Archive for August, 2007

Publishing files for local deployment in Flash 8

August 22, 2007

In order to deploy an swf file that is published to Flash 8 and is wrapped in an HTML file on a CD/DVD or thumbdrive, the following set-ups are required:

1.  “Local playback security” needs to be set to “Access Local files only” if the swf needs to load other local swf files. If set to “Access network only”, the swf will fail to load/access other local swf files. “Local files” means the files that reside in local FileSystem or in movable media that are accessable to the FileSystem. They are usually accessed via something like “file:///C:/folder/fileName.swf“, etc, or a relative path like “./folder/fileName.swf”. Network files are those that are accessed via an iNternet protocal such as “http://sub.domainname/fileName.swf“.

2.  If “getURL” is used to access files on the Internet or call pseudo JavaScript functions within Flash Player, “allowScriptAccess” needs to be set to “always”. Otherwise, both IE and Firefox will fail those getURL calls.  If not wrapped in HTML, request to access an HTML page via “getURL” will be successful, provided the file exits, but the browser window will be obscured by the projector or the swf file and will not get auto focus as if the call is made via a web browser.

The above statements are true when tested on FP 8 and FP 9.


MVC/UMC revisit

August 7, 2007

I was doing a simple audio clip playlist a couple weeks ago. Such a small thing got to the level at which I almost couldn’t handle it when I started to try to put some buttons to indicate and prompt users to toggle “repeat all” on and off. It works like this:

If no clip is playing and “repeat all” is off, the player displays “toggle off” icon while displaying a “play all clips” button when mouse rolls over it. If a clip is playing and “repeat all” is off, the player instead displays “turn on autoplay” when mouse-over. When “repeat all” is on, the player displays “repeat all on” icon on “mouse-up” state and “turn off autoplay” on “mouse-over” state.

I got into interwinded “if-else” conditions and was almost lost. I spent almost a whole day and still couldn’t get the buttons work together nicely. I think to myself hey if such a small thing takes me so long to figure out, what else could I possibly do? I have been doing ActionScript full-time for about 3 years now; although most of my stuff is not stellar in terms of the coding level, I’ve gained quite a bit of confidence in coding UI and animation. So I decided to give myself a little break and went to the library and Borders. After reading and thinking, I finally think I not only got the stuff done, but also showed a bit of logic in my spaghetti code. At the end I list the books/articles that I have been reading for the past couple of days. Not that I feel I have been elevated to another level, but here are just some thoughts:

Basically Control is like a hub. It knows about both View and Model and connects View and Model. It contains the logic of the UI. The responsibility of Control basically is about coordinating all the events and actions that affect both UI and Model. Many of them (but not all) are about how to react to the requests from UI and updates it; if there is anything it cannot decide by itself, it relays them to the Model. The rule of thumb here is that if an interaction won’t affect other UI components and doesn’t request anything from Model, Control will handle it directly (like most of the roll-overs) without consulting Model.

If the UI is requesting anything from Model (like asking to play another song or movie), Control would not change View directly. Instead it sends the request to Model and Model makes changes accordingly and then signals Control about the updates. Control/View receives the information about the changes and View is then updated (like dehighlighting the currently highlighted button and highlighting the button that is associated to the clip that is requested, assuming the clip starts playing.)

The change of UI doesn’t necessarily initiated by a request from UI. It could be a system event, like a song completes playing, or a change from Model, like a new song has been added to the pool of all the available songs on the server, etc.. Control should have routines to handle all these changes (such as “handleClipComplete()”, etc..)

It’s just a matter of your coding style and about the complexity of the entities you are dealing with when it comes to whether to use classes (each of which present only a segment of MVC), or whether to have one UI component per View, or just to stuff everything in one big chunk of code. It’s the thought behind it that matters.

It helps me a lot since I’ve started to learn to draw diagrams of the states (statechart) and then to figure out all the classes and operations/attributes from it. I am using paper and pall pen instead of any of the UML tools. It helps me to concentrate on the thinking.


Adding index for a table

August 3, 2007

In Excel:

Approach #1: In the cell 1 of the index row, type in “1” and cell 2, type in “2”. Select both cells. Excel highlights both cells with bold-stroked box. There is a small handler on the bottom-right corner (Supposed to be a feature called “Auto Fill”?) Hold and drag the handler all the way down to the last cell on the index row. Excel should autofill all the rest of the cells.

Approach #2: In the cell 1 of the index row, type in “1”. In cell 2, type in “=ROW_NUM1+1” (For example, if this is row A, type in “=A1+1”. What this formula does is to increment the value of cell 1 by 1, which equals to 2.) Select cell 2; drag the small handler on the bottom-right to all the way down the last cell.


Approach #3: Create a column which is primary key, data type of int and attribute of AUTO_INCREMENT.