[R] R GUI For Which User?

Philippe Grosjean phgrosje at ulb.ac.be
Thu Jul 11 10:37:51 CEST 2002

Three points:
1) Let's try to be constructive. Stating "I am right and the other are wrong
(and incompetent)" is just a sterile accusation.

2) I am setting up a web page to introduce the various R GUI projects,
aiming to show the possibilities that are (and will be) offered to R users
looking for an alternative interface. So, please, if you are working on a R
GUI, contact me. I try also to clarify things and summarize ideas discussed
here, which leads to the third point:

3) I believe that your passionate discussion around CLI versus GUI is biased
by your vision of the GUI "a la Splus". However, there are other possible
GUIs. The basic aspects for a GUI are certainly the use of a graphical
device for output (the VGA, or more, screen) and for input (the mouse).
Typically, interaction with the program is event-driven: the user, not the
program, decides what is the following part of the code that will be
executed. In GUI,this is achieved that by selections in menus, dialog
boxes,... By the way, the question of who decide what is a key aspect. The
more freedom for the user, the more flexible is the interface. With CLI, all
the decision is up to the user. With GUI, the user's freedom could be more
or less limited, depending on the way the GUI is designed. A restrictive GUI
is also easier to learn, because it "tells" to the user what he has to do
next. The bad side-effect for data analysis is that a "lazzy" user lets
drive himself by the program without knowing what he is doing and still get
results, although these are most probably irrelevant. The most rigid GUI is
the 'assistant' (a series of dialog boxes to be filled successively: step 1,
step 2, ..., congratulation you did it [even if you are really, really
stupid]!). There are several other GUI concepts than the 'menu/dialog box'
or the 'assistant'. One is the 'notebook', like in Mathematica or Mