>    my natbib. It would therefore be better to recommend the standard
 >    packages that one should  be able to use, such as the tools collection
 >    and amsmath. That is,  a paper using the model journal class could
 >    include all these packages  without having to submit them separately
 >    to the journal; he can assume they will be at any installation that
 >    contains journal.cls.
i strongly support the thinking behind this. an awful lot of problems
can be solved simply by conventions, like `everyone should have the
contents of CTAN:macros/latex/packages installed'. then its a question
of which packages to move into that area