User:Sdowne/summer 2010 worklog august

From CDOT Wiki
< User:Sdowne
Revision as of 11:11, 25 August 2010 by Sdowne (talk | contribs)
Jump to: navigation, search

Tuesday August 3nd

  • Final touches to lextLeading
  • back onto popcorn and open video
  • started to research a new demo, brainstorm ideas.

Wednesday August 4th

  • continued with research for a new demo. I am hitting a lot of dead ends, and not sure which dead end to turn into an opening.
  • Talked with Brett about popcorn, about where it is heading, what is important, and to get ideas for the current and future demos. I think I need to focus on, as much as I can, breaking that wall between flash and the web. I think a strong argument with using javascript over flash is I can have the library, made in a way that is easily integrated with familiar things like jquery, processing, and any javascript library. Something the xml limits us on is that it is just xml, and not integrable with any and every javascript library, but also using xml in an of itself is an advantage, and maybe supporting both is the best way.

Thursday August 5th

  • First thing I did was re read yesterday's conversation with Brett, to help it sink in and warm me up.
  • Researched a new demo, and continued with popcorn research

Friday August 6th

Monday August 9th

  • Did an unofficial review of js only processing.js patch
  • helped anna with xml stuff The specifics of what she was doing was beyond me, but, I was able to parse data with some regex stuff I helped her with. Knowing little as possible; I had x data, and I needed to make it into y data, and made regexs to do this. I love regex :)
  • Swing back into processing.js textAscent and textDescent. Anna failed them, which is good. I started to fix them up.

Tuesday August 10th

  • Worked on textAscent and textDescent. This is going to require a rewrite. What I am doing is all wrong, and stems from the unavoidable difference between java fonts, and browser fonts. TextLeading will also have to be fixed. Ugh... I hate fonts...

Wednesday August 11th

Thursday August 12th

  • I think I am trying to balance too much at once, and it's making it hard for me to concentrate and focus on something. Personally, I work best when I can get lost in something, and obsess over. I need to start putting things to rest, and get lost in popcorn. Anyway, moving towards getting html5 current time, and a demo put to rest. Got started on that.
  • Ended up doing a review of updates to patch init.js

Friday August 13th

  • Was stuck all day on changing a html5 video's current time. I wrote a blog post on it next Monday (just went through time a little :P). The fix mentioned in the post was discovered a few minutes before the day was over.

Monday August 16th

  • Worked on the video demo Brett from Mozilla started while he was in Brazil, working on the fix discovered in my last worklog entry.
  • There is a problem with loading the video based on time stamps, instead of having many multiple smaller videos... it's just as slow.
  • Popcorn conference call with Brett
  • Going to try the brazil demo with multiplel videos again, but instead of creating the videos, have all videos loaded, and hide and display them as needed. Might boost the performance.

Tuesday August 17th

  • Finished up a refactoring of popcorn.js that I have been slowly working on while balancing other projects...
  • Started phase two of popcorn refactoring
  • Blogged and did final research about HTML5 video currentTime problems. Hopefully I can put this to rest and move on to popcorn.

Wednesday August 18th

Thursday August 19th

Friday August 20th

  • Continued working on text leading
  • I assessed the main problem with textLeading. it has to do with infinite recursion because of calling processing instances inside createGraphics which calls another processing instance, etc. This is caused by the way I solved ascent and descent. Blogged about it.
  • the other problems with textLeading was some code I wrote to make the results looks the same as p5, which I shouldn't of had to do, the reason they looked off was because of my previous ascent and descent were incorrect. Now that I fixed those a couple days ago, some of text leading can use the previous code. Basically, I took my failed code, and the original code, merged what works in both; it wasn't too hard.

Monday August 23st

  • Ok, text leading is going to require a small, but efficient class to emulate the text drawing of processing. This will also be faster.
  • I got the class working in a reduced little demo, I just need to add it to processing and I will be done.

Tuesday August 24nd

  • finished and submitted textLeading
  • Popcorn conference call with Brett, Corban, Joel, Anna, and Jason. We talked about Popcorn's future, using a wrapper for the video element, to emulate the html5 video element's functionality, but to wrap multiple videos into seemingly one video. Could also be used for flash or canvas(to create a test suite). Also talked about switching from xml to JSON, and exporting xml or JSON.
  • Started commenting the processing.js parser.

Wednesday August 25rd

  • Did an early fix of a small bug in text leading
  • blogged about my recursion fix
  • continued commenting processing parser