Well, many hands are involved here, and who to "blame"
is an open question.
As the help for -glcurve- documents, I had a hand
in its last revision, and it may well have been
me who wrote that line, as I tend to code using
-legend(order())- and not -legend(label() label() ... )-.
Then Philippe and Stephen Jenkins are the authors,
so there they are.
Then there is whoever wrote that line in the manual.
Writing stuff like "considered better style" is
great fun when you are producing documentation.
I've done similar things myself, but who is thrusting their
prejudices upon us and quite what canons are being
invoked and why? Probably Vince Wiggins....
Talking of "blame" is facetious here, as you may
have gathered. Concretely, next time -glcurve-
is publicly revised, the point is probably worth a tweak
to the code or a comment in the help.
Nick
[email protected]
philippe van kerm
> >From: "Nick Cox" <[email protected]>
> >Philippe underestimates his own program!
> >
> >One of the small problems here is that -glcurve-
> >internally uses -legend(order())-, but to
> >overwrite that you just need to use the
> >same. At the same time you can label
> >the extra diagonal and fix the line patterns.
> Well spotted Nick!
>
> I'm not sure now why it was coded this way in -glcurve-, but I read in
> -help twoway_legend- that
>
> "You may also specify quoted text after # to override the descriptive
> text associated with a symbol. (...) It is considered better style,
> however, to use the label() suboption to relabel symbols."
>
> So -label()- would have been a more proper way to proceed. One lesson
> is that, when both order() and label() are used at the same time, the
> former takes precedence.
>
> Thanks for pointing out the solution.
*
* 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/
