Code Zen

Last night Mumbai suffered what some news reporters called the worst act of terrorism India has ever seen. Mumbai unfortunately, is no stranger to acts of terrorism but these attacks were of a completely different kind, the attackers definitely suicidal who just wanted to kill as many people as they could before they themselves were killed. Even after 24 hours, the drama still is not over with the security forces still trying to rescue people who have either been taken hostage or have been stranded locked up in their rooms.

I was glued to the NDTV live steam for most of the night and since I have woken up this morning, and on tracking the story on search.twitter.com as well as occasionally jumping to BBC news or CNN. My first emotion was anger. Mumbai was my home for a while and its still my favorite city in the world. And to see it virtually being held hostage by these loathsome assholes enraged me. My sentiments towards the government and the security forces were those of sympathy almost, they are trying their best arent they?

But I talked to my brother over chat then and another friend who also often visits Mumbai, and they were less so. The attacks in Mumbai weren’t unexpected, after the recent attacks in other big cities where Mumbai was conspicuously out of the list. Its the job of the government to keep its citizens safe, and it failed miserably this time. While the on site journalists kept insisting that the city will bounce back and take it in its stride, I really hope this wakes us up as a country and makes us ask some serious questions of ourselves and of the people we put into power to make sure things like this don’t happen. Its time we stopped being reactive and “politically correct” and just root the sources of such terror out. Lets stop taking the safe and conventional responses. Give the security agencies more rope and let them go after the core of the threats irrespective of political sensitivities. We need more supercops like KPS Gill

The drama still goes on in Mumbai but I hope this is a wake up call to us all.

I got back from San Francisco this morning taking a couple of extra days at San Francisco to do some tourism. So I figured I’d put down what I had learned this year before the workweek clouded all that to a distant blur.
As usual, MAX was f-ing awesome. I have tried to make it to MAX every year since I graduated and joined the corporate workforce and Comcast has been pretty awesome about making sure I could do that on their dime. This year was no exception and most of the Flash team from CIM was there at MAX as Adobe showed of the cool stuff they have been working on.
My strategy for MAX this year was a little different than earlier. Traditionally I have tried to get deeper into technologies that I work with everyday. This year, with so many new ideas being shown at MAX, I decided to take sessions on the things that I did not work with between 9-5 and would be really easy for me to miss. I have to say it worked out pretty well.
Let me start off by saying: HOLY S**T, the Flash Platform is AWESOME and the pace at which its evolving is just incredible. Below are some of the best sessions I attended this year:

Using C++ Libraries in Flash Player and Adobe AIR
This was definitely one of the best sessions I attended at MAX this year. The session introduced Adobe’s Alchemy project that compiles C++ code to Actionscript bytecode that can be run on the Flash Player just as any bytecode compiled from AS3 code would. The project isn’t mature yet but some of the examples that were shown like ports of C++ OggVorbis codec, motion detection libraries, cryptographic libraries, etc were definitely impressive. I guess its time for me to brush up my C++ ;).

Hands On: Creating Effects with Pixel Bender
I think I got most excited during this lab. Pixel Bender, as a lot of you may know, is a technology that allows developers to write Shaders in a language called Hydra. Hydra is has a C-esque syntax but is really easy since you dont do really complicated logic in it, but rather simple pixel manipulation. These Shaders can then be applied at runtime to Flash DisplayObjects and also be used within Photoshop / AfterEffects as custom Shaders. I had read a bit about this before but the syntax and the prospect of having to learn another language was keeping me away from writing a custom Shader. However, it was really surprising how ridiculously easy this was. Hopefully I’ll write a couple of filters soon and release them on this blog.

Extending Flex Builder
Another awesome session, almost perfect for me. This session introduced the object model that the Flex/Actionscript plugin uses for all the language intelligence within the IDE. The session was really useful since I had a couple of ideas to extend Flex Builder but while I was able to learn about Eclipse itself, I had no knowledge of the details of Flex Builder plugin itself.

Hands On: Getting Started with AfterEffects CS4
Programmatic Animations in AfterEffects with Expressions and Scripting
On day 1, I attended part of the AfterEffects workshop in the morning. I have never worked with AE but have seen some designers I work with come up with some incredible stuff with it. The session was perfect since it taught me enough to now let me go play on my own. Unfortunately I had to get out of that session half way since I had to get to a meeting (yeah I was also working there) but the pace of the session was perfect.
On day 2, I attended another session on programmatic animation in AE. Since most of the animation I do in Flash in programmatic, this really made me map that knowledge to the AE domain. Between these two sessions, I think I really got a lot clearer understanding of AE than before.

Adobe XD: Designing Design
Adobe XD: Building a Next Generation News Reader
I loved the XD sessions. Seeing how they go around developing apps and how they make decisions around user interfaces is always interesting. The session where Jeremy Clark explained the process of developing the New York Times news reader application was particularly informative, where they even showed the process of prototyping interaction design, something I really feel all companies should do more of.

But more than the sessions, it was meeting other attendees that was just plain amazing. I had some amazing conversations around the future of LogBook and OpenPyro with quite a few people I have talked to on twitter and email but seeing them in person was very exciting. These were almost as educational as some of the sessions.

Max 08 was just amazing, and am definitely looking forward to Max 09 in LA.

