In my view 3-d graphics are only useful when they convey a message that 2-d
graphs cannot.
Incidently, Nick mentions contour plots. That would be a very nice addition.
Matt
-----Original Message-----
From: Nick Cox [mailto:[email protected]]
Sent: Friday, January 17, 2003 8:57 AM
To: [email protected]
Subject: st: RE: Version 8: 3-D graphics?
Hoetker, Glenn
>
> First, let me say that the new graphics capabilities are stunning.
> Slow, but stunning. Thank you to everyone who contributed to them!
>
> I had hoped that the improvements might include decent tools for
> graphing 3D surfaces. Nothing fancy, just being able to show how Y
> varies over combinations of X1 and X2. There are a few
> user-contributed
> ado files, but nothing that easily produces
> publication-quality graphics
> (certainly not at the level that version 8 does for other types of
> graphics.)
>
> Is this something that the new graphics capabilities would
> make easy to
> develop (either by Stata or a user)? Or, am I going to
> have to deepen
> my acquaintance with GnuPlot (which is a great tool and
> only moderately
> user-hostile-I just prefer to limit the number of tools I
> have to use
> for a project).
>
Poor Stata Corp! No sooner have they
issued a new release packed with extraordinary
things that the wishlist for the next release(s)
starts anew.
So let me join in!
I endorse a call for 3-D graphics -- and I
echo the view, which I take to be tacit here,
that this means primarily graphics for data
triples, not bars or cones or whatever on
a sloping surface with arty effects. (Other
applications readily accessible to most Stata users allow
that kind of thing in any case.)
Speculating on how difficult this might be is
challenges Stata Corp to be open on how
difficult it really will be -- and to do the difficult
at one and the impossible next week.
My wild guess is this, and this is an inverted
pyramid of conjecture resting on a small base
of knowledge:
1. A nicer version of Bill Gould's -gr3-
(in my experience the most useful of the
non-official add-ons) will be relatively easy,
under a day's work for Stata Corp's graphics
programmers.
2. Perspective views of surfaces are nearer,
but a way off.
3. Contour plots are nearer, but a way further
off.
In both cases the algorithms exist, but someone's got
to code them. There's an extra issue whether
the coding really needs to be in C for speed.
I don't know, but that what place it beyond users.
1. certainly is ado-able. It's been ado-ne
before, of course.
Nick
[email protected]
*
* 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/
*
* 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/
