textfield for example) on the stage because you could have a worst result.
  • Use cacheAsBitmap and cacheAsBitmapMatrix
    If you have vector object that you’d like to animate in your project remember to cache them and then animate; remember also to set your application with GPU acceleration and you can see a really good performance with this technique.
    Avoid to cache objects that you need to remove from display list, it will be so expensive for your memory.
  • System.gc() works!
    I tried in few sample to use it on Android and I saw a good result, so the “old” tip to call System.gc() twice in a try/catch statement works on mobile device too (only on AIR apps)
  • Take care with multiple animations on iPad and iPhone
    On iOS devices we don’t have AIR runtime so LLVM translate our AIR project for us in Native Binary so it could help if you move few objects per time in particular if you have big objects to move like a background or something like that
  • Greensocks tweens are the best
    I tried tweener and other tween libraries for AS3, but the best one for me are the Greensock tween library, in particular on iOS devices.
  • Last but not least, remember to test your animation on the device because you could see “funny” results
    Sometimes happen that on your computer everything works well but when you port your content on a tablet or smartphone everything works not so well.
    Before hurt your head, remember to test a lots of times your mobile content on the device, it could save your projects!
Finally I suggest to take a look at dev center mobile development zone of Adobe site because you can find many tutorials and helpful tips on mobile development.
That’s all folks for now, I hope you enjoy those tips.

Multicast UDP socket in Adobe AIR with Python

In my last project I solved a big issue that I needed to receive notification from an hardware via a Multicast UDP socket.
If you want to know more about Multicast and Unicast in Flash Platform I really suggest to read Flashrealtime blog.
Like you know the Flash Platform can use multicast with RTMFP protocol so you can use it Flash to Flash or Flash to Flash Media Server but not Flash to hardware for example.

In my case I needed a Multicast UDP socket in a local application that communicate with an hardware that was my server, so I thought to realize a Python application for Mac OS X and Windows that could help me to solve this problem.
Python is my second favorite programming language after ActionScript (obviously) and this time Python saves me to accomplish my project.
I want to start with Python side, my goal is to connect to a multicast socket and get all data and then send them to a unicast UDP socket server that we will create in Adobe AIR in next example.

#!/usr/bin/env python
# encoding: utf-8

import socket, select

ANY = '0.0.0.0'
MCAST_ADDR = '239.0.1.1'
MCAST_PORT = 10300

sock = socket.socket(socket.AF_INET, socket.SOCK_DGRAM, socket.IPPROTO_UDP)
sock.setsockopt(socket.SOL_SOCKET,socket.SO_REUSEADDR,1)
sock.bind((ANY,MCAST_PORT))
sock.setsockopt(socket.IPPROTO_IP, socket.IP_MULTICAST_TTL, 255)
status = sock.setsockopt(socket.IPPROTO_IP, socket.IP_ADD_MEMBERSHIP,
socket.inet_aton(MCAST_ADDR) + socket.inet_aton(ANY));

sock.setblocking(0)
print 'socket UDP multicast ready'

udp = socket.socket(socket.AF_INET, socket.SOCK_DGRAM);

while 1:
	ins, outs, errs = select.select( [ sock ] , [], [], 1.0)
	try:
	    data, addr = sock.recvfrom(1024)
	except socket.error, e:
	    pass
	else:
	    print data
	    udp.sendto(data, (socket.gethostbyname('127.0.0.1'), 10303))

socket.close()

First few lines we define the multicast socket server ip (that in my case is pointing to the hardware) and the socket port to communicate with.
Then I create the UDP multicast socket and the UDP unicast socket (udp var); in the while loop I set the socket timeout and if I receive data trough the multicast socket I send everything to unicast one.
After that to create an application without any dependency in Windows or Mac OS X, you need to use a couple of Python libraries called py2app and py2exe.
Both allow you to create an executable file from your Python script for mac or win without any dependencies, in Windows side you have only to remember which kind of Python dll you have to incapsulate in your AIR application but we take a look at that in a while.
To create the executable file you have to create a setup script in Python for both operating systems, I suggest to create something like that:

FOR WINDOWS:

from distutils.core import setup
import py2exe,sys,os

origIsSystemDLL = py2exe.build_exe.isSystemDLL
def isSystemDLL(pathname):
        if os.path.basename(pathname).lower() in ("msvcp71.dll", "dwmapi.dll"):
                return 0
        return origIsSystemDLL(pathname)
py2exe.build_exe.isSystemDLL = isSystemDLL

setup(windows=['myPythonScript.py'])

