jesusabdullah edited this page Oct 19, 2010 · 55 revisions
Clone this wiki locally

To-do List:

  • Review derivation of equations, double-check square term and figure out how a non-squared term would make sense (if it does).
  • Order some stuff!
  • Figure out what’s causing segfaults
  • Change sim to do less runs before save, in order to repeat less work
  • Figure out how to solve the problem space :v

To-Done List

  • Update resume some moar. (YAML for data file(s)?)
  • Find a third committee member. See Ask A Mathematician.
    —> Rawrik suggested someone from CNSM, particularly one with a solid background in tensor algebra
    —> Dr. Bueler? At this point, probably. He agreed to help, and has been, well, of pretty good use already.
  • Actually, y’know, make the model. XD See parallolism.
    —> It turns out that there are bugs to the underlying “worker” code that need solving before parallolism becomes a legit concern. Chances are simpler code will save more time than using multiple machines.
  • Abstract is being submitted without any results. D:
  • Submit an AGU abstract by September 2nd
  • Mebbe ask Dr. Johnson about the other papers? Could be a good idea.
  • Given time, make some more scripts for Augustine.
  • Attempt to get Sikuli working on ARSCputers. sikuli is pretty buggy all around imo
  • Attempt to compile Sikuli on workstation
  • Attempt to get Haskell working on ARSCputers.
  • Come up with a system for local installs and share via Github. Also share any other tidbits I picked up.
  • attempt gmsh install on arsc for fastness Worked, but not as stable as comsol from experience
  • Rewrite dataset generator to use -CELP tools I found. lessons from CELP and tools from Methlabs. Someday. It works now, so :/ Why fix it? XD-
    → -New plan: make dataset generator only deal with k-values. Generate angles separately.
    → -addendum to new plan: Use python. It’s your tool du jour anyway rite?
  • bashreduce weird netcat version, don’t feel like trying to port to obsd netcat
  • Look at the INFORM problem! (and register!) Lack of time and interest, unfortunately.
  • process supervision frameworks: supervisord , god , perp, monit
  • experiment with reverse ssh action ARSC sucks

To-Deferred List

  • Given time, make some scripts that clean up Augustine. No time
  • Given time, process information from Augustine on ARSCputers, automate as best as possible. No time
  • Write up the generalized geometry form of the k-value stuff I worked out earlier today.
  • Write a program that computes theoretical k-values based on geometry.
  • Rework any k-vals plots I have into something useful. Maybe use Haskell’s charts for learning! Plots not that useful imo
  • Possibly experiment with sfepy. Even if I don’t use it for this project, cause we liek pythons. Can also now take comsol meshes. :3 meh nvm
  • Supposed to chat with CFC re: MTUAF.
  • Throw some designs and a workshop list onto Github
  • Figure out what ARSC’s beef is with github. ugh screw that noise
  • Get ahold of Epic.