Bluntly, this depends -- partly -- on how far one cares
about making one's programs available to others.
Richard has himself put stuff on SSC, so he does
care. I also put stuff on SSC, so I care, although
in practice life is sometimes too busy to accommodate
all possible users.
Thanks Nick, Ben, Scott and others. I imagine this open system where
anyone can write their own programs creates headaches for Stata as it tries
to maintain compatibility across versions. I wonder if you really can be
sure that using the version number command will maintain
compatibility? Suppose, for example, I write something today in 8.2, but
it calls an ado file that gets changed in 9.0? I imagine Stata works hard
to avoid problems with the stuff it releases, but user-written routines
might not be so careful about not breaking something. Nick cautioned me
that the seemingly minor changes I made in my -pcorr2- command might not
automatically be embraced as an official part of Stata, and I'm beginning
to understand why.
