Richard Williams
> To clarify my concern - the -spost- routines have this painstaking
> customization where they look for specific program names, and
> if they don't
> find that name they don't work. So, for example, they might
> work fine with
> -prog-, but if -prog- gets renamed to -prog8- they won't
> work.
I can't speak for -spost-. But in general user-written programs
that depend on other user-written programs will indeed be broken
if those other programs become unavailable. But how would
this arise?
Suppose I announce on Statalist that my program
-foo- has been upgraded to exploit's Stata 9 new commands
and that the old -foo- is now -foo8- and frozen as is. If
you are still on Stata 8, you clearly should not replace
the -foo- you previously installed by -foo8-. No, the -foo8-
is just for people who see this and are interested in the
program but are still on Stata 8. If they want to
build something that calls -foo8- then that is
surely their concern. This is all one reason for this
numbering convention.
They may also be broken if the program name is adopted as
an official command name.
The solutions are various and include fixing the program.
In some cases copying the user-written program wholesale and
making it a subroutine is good practice, which I've often
recommended, subject to permission from the original author.
> But of course that is a potential
> issue with
> Stata's own updates, e.g. Stata might change something about
> -mlogit- that
> could zap -spost-'s ability to handle it.)
Similar comments. Often oldstyle stuff remains available
under version control.
More generally: the downside of user-written programs
is that other users have to be circumspect about what
they install, and aware that, for example, users
disappear or fail to fix their programs.
Nick
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/