CALMOD falling down slow

Bob Nolty (nolty)
Tue, 22 Aug 1995 16:50:40 -0700

Hi all,

We've known for a long time that the way CALMOD handles its command
line is different than the normal KUIP philosophy, and various
releases of CERNlib have tended to break CALMOD and/or its KUMACs in
subtle ways. I have found in the past couple of days that I must
compile CALMOD with CERNlib 94a or earlier to get it to work with the
standard calibration KUMACs. (With later versions, any command issued
from the ERP prompt rather than the CALMOD prompt does not process its
command line arguments correctly.) To get CALMOD to be a normal KUIP
application and still be able to process all the old KUMACs will
certainly be painful, and may even be impossible, and no one has been
willing to try in the couple of years we've known there is a problem.
In the meantime, I think ERP_CALIBRATOR should be modified to produce
KUMACs that can be executed by a CALMOD built with CERNlib 95a -- in
particular we should replace output like:

ERP
BOX 1B01
TICK 0A .2

with:

ERP BOX 1B01
ERP TICK 0A .2

BTW, apparently the FORTRAN standard requires all data statements
appear after all type declarations -- most FORTRAN compilers relax
this requirement, but to get FORTRAN programs to work on Linux we are
having to move data statements around. I will soon have a version of
CALMOD (2.53) with the data statements in the right place. Please be
aware of this requirement when writing collaboration software -- MACRO
now has at least 6 Linux boxes at 2 institutions.

Bob