2005-4-2

Add common /parms/ to fun3b.f and rhodef.f so that rhoexp no longer
returns non-existent values.

2005-3-31

Rewrap all stop 'message' commands with call intpr, but pathmix(4, 
remains to be fixed.

2005-3-30

Tidy up *.f. It turns out most problems were caused by arrays ncount and
bin in pxfamr.f -- with sucess of path3a after this fix.

2005-3-29

Add do i=1, ngroup / ncount(i)=0 /enddo to pxfamr.f. First dry run of 
pathmix(3,) and pathmix(4,) with istat=-1

2005-3-28

Redo ALPHA = -1.0 within almini.f so that gemtest.f works. It remains to
fix data I/O.

2005-3-24

Note that reverse to "implicit double precision" actually makes almtest
work but not gemtest; this is after fix calls to alfun via iop. Also 
note that -O2 (as in R CMD check) actually does away with the f77 and g77
problem come across earlier.

2005-3-23

Note that there are HUGE differences between g77 and f77 therefore mask
'implicit double precison', also add -w to FFLAGS in Makevars. It is 
further noted that switch defined by iop is not accessible to alfun.

2005-3-22

Inform Kathryn V Steiger <ksteiger@illumigen.com> about POINTER and PATHMIX
Reuse Makevars so that .inc can be used, all *.F remain as *.f.

files.f, path3a.f, path3b.f:
Add FILNAM1, FILNAM2, FILNAM3, FILNAM4
inst/doc:
Add pathmod.dot, pathmod.ps and pathmod.pdf

2005-3-21

Attach '3a', '3b' to defalt.F, input.F, output.F, path3a.F, path3b.F.
It appears the key element is FUN(), therefore add 'iop' to ALMINI and therfore 
ALMIZE, with iop=1,2,3,4 for almtest, gemtest, path3a and path3b. There is 
insufficient argument with 'call dtime (TArray)' and therefore change it to 
'call dtime(TArray, result)'. As the data, job, prolix and terse files are
specified through files.F, it is simpler to call path3a and path3b. 

almini.f:
Alter branch to "200     ALPHA = -1" within if/endif block

rdcorr.f:
Reorder CORDAT

path3.inc:
/CORDAT/, /GOF/ have different sizes from gof.f and rdcorr.F

Note that *.F are required since they have #include directive, but R would
not take *.F therefore the first pass was gained with all *.F as *.f.

2005-3-20

Redo compiling under Sun with the original pathmix and create the template for R
It remains to fix the differences between path3a and path3b w.r.t. the following 
subroutines,  

defalt.F, fun.F, input.F, output.F

The same is true with almtest.f and gemtest.f which contains other definitions
of INPUT. A method is necessary to get over these problems, possibly via an extrat
argument for these subroutines, or ignore these tests entirely.
