towards feff8.2 release

[Bruce Ravel <ravel@phys.washington.edu>]

One comment:

In reference to modular code, Alex said:

 ALA> but there is big disadvantage: it needs run scripts for every
 ALA> possible machine (=>portability problem).

This is just not true.

You need compilation utilities for every possible machine and
compilation utilities differ from machine to machine.  But once code
gets compiled to object libraries, it is small work to write an easily
portable program in C or python or perl that makes use of the object
libraries.

Let's not obfuscate the issue.  Properly designed libraries are easy
to port and easy to use in portable ways.  The hard part is making the
effort to see to it that the well-designed libraries actually compile
and install smoothly on many platforms.

So, in truth, modularity is no more or less portable than
monolithicity and it is disingenuous to claim that modularity is a
disadvantage.   What is true is that modularity is more work than
monolithicity because an API has to be defined, adhered to, and
maintained.

B


-- 
 Bruce Ravel  ------------------------------  ravel@phys.washington.edu
 U.S. Naval Research Laboratory, Code 6342      phone: (1) 202 767 5947
 Washington DC 20375, USA                         fax: (1) 202 767 1697
 WWW: http://feff.phys.washington.edu/~ravel --------------------------