FOR MAC:

from setuptools import setup

APP = ['myPythonScript.py']
DATA_FILES = []
OPTIONS = {'argv_emulation': True}

setup(
    app=APP,
    data_files=DATA_FILES,
    options={'py2app': OPTIONS},
    setup_requires=['py2app'],
)

Then if you want to create your executable files you have only need to launch the command on command prompt or terminal (you can find more informations on how to customize those setup scripts in each library’s website):
FOR WINDOWS:

python setup.py py2exe

FOR MAC:

python setup.py py2app

Now we can start with the AIR part, like you know with AIR 2 you can work with Native Process and this is the case to use them:

const WIN_PATH:String = "win/socket.exe";
const OSX_PATH:String = "socket.app/Contents/MacOS/socket";

var nativep:NativeProcessStartupInfo = new NativeProcessStartupInfo();

var finalPath:String;

var f:File
if(Capabilities.os.substr(0, 3) == "Win"){
   finalPath = WIN_PATH
}else{
   finalPath = OSX_PATH
}

f = File.applicationDirectory.resolvePath(finalPath);

nativep.executable = f;

var process:NativeProcess = new NativeProcess();

process.addEventListener(ProgressEvent.STANDARD_OUTPUT_DATA, onOutputData);
process.addEventListener(NativeProcessExitEvent.EXIT, onExit);
process.addEventListener(IOErrorEvent.STANDARD_OUTPUT_IO_ERROR, onIOError);
process.addEventListener(IOErrorEvent.STANDARD_ERROR_IO_ERROR, onIOError);

process.start(nativep);

var udp:DatagramSocket = new DatagramSocket();
udp.addEventListener(DatagramSocketDataEvent.DATA, getData);
udp.bind(10303);
udp.receive();

function getData(e:DatagramSocketDataEvent):void{
	trace(">>>>>" + e.data.readUTFBytes( e.data.bytesAvailable ))	
}

function onOutputData(event:ProgressEvent):void{
   trace(process.standardOutput.readUTFBytes(process.standardOutput.bytesAvailable)); 
}

function onErrorData(event:ProgressEvent):void{
    trace(process.standardError.readUTFBytes(process.standardError.bytesAvailable)); 
}

function onExit(event:NativeProcessExitEvent):void{
   trace(event.exitCode);
}

function onIOError(event:IOErrorEvent):void{
  trace(event.toString());
}

In this simple script I’m choosing the right executable file for the operating system where my AIR application is working on and then I launch the native process to start the multicast socket.
Finally I listen for the unicast socket and I trace on the output panel the messages that I receive from the hardware.
I think this is an interesting way to extend Adobe AIR with Python that open new possibilities on the desktop side, think for example to create a Python bluetooth extension for Adobe AIR, it could be so interesting add this feature to AIR apps isn’t it?

Presentation Model design pattern: multiple screen solution – part 1

Today I’d like to talk about Presentation Model design pattern because it’ll be so useful for anyone that is working on multiple screen project with Flash or Flex.
In latest months we lived in phrenetic mobile world where we have tons of new and powerful devices on the market with different screen sizes, different hardware and so on, on the other hand we have clients that are looking for new business opportunities and they are seeking a way to spend less but earn more.
So, one way to solve those problems could be find a way to deploy with the same technology in different devices (tablets, smartphones, computers…) saving time with great results!
That’s not a dream, it could be made with knowledge, architecture and a good developer or a team!

Our goal

For a developer find a way to have a good project architecture to maintain it’s so important, usually when you start a project you try to define the parts that could be re-usable in different part of the same project or in different project too.
So our aim is find a way to write less code but it will cover the project needs and that could be portable in different screens / operating systems easily.

What is Presentation Model design pattern?

