0845	Arrive. Working on blank100 and blank20 as regions to practice the
		histogramming-->fitting-->etc method...
	in GASSright/blank/
		at 19:06:50  -43.34.10
	new script: cfit.g - for cool fitting!
	worked out histogramming the ratio image...
	fit line to x=HI,y=IR, y = mx+b  or IR = m(HI) + b
	when b = 0, then the histogram of y/x values will peak at m, the slope
	when b != 0, then histogram y/x peaks at m + b/x. should I be
		removing this offset, somehow? like subtracting b from the
		ratio image?
1200	Lunch
1240	back. Doing a side experiment... seeing how correlation improves/
		breaks down when looking at smaller and smaller sections
	Added in functionality to corr.g and cfit.g to accept a number, which
		is the edge to crop of the image - edge of 10 takes 10 pixels
		off each side, so a 100x100 image would become 80x80
	worked a lot on GASSright/blank/ field, both in 20 and 100 sizes.
	now going to work on f, and fh
1503	Snack time!
1530	back. yum. jay says let's meet tomorrow when sophia's back.
	continuing the new scripts on fh and a in GASSright/
	in field a:
		a_allv_corr.ps shows a clear split in the higher HI values:
			, so I'm going to put a cut, before histogramming,
			at N_HI >= 28x10^19 N_H cm^-2
		this makes new ratio histogram by means of cfit2.g, which
			combines some of corr.g to get the input pixels
			and cfit.g which plots the histogram...
	in GASSright/a/:
		a_allv_H28_yxhist.ps is a histogram of values of the ratio
			of the integrated HI to IR, only using values of HI
			greater than 28 x 10^19  -  TWO-PEAKED!!! HOORAY!
	going to figure out where these pixels are, with this small script:
		im := image('a.img_v_-250_0250_mom_0.fits')
		cs := im.coordsys()
		ima := im.getchunk()
		s := im.shape()
		for (x in 1:(s[1])) {
			for (y in 1:(s[2])) {
				if (ima[x,y] < 28) {
					ima[x,y]:=0
				}
			}
		}
		im2 := imagefromarray(pixels=ima,csys=cs)
		im2.tofits('a_allv_clip28.fits')
1644	summed up the day's progress on this stuff in an email to DJ, with
		GIFs to illustrate the progress on the histogramming and
		ratio-ing.
	now going to work a little more on the pretty picture...
	applying new scaling images by sin(latitude), with this script:
		im := image('chunk_3_6_log.fits')
		cs := im.coordsys()
		s := im.shape()
		nx := s[1]
		ny := s[2]
		ima := im.getchunk()
		for (x in 1:nx) {
			for (y in 1:ny) {
				if (sprintf(ima[nx,ny])!='nan') {
		
	can't figure out the measures.... but need to find a way to convert
		from J2000 to galactic coords... something with measures
		tools.... for tomorrow!
1745	heading out
		NEED to finish /home/scratch/sjanowie/ 13 and 14...