Dear Modules:

I wrote/maintain a perl-based thing named NMQual
(http://nmqual.googlecode.com).  It creates value-added installations
of NONMEM(r), a licensed fortran program for fitting nonlinear mixed
effects models.

One problem I have is that my scripts don't know unambiguously where
to find my subroutines (usually a .pm file in same directory).  "use
lib $FindBin::Bin;" seems to be guessing, and "rewriting" the scripts
on an installation-specific basis seems to be a hack.

It occurred to me that perhaps I should communicate my routines
through CPAN, and let perl just find them on @INC.  If that is a good
idea, please comment on appropriate module name.

Googling "nmqual" returns a first page with all-appropriate hits.
Should I consider NMQual (Nmqual?) as a top-level namespace?  More
than half my routines are convenience abstractions, borrowing heavily
from File::Spec.  A large chunk implements patching with context-style
diffs in a platform neutral way (relies on Text::Patch) so that I
don't have to worry about Windows users not having the system patch
utility (although I may have to worry about ActivePerl users not
finding my package).

Thanks,

Tim

Reply via email to