Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

RE: John's asinh split versus convert-to-sortable...



I like this reformulation, it only used sqrt which is usually very fast. 

-----Original Message-----
From: stds-1788@xxxxxxxx [mailto:stds-1788@xxxxxxxx] On Behalf Of Dan Zuras Intervals
Sent: Thursday, March 22, 2012 12:49 AM
To: stds-1788@xxxxxxxxxxxxxxxxx
Cc: Dan Zuras Intervals
Subject: John's asinh split versus convert-to-sortable...

	Folks,

	I think if you look at John's split based on asinh,
	you will find that it already contains, at level 1,
	all the properties you are seeking in a convert-to-
	sortable split.  And none of the faults associated
	with a split defined at level 2 only.

	To review, I will formulate it as follows:

		split(x,y):
			u = asinh(x); v = asinh(y);
			t = (u + v)/2;
			return s = sinh(t);

	where John's fully general split with the scale
	parameter L is L*split(a/L,b/L).  On can use these
	definitions:

		asinh(z) = ln(z + sqrt(1 + z^2))
		sinh(z) = (exp(z) - exp(-z))/2

	To turn that into:

		split(x,y):
			u = x + sqrt(1 + x^2);
			v = y + sqrt(1 + y^2);
			t = sqrt(u*v);
			return s = (t - 1/t)/2;
...