! Do x-coord first -- easier?!?

----------------------------------------------------------------
k Run phase 1 once for each k
k Run for n = 50, 60, 70, ...
k Get medians from PMFs

k Find a's and b's from graphs:
  - 1st, (relative) a's to level the heights
  - 2nd, b's to get the origins to zero
  - use try2.py to get pairs of n's right
  - then use tryrescale.py to look at all n's

k Get shifted PMFs

! Trim down to top 1100 lines or so for reduced du?
  - Make a script for this.

o K-S test to see if shifted plots match one another

k Lin fit to get alpha and alpha*ln(k)
k From that, get k

k Then go back & do plots scaled by 1/(median*k)
o K-S test to see if these match the theoretical distribution ...

----------------------------------------------------------------
... Theoretical distribution:

* c? alpha?

* google stable subordinator; stephen lalley?

----------------------------------------------------------------
R:
* install fBasics package via synaptic
* R prompt:  library(fBasics)
* http://hosho.ees.hokudai.ac.jp/~kubo/Rdoc/library/fBasics/html/StableDistribution.html

* x = seq(0, 2, 0.001)
* plot(x,dstable(x=x,alpha=.55, beta=0.9999, gamma=.2, delta=0.2))

* http://hosho.ees.hokudai.ac.jp/~kubo/Rdoc/library/fBasics/html/StableDistribution.html

* plot(x1, y1, type='p', xlim=range(x1,x2), ylim=range(y1, y2), xlab='x', ylab='y')
* points(x2, y2, type='p', col="red", xlab='x', ylab='y')

* maybe get a better fit (alpha ~= 0.75?) w/overlay plot & R I/O from rescale
  data.  Find out how to do that.

* ac = read.table('all_columns.txt')
  plot(ac[,1], ac[,2])
  points(ac[,31], ac[,32])

!! stableSlider()

----------------------------------------------------------------
R misc:
* x=c(1, 2, 3)
* names(x) = c('a', 'b', 'c')