Senior Project Week 9

Richard Banister
CSC 4790 Senior Projects
Work Log

Week 9: 27 October - 2 November, 2008

Put together a real solid ant build script; project folder set up for proper building. Code refactored properly and building correctly. 2.75h

Extracting video information from a GET. Probably best to use PHP for this, but that means I need to install it to test locally. Will hardcode for now, and do this when I really need to. 0.25h

Subversion working at home; project committed to repository on minimumrage.net. 0.75h

Subversion working at school. svn+ssh to minimumrage.net using PuTTY, Pageant, etc. Ant build working at school. Workarounds needed due to not being able to change system settings (environment variables, etc.). Unfortunately this would need to be done for every machine I wanted to develop on. I've been using just one, but it happens to be right next to the window and is thus the coldest possible place to work in all of Mendel, and winter is coming. Setting up a development environment at another workstation might need to happen. It will be quicker now that I know how to do it. 1h

Working video slider/scrub control. Working standalone video screen component. Seeks according to requirements document, including interactive objects. Much better understanding of component extension, event model now. Also organization of project classes is much more sensical. I probably should have aimed for strict MVC from the beginning, but I had to get through a lot of stuff in a frenzy. I'm closer now. 5.5h

Custom relative seek control working. Created custom events, dispatching, etc. Need to revise scrub tool to use custom events now. Also need to connect the video screen and the controls better. Right now all the seeking/scrubbing code is handled in the master class, which I don't think I like. What would MVC do? 1.5h

Communication with html/js wrapper. Still need to get PHP working; but this is more necessary now - ran into some sandbox problems when working with wrapper communication that would be solved by running off a localhost server. I may have a PHP server thingy installed somewhere, now that I think about it, but getting that working or some alternate thing may take a little doing. But regardless. Encoded a better video. Wanted to use h.264 but encoding software's not there yet. Maybe when I get to CS4? The On2 VP6 stuff looks good enough for now (although not great - but it's a code release, I doubt anyone will care very much). Resized video controls to be reasonable on a page. This puts the controls in perspective - much less space to work with. But it will be okay. The seek bar doesn't need to take up much room, so I'll put that on top of the scrubber, with marker controls, etc. to the right. Then the scrubber can extend all the way to the right, making it more useful. 3h

I do wonder if this will prove to be a different beast once I put it on a server and have to worry about bandwidth.

Setup PHP server locally for testing. Get video info from GET variables, pass through javascript to SWF, which loads the XML file with video properties in it, then load the video and all associated objects. This is turning into quite the multicultural affair. Linking should be quite easy now. I wonder if some info should be passed via a POST and not a GET? Does it matter? 1.5h

Handle case when video is not identified/improperly identified in url (disable video player). Possible security problem here? Ran into problem with caching of SWF which took up an unfortunate amount of time. Fixed by addition of a couple meta tags to wrapper file. Video player more properly sized at startup. Removed redundant/conflicting resize code. 1.5h

Better notification of video time + status (although not really perfect). Overloading methods (see this post in the development blog). Caching XML problem fixed (see this post in the development blog). Dynamically grab title from XML document and update wrapper HTML (calling Javascript methods). 2.5h

Interactive buttons overlaid on video. This will be the basis of interactive regions. Architecture of same. Movement specified by waypoints. Pretty awesome. Movement is still jerky, but that's the least of my problems. 5h

Built a better debug system. Before, all debug information was going to a log text file for flash in an obscure directory, making it difficult to get to debugging information when you need it. Now I have it so you can turn debugging on in the URL, in which case debug output gets written directly to the page on which the SWF resides. Much better! Also tried to import my Ant build file into Eclipse, but failed after much hair-pulling and gnashing of teeth. My build works perfectly on the command line; I don't know why it's so hard to get Eclipse to just run it. 1.5h

Restructured interactive objects. They will now have a common base class of SutekiObject (extends mx.core.UIComponent), which will handled a lot of the work in placing them, moving them, and dealing with interaction (since much of that is common to all interactive objects). So each extending class will define its look and feel and anything else it needs; if, for example, a label is required it can create a Label on itself. 0.5h

Region interactive object. Actions for interactive objects. Framework and XML structure established. Transparent popup window for textual content, with link to URL for further explication. Successful placement and use in video. 5.5h

Created css stylesheet and layout for application; looks much better now. Always a lot of struggling with laying out things in css. Show/Hide interactive objects through javascript on page. 2h

Working on subtitles. Should have been easy to extend my interactive object class, but ran into a problem that I couldn't figure out for much too long. Of course it turned out to be something incredibly stupid and sad. But still struggling with getting the text to look like text. Nothing works like it should. It's very strange. Any other day you might add a Label to a container and the label would properly size itself based on its text. But not for me. Something must be going wrong somewhere. 4h