TODO

put in a alpha based window taper

joint96 in ddupdt and wcupdt  I set both m and m+mmax e.g., V and Q?

08 DEC 2002 ----- get rid of gaussalp weighting in jramat
	also test new version of joint96 currently bombs 
	because of tmpsrfi.00 format

	CHECK how vars varr are computed, especially vars
	when use/not use observed std error

10 JAN 2003 - added option 50 to change the RFTN trace weight
	also gid rid of the two sources with different ssytem/system
	calls

	Instead I make put this into the makefile
	Note Makefile.SLE actually used the system and not the ssytem
	so I will continue this but check on the SPARC

16 JAN 2003 - seems OK Now joint96 with jamat is the prototype of
	the rftn96 and surf96 since these are actually subsets of the
	joint96 program. This means that I plae the complexity in
	jamat ramat samat and but all use jsaqmat and jramat unchanged

	surf96 is joint96 but with a call to ramat or samat instead of
	jamat

FOR A WATER LAYER - IF P = 1.45 WHY DOES THE WATER DENSITY CHANGE
FROM THE INITIAL 1.45 1.234 to 1.45 to 1.00?