For as long as I have been a Flash developer, I have been completely addicted to it. And its only getting better with age. But off late, a lot of people have been eager to prophesize the death of Flash (usually siting poor Flash implementations as reasons). However there are certain features (not counting the wow effect) in Flash that make it a good choice for certain implementations. For one thing Flash Player (esp greater than 9 which 98% of the users now have) is much faster in processing data than javascript, especially on older browsers. Another reason is that a self contained swf file is a lot more portable when moving from one environment to another. In today’s world of Facebook widgets, OpenSocial containers, etc etc, deploying a swf is a lot easier than packaging a suite of javascript files. Actually a lot of those containers won’t even let you deploy JS on their servers and force you either to use an html extension like FBML or use an IFrame (and again there are limitations set on where those can appear on the sites). So it was really heartening to see that officially be posted on the Yahoo Developer Networks blog by Jonathan LeBlanc, Senior Software Engineer / Technology Evangelist from Yahoo. So do give it a read and use that as ammo if you are in another one of the “Why Flash” conversations.

The post can be found here.

Also take a look at the Yahoo Open Strategy (Y!OS)’s Flash SDK here (Yahoo has only released a Flash and a PHP SDK).

Go Flash !

Mike Chambers, Principal Product Manager for developer relations from Adobe, has been pinging the Flash community about whether AS2 based development is faster than AS3 for Flash projects. Having spent a long time in both languages, I thought I would post my 2 cents on this blog. So my answer is: Yes. Of course I’ll qualify my statement by saying that 1) There are a lot of capabilities in Flash 9+ that are amazing but I want to disconnect the language (pure syntax) from what it enables and 2) This post assumes you are as comfortable with AS3 as AS2. I am not factoring in the learning curve for a new language, which happens anytime there is a drastic change.

AS3 can use some help. Let me list out some reasons why I say that:

1. Loading XML is a complete pain

Calling an external URL is AS2 was simple (although having to use delegate because of the lack of proper closures was retarded). The code was something like this:
[AS]
var xml:XML = new XML()
xml.ignoreWhite = true;
xml.onLoad = function(success){
if(success){}
else{}
}
xml.load(‘doc.xml?foo=bar’)
[/AS]

Very readable, plus all errors were available in one check. Lets compare that to AS3:

[AS]
XML.ignoreWhite = true;
var urlLoader:URLLoader = new URLLoader()
var urlRequest = new URLRequest(‘doc.xml’)
var urlVariables = new URLVariables()
urlVariables.foo = bar
urlRequest.data = urlVariables
loader.addEventListener(IOErrorEvent.IOError, onError)
loader.addEventListener(SecurityErrorEvent.SECURITY_ERROR, onError);
loader.addEventListener(Event.COMPLETE, onComplete)
urlLoader.load(urlRequest)

[/AS]

How verbose is that? I am still trying to figure out why every request needs a URLRequest object, seems unnecessary. just calling urlLoader.load(“data.xml”) could automatically create a URLRequest object, if you ever want to examine a request post it being called. Another huge problem is most developers I know dont actually add the error event listeners, just cause adding it is so tiresome. So when an error occurs later, they are usually clueless to what happened (was it a code bug or did some http call out just did not work?).

The URLRequest param for navigateToURL seems another one of those requirements that seem unnecessary.

2. Whoops, forgot my addChild

I think this is where I have lost countless hours, I would create a sprite, and set its properties, but forget to add it to the stage…and then spend too much time to figure out what did I do wrong. The only way to create a new UI element in AS2 was by referencing an already added MovieClip (remember my_mc.createEmptyMovieClip() ?). I would love to have a createAndAdd function in DisplayObjects that becomes the default way of creating new Sprites.

3. Dynamic drawing: The graphics object

Drawing in AS2 was simpler: MovieClip.lineTo(….). In AS3 it has to be on the graphics context of the DisplayObject (mySprite.graphics.lineTo() ), but why? There isnt any other context to draw on, and the graphics context isnt swappable, so why do I have to type that in?

4. No numeric depths

In AS2, you could set explicit depth numbers to elements and reorder elements under those depths. So for example, if you were writing a button control, you could set the label field up on depth 1000 and keep swapping out the children under it. In AS2, you have to track where in the display tree your element is.

5. useHandCursor vs ButtonMode + useHandCursor

I dont know what buttonMode does besides actually allow me to useHandCursor property, which I have to set anyway

6. No onReleaseOutside

This one also is weird. The way we now have to do an onReleaseOutside is via listening to a stage click event. But wait, you may not be on stage yet, so you need to check : if(stage), and also write an event listener to the addedToStage event. Verbose.

7. addEventListener vs. onRelease

onRelease (and similar “on…” functions) is not a bad default hook to have. Considering how often we write these functions, having a quick shortcut to the action was nice. Not that I am saying the event model is bad, but I end up writing a lot of addEventListeners and removeEventListener throughout the code.

I think overall, when I look at an AS3 file, it just seems longer than AS2 code used to. I think everything needs to be very explicitly set in AS3 with no smart defaults. Maybe that’ll change in AS4 now that we no longer care for the ECMA spec ;). There are a lot of improvements in AS3 as well though, like actually having private and protected namespaces, but I am only listing out my pet peeves here hoping Mike takes note :). Flash is an amazing platform, and I love what it enables me to do.