RE: st: RE: RE: RE: Undocumented limitation of describe (and hence -ds-) in Stata 10?

["Nick Cox" <n.j.cox@durham.ac.uk>]

I don't worry that much about taxonomy, but 

r(varlist) is, or is equivalent to, a local macro in at least two
senses: 

1. You define it via -return local-. 

2. You can invoke it via `r(varlist)'. 

It is not a local macro in the sense that -macro list- won't list it.
-return list- will. 

Also, any r-class program will zap all existing r-class results. To zap
a local macro, you have to blank it out explicitly. 

Loosely similar points could be made about e() stuff. 

In Stata, I think the word "function" is best reserved for things so
named in the help for -functions-. Also calling -egen- functions
"functions" was arguably a misleading choice, but there are only so many
words that would fit well.  

Calling all sorts of other things functions -- not least commands, as
often done on this list -- does no great harm, but does not much help
communication either. 

The most congenial terminology is usually that of the software one
happens to know best! 

Nick 
n.j.cox@durham.ac.uk 

Michael I. Lichter

Thank you, Nick and Martin. I set myself up for a sucker punch in my 
last message.

Still, even if single quotes can evaluate lots of different kinds of 
things, and r(varlist) is not a local macro kind of thing, what kind of 
thing *is* r(varlist) and what are the rules governing its use? Is it a 
function returning the contents of a macro, or something else? (And what

are the other kinds of returns, like e(b), which acts more like a matrix

than like the contents of a matrix, and yet does not support
subscripting?)

Nick, thanks for the tip about your tips book. I do have a copy ... if 
only I could find it ... :(


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/