Comments on: MBH98 Source Code: Status Report

By: Sam Urbinto

Sam Urbinto — Wed, 07 Nov 2007 17:33:44 +0000

Mann’s code: Based on a question from Steve and what Jimmy Smith said and steven mosher said, I’m bumping this one. Have at it gents.

By: Jim Mitroy

Jim Mitroy — Mon, 08 Aug 2005 09:12:02 +0000

The ftnchek program is designed to check the syntax of program
vs strict fortran standards. So programs which fail the checks
can potentially give different results depending on the
compiler/CPU mix.

Uninitialized variables are usually set to zero, and most
compilers do have an option to enforce this. But, the people
who write programs with unitialized variables generally do
not bother fiddling with compiler options.

The 470 or so such mixed REAL*4/REAL*8 mode expressions
mean the program is more susceptible to round-off error
creeping in and possibly polluting the quality of any
computation. I noticed that the code had an SVD in there,
if I was doing an SVD I would generally like to maintain
the integrity of all the 15 digits of REAL*8. It should
be noted that there are compilers with options that can
force everything to be REAL*8.

For the record, I like to know what is happening to all
15 digits of my floating point numbers as I add, subtract,
multiply and divide them. Mixed-mode arithemetic and/or
unitialized variables do not constitute good programming
practice.

Jim Mitroy

By: Jim Mitroy

Jim Mitroy — Sun, 07 Aug 2005 03:21:52 +0000

Since the economy of the free world apparently depends on the
hockey stick I decided to run the code through the the FORTRAN
syntax checker ftnchek.

There were close to 100 screen fulls of warnings concerning
mixed mode (mainly REAL*4/REAL*8) arithmetic. The tail-end
of the output is enclosed below.

This is what I got

……………………………………………………..
……………………………………………………..
Warning near line 2960 col 13 file multiproxy.f: real G
promoted to real*8 S: may not give desired precision
3012 570 RETURN
Error near line 3012 file multiproxy.f: missing END statement
inserted at end of file

2 syntax errors detected in file multiproxy.f
476 warnings issued in file multiproxy.f

Warning: Subprogram CSVD argument usage mismatch at position 6:
Dummy arg IP in module CSVD line 2749 file multiproxy.f is used before set
Actual arg IP0 in module %MAIN line 1138 file multiproxy.f is not set

************************** END ftnchek output ********************

I really would be happier if decisions concerning the future of
the world economy did not depend on whether the compiler sets
unitialized variables to zero.

Steve: Could some of the diagnostics be compiler-specific? He’s produced results, so at some level I presume that his program must at least compile.

By: Larry Huldén

Larry Huldén — Wed, 03 Aug 2005 05:34:36 +0000

Is it possible that Mann originally believed in his hookey stick that much that he made the calculations quickly to get the results published and expected that they would be more thoroughly verified later? Because nobody has later been able to verify what he did he doesn’t dare to disclose everything because the original reconstruction was not properly done. I feel pretty sure that he will never disclose completely the methods and data used in MBH98 & 99.

By: fFreddy

fFreddy — Tue, 02 Aug 2005 15:21:42 +0000

Steve, the link to the lexicon is bust. Steve: Fixed.