Thanks for this. You raise an interesting --
although difficult to answer -- general question
on what is, or should be, regarded as core Stata
that everyone should (want to) know. There is one answer
in [U] 27.1, but as one fairly experienced user
I find that some of the commands there are ones
I have hardly ever used; that's more or less inevitable
with such lists. The idea that -graph- is optional
was presumably not intended, although it wouldn't
surprise me if indeed some users never produced graphs.
Although some people no doubt use macros for shortcuts as in
. local myxvar "x1 x2 x3 x4 x5"
. regress y `myxvar'
the main rationale for locals is for use within programs,
and not mentioning stuff to do with programs has long been
a criterion for most introductory Stata documentation.
Conversely, some books by Stata users include token chapters
on programming and I often wonder how effective such chapters are.
Conversely, again, this criterion confines -forvalues-
and -foreach- to the outer darkness of the Programming
manual.
More seriously, while StataCorp continues to make money out
of the manuals, and it's also important to keep individual
manuals to moderate size, there is no optimal solution to all
this.
My own prejudices on this are to some extent reflected in
the "Speaking Stata" columns of the Stata Journal.
Nick
[email protected]
George Stimson
> Thanks for the response and the reference.
>
> While I'm new to Stata, I've been using other advanced statistical
> packages for a couple of decades, so perhaps I was
> over-confident that I
> could just read what I thought were the relevant sections of
> the user's
> guide and reference manuals out of sequence. Of course, it
> would not be
> practical for Stata to repeat common warnings about such things as the
> distinction between left-quote and right-quote in every section of the
> manuals, so my approach of sampling individual sections carries risk.
> Since I was not using a local macro, I did not read the
> section on local
> macros you mentioned. I did, however, spend some time reading various
> other sections I thought would be relevant, doing searches
> and studying
> published do-files on the web. In every case, I just missed
> the obvious.
>
> I didn't mean my comment about the manuals to be a complaint. I like
> Stata very much and am finding the manuals to be quite clear
> generally.
> I would think in this case, however, that the importance of
> distinguishing between left and right quote marks might be highlighted
> in the "Getting Started" manual, since, as you said, "Many people have
> been bitten by this at precisely your stage." This is,
> however, just a
> quibble and not a serious complaint about a wonderful package or its
> documentation.
>
> Now I think I'm going to force myself to work through the User's Guide
> in sequence (at least skimming it), as I probably should have
> done from
> the beginning. : )
Nick Cox
> Many people have been bitten by this at precisely your
> stage, but it is wrong to blame the manuals. I find
> at [U] 18.3.1 (p.200 of Stata 9 edition) that the key
> difference between left and right single quotes is explained
> when local macros are introduced.
>
> Note also that Stata is not being awkward here. As references
> to local macros can be, and often are, nested, without some
> asymmetry of delimiters syntax like
>
> 'a'b'c'
>
> would be hopelessly ambiguous.
*
* 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/