My mentor Martin Fowler describes it with those words: “Represent the state and behavior of the presentation independently of the GUI controls used in the interface”  (here the full article of Martin Fowler).
In fact with this pattern we divide UI (components, movieclips, sprites…) from their behaviors; each view will have one and only that one presentation model and only the presentation model will interact with the whole architecture (like with model or proxy…).
Probably with this image you can easily understand better this concept (from Martin Fowler website):

Like you can see we have 3 class, AlbumTitle that is a view with a textfield, AlbumPresentationModel that is the presentation model of AlbumTitle and it has the copy of the view but storing datas inside and finally the main model Album where we have the data that could be used for the whole application.
There aren’t any connection trough the views and the application model because only the presentation model has access to the whole project, so the view is only a bunch of components and custom graphics, this it means that if you have to change UI on a mobile project for a different device, changing the views and anything else, you will have done your job.
In fact with this easy technique you perfectly solve our problem and you should create the same content for different devices changing only the views.
Probably your application in different devices will have the same functionalities but with a different UI dedicated for the OS is running on.
With this sample design pattern you’ll have a solid infrastructure that will solve the big problem to port the same application in different screen sizes.
So in next paragraph we can take a look on how to organize our files project.

Manage a project for different screen sizes

Another important thing before starts the project is understand how to organize the project for different OS, in fact if you work with Adobe AIR on Android you’ll have only the XML descriptor with Android permissions described in this file, on Playbook you’ll have another XML file dedicated to this platform and so on.
So, my suggestion is to organize the project in different projects that work together.
In this image you can see how I organize it for our final scope, I’ve a common project (Runtime Shared Libraries in this case but you can use also an Actionscript or AIR project if you work with Flash for example) where I’ll put all classes that are common for different target, so in my case all the presentation models, main models, utils classes, my notification bus and views that are common for different projects like components for example.

In the target device projects I add only what I need for that device, in this case only a couple of views and assets but in big projects could be more stuff that those one:

When you have a specific behavior only for a target device, you can easily extends the common presentation model and add new functionalities or create a side class that you’ll add in the specific view.
So with this infrastructure you can solve bugs and change stuff directly on the common project and all the platforms will be ready to package the project and upload to their store. That’s cool, isn’t it?

Summary

So in this post I hope to give you some ideas on how to solve the big problem to create a common base code that could be useful for different purpose.
In next post I’ll show you how to implement it in practice, for now if you have any feedback or suggestions for the second part of this article please add a comment to this post.

New AT&T Research Reveals Clues to Smarter App Development

This time I’d like to publish a post written by Doug Sillars member of the AT&T Developer Program team, he asked me to publish it on my blog so here we are, great job Doug!

Nothing chases away users (or invites bad reviews!) like an app that is slow or drains battery life too quickly. New research from AT&T and University of Michigan offers insight into how applications can be modified to prevent these issues from occurring. Historically, wireless optimization best practices have focused on how to optimize content for smaller screens and slower processors. In this article, the researchers describe how mobile wireless networks are different than wired networks, and how simple changes to your app can work WITH the network to improve your application’s efficiency.

What is truly fascinating about this study is that developers are given a peek at network architecture and lower-level protocols that are usually hidden from them. The researchers describe in detail the connection latencies and state machines that exist in wireless networks, and how much longer these timeframes are than in traditional networks. The main takeaway for developers is that all connections open and close on the order of seconds (as compared to milliseconds in the wired web). These latencies make content downloads appear slow, and can drain the battery quickly.
The researchers point out that it is more efficient to utilize fewer connections to send more data per connection. For example: 5 connections of 1 KB each utilize 90 seconds of radio, while one 5 KB transfer uses 19 seconds (a 79% decrease in battery usage for the same amount of data!).

Each network connection causes a power drain to the battery, and can make your application appear slow to the end user. By sending more data per connection (and using fewer total connections), you can drastically improve your application’s performance. The researchers provide the example of a news application that downloads data on demand (e.g., images are downloaded when the user scrolls to them). They found that prefetching images (sending many files in anticipation of future use) can reduce power consumption by 15-18%. Since the images are downloaded in advance, they are available for display immediately, and your application appears faster. However, the researchers are quick to point out that prefetching data is a delicate balance, as downloading content that is never used also wastes the battery (and can cause your users to exceed their monthly data caps).

In the case of a popular music streaming service, periodic audience measurements were draining battery life, causing the app to use more than 40% of its power to transmit just 0.2% of its data. By lengthening the time between these measurements, it is possible to greatly reduce the battery drain, and allow users to listen to more music (and at the same time view more revenue-producing advertisements)!
AT&T has begun working with developers of popular applications (like Pandora and Facebook) to enthusiastic responses. The biggest response has been that the solutions to optimize wireless applications are straightforward and benefit both the developer and the end user!