I don't know about "must be": "should be",
perhaps, but then again "should not be",
perhaps.
It is always tricky even for StataCorp to
know what should be an FAQ. The difference
between copying and evaluating is quite explicit
in [U] 21.3.4; this is presumably where most
Stata users first learn about macros. Perhaps
it needs a stronger warning: pay attention, as
otherwise this really will bite you!
In addition, the limits on various operations are
at [R] limits and in the corresponding on-line help.
So all this is documented fully.
The difficult trade-off here is based on some
elementary but also fundamental facts:
1. Most people, naturally, don't absorb
all the information they need from a single
reading. So, it helps to have multiple
sources of information and some repetition.
2. But given the massive manuals,
on-line help, FAQs, etc., etc. there is
also, in an equally obvious way, already
too much for many people to (want to) scan
for an answer. (A minority use of Statalist
seems close to "I can't be bothered to read
the manuals, so can somebody give up their
spare time to do my work for me?".)
The FAQs grew up as extras, plugging gaps
in the documentation, expanding on tricky
points, etc. They should not be used
just to repeat information. That just
bloats the available materials.
Another basic tip is just to re-read
the manuals, cover to cover, from time
to time. (You are allowed to skip and
skim, when you know that a command is of
no present interest to you.)
You may blench at the time investment, but
if you use Stata almost every day, it is
likely to be repaid.
Nick
[email protected]
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]]On Behalf Of Michael S.
> Hanson
> Sent: 16 July 2004 15:11
> To: [email protected]
> Subject: Re: st: Tests on sums of coefficients?
>
>
> On Jul 12, 2004, at 2:51 PM, Nick Cox wrote:
>
> > Be very careful here. As numerous threads bear
> > witness, looping around
> >
> > local testsum = "`testsum'+`var'"
> >
> > will truncate your macro at some limit. It
> > is much safer to copy, not evaluate:
> >
> > local testsum "`testsum'+`var'"
>
> Thanks for that suggestion -- it might help explain a
> problem I have
> seen cropping up with a similar construct.
>
> I had searched the archive for similar questions before
> posting the
> original question, but found none like you refer to. I guess I had
> been looking the wrong way....
>
> Thanks again for taking the time to respond to what
> must be a FAQ.
*
* 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/
