question from a beginner

On Thu, 31 Jan 2002 hzi at uol.com.br wrote:

On Wed, Jan 30, 2002 at 06:49:00PM -0500, Ben Bolker wrote:

  No, not really.
  There are some facilities if you want to *design* GUIs for other people
to use, e.g. the Tcl/Tk extensions or Rweb (www.math.montana.edu/Rweb/),
but nothing where you can start from scratch with an arbitrary set of
data.  R is much more process-oriented: if you know what you want to do
(linear regression, ANOVA, etc.), it should be fairly easy to find the
commands that do what you want.  S-PLUS has GUI facilities, although I
haven't tried them.  I think there is a fairly strong philosophy among
most R developers that the "right" way to do things is to learn some of
the underlying foundations so that you can really understand what you are
doing.  It is easier to go wrong with GUIs -- although there are lots of
confounding factors, and GUIs can be designed in better or worse ways.

	As true as it what you've said may be, one sometimes wishes for a little GUI
functionality.
	If one __were__ to develop (say something with Tk), or any other contribution
in that sense, how could one distribute it)?
	Little tedious things like summary and histograms...

	Cheers,

	Henry L

I agree, some basic functionality in a menu would be handy, like,
for example, being able to print or save displayed graphics or geting a
window with an object browser. 

Agus

Dr. Agustin Lobo

On Sat, Feb 02, 2002 at 02:02:10PM +0100, Agustin Lobo wrote:

On Thu, 31 Jan 2002 hzi at uol.com.br wrote:

On Wed, Jan 30, 2002 at 06:49:00PM -0500, Ben Bolker wrote:

  No, not really.
  There are some facilities if you want to *design* GUIs for other people
to use, e.g. the Tcl/Tk extensions or Rweb (www.math.montana.edu/Rweb/),
but nothing where you can start from scratch with an arbitrary set of
data.  R is much more process-oriented: if you know what you want to do
(linear regression, ANOVA, etc.), it should be fairly easy to find the
commands that do what you want.  S-PLUS has GUI facilities, although I
haven't tried them.  I think there is a fairly strong philosophy among
most R developers that the "right" way to do things is to learn some of
the underlying foundations so that you can really understand what you are
doing.  It is easier to go wrong with GUIs -- although there are lots of
confounding factors, and GUIs can be designed in better or worse ways.

	As true as it what you've said may be, one sometimes wishes for a little GUI
functionality.
	If one __were__ to develop (say something with Tk), or any other contribution
in that sense, how could one distribute it)?
	Little tedious things like summary and histograms...

	Cheers,

	Henry L

I agree, some basic functionality in a menu would be handy, like,
for example, being able to print or save displayed graphics or geting a
window with an object browser.

Agus

Dr. Agustin Lobo

Dr. Lobo-

	XLisp-Stat has some very nice functionalities, like choosing to save
	the image in PostScript.

	It wouldn't be bad at all to have that in R...

I agree, some basic functionality in a menu would be handy, like,
for example, being able to print or save displayed graphics or geting a
window with an object browser.

Agus

Dr. Agustin Lobo

Dr. Lobo-

	XLisp-Stat has some very nice functionalities, like choosing to save
	the image in PostScript.

	It wouldn't be bad at all to have that in R...

It isn't bad at all that we have it on the GUI interfaces, then ....
It is on menus under Windows (surely the most used port of R) and I believe
under MacOS and GNOME too.

Brian D. Ripley,                  ripley at stats.ox.ac.uk