Hello there, Flash cowboy!
If you can read this message, you should apply for a job at blip.tv.
Send an email to careers@blip.tv with your resume and make sure to
let us know that you saw this message.

Love,
The blip.tv dev team

What a clever way to look for flash developers, if they see the message they are likely developing in flash. Clever.

Recently I had this happen again, and apparently it was because my computer at home had its date set later, or my computer at work had its date set earlier. So when I brought the files home and worked on them / saved them, then brought them back to work I started getting these errors again. I finally found some developer in Asia working remotely with some company in on the East Coast having the same issue (because he was in the future). He said once he had the company on the East Coast change their date on their computer to the next day, all the errors went away.

So I tried the same thing, and boom my errors were gone. Just had to change the date in order to publish .as classes from the “future”. What an odd flash bug.

Recently at work I finally got the upgrade and opened one of my Flash 8 ActionScript CS2 (.flp) projects in Flash CS3. I immediately published the project, as its one of the options from the project window. It came back with 72 errors, something about a naming conflict in my classes.

Well this is a lot, and although Flash CS3 does a great job directing you to where the error is (it now opens the class file and highlights the line with the issue), I for the life of me could not figure out what was causing this. So after poking around at various files for a bit, I recompiled the project and only came up with 12 errors. The really funny thing was I really didn’t change any of the code. And of course they were all classes I wrote for this particular project. So initially I’m thinking that I’ve royally screwed something up, which is always my first instinct.

So I poked around some more and eventually my errors all disappeared. My first thought was that maybe I didn’t restart my computer after I installed Flash CS3, which is always a good idea. However this was not the case (although yes I’m working on a PC at work).

So next I opened a similar project at work that has the same kinds of errors. This time I was working on this project with another coworker that was having no issues publishing the same files (we were using the check in/out setup from a local raid server). So after a lot of going back and forth, and even successfully publishing the my files on his computer (via a shared folder over the network) I finally figured it out.

So here it is, if you don’t actually open the .fla file (from Flash 8/MX 2004) in Flash CS3 before you publish it, the compiler will throw error for naming issues in your classes (it will say something like “The name of this class, ‘com.domain.slide.MonographBack’, conflicts with the name of another class that was loaded, ‘com.domain.slide.MonographBack’.”). Opening my Flash 8 .fla files in Flash CS3, then publishing took care of all these errors. How funny.

http://proto.layer51.com/d.aspx?f=147
Robert Penner, Dave Yang, and a number of others contribute some other various methods that accomplish the same thing are listed here. Robert Penner’s array.shufle method was rated the fastest, but Dave Yang had a modified version that might run a little faster. One thing to note here that would be a big help with running this on a mobile device is that the old deprecated random() function runs faster than Math.random().This post was based off of what ended up being a competition on Flash Kit. You can find the post here (start on page 2).

http://www.actionscript.org/

This is the other threat I found kind of interesting. It started out with a simple question on how to make a “Who wants to be a millionare” flash game bo some n00b. You can find a varioation on the Penner method here and an interesting version with objects.

Two of the biggest problems working on the Leapster were overwhelming its little prosseser, and fragmenting memory. It was quickly evident when you overwhelmed the processor, because the game would just freeze until it finished processing whatever that task was. But fragmenting memory usually wouldn’t come up until much later when QA was really hammering on the whole title and red squares would show up on the screen. For me the problem was usually being lazy with my array creation and deletion. Even so, you only really wanted to use an array if you really had to. This for me was like putting the training wheels back on my bicycle.

Let me just preface this by saying the Leapster runs on a modified Flash 5 engine. When I started working on these games the method was to use a for loop to parse items from XML nodes into one array, then randomly parse that array into another. I did some searches at the time, and coulnd’t locate a better method. So I used this for my first title, and ended up with some memory fragmentation issues I was eventually able to solve.

On my second title, I wanted to do something a bit more efficient. Plus my boss was getting on everyone’s case for using any array’s at all.

this.numOfQuestions = 6; //number of randimized items to create

this.question_array = new Array(); //create your array

for( var f=0; f < this.numOfQuestions; f++ ){
//pick a random item placement in the array
var randomPlace = random(this.numOfQuestions);

//if that random array placement has no value, fill it with f
//otherwise finish this iteration of the loop by decrimenting f by one
if(this.question_array[randomPlace] == null){
this.question_array[randomPlace] = f;
} else {
f--;
}
}
trace("this.question_array = "+this.question_array)

Basically what this does is force the for to skip array placement of a value if that array item already has a value, and the loop runs a few extra times. Its not a terribly efficient for loop, however its far nicer to memory usage by not creating the extra array. And when your working on devices this tradoff has tremendous impact on your final product.

From there its just a matter of cycling through the question_array till you use every item, then just run this for statement again (most likely in some function or class).

Simple example:
//using the deprecated random() function in Flash
var aRandomNumber = random(5);
trace("my random number = " + aRandomNumber);
//in the output window you should see something like:
//my random number = 3
//
//note: you will see results within a range of 5,
//however this range starts at 0,
//so you will get any of the following results: 0, 1, 2, 3, 4

Imagine it like that classic 80′s game Simon. But just like in that game random doesn’t always mean unique. If you remember playing that game each game you played was unique, however you might hit red 3 or 4 times in a row. This is very typical of a random setup, overall your experience is unique, but you may end up with the same result several times in a row. Conversly you may end up never seeing certain results (i.e. you may never see blue used in a round of simon).

Random exhaustive is sometimes the answer to this, depending on your problem set. Random exaustive can give you a randomized problem set. Imagine random exhaustive as a new deck of cards. When you open the box, all the cards are in order. Then before you play, you shuffle the cards so they are in some kind of random order. You deal the cards and play until you use up the deck, and shuffle again.

Now using the example of the game Simon, this could take away some of the flavor of the game itself (red, blue, blue, blue, yellow, red, green, etc…). However for longer problem sets a random exhaustive problem set can be an advantage, not only for the user experience but also for QA testing (since you can test every problem in the problem set if you play long enough).

So thats the idea. Following shortly I’ll be writing an article on how to accomplish a random exhaustive problem set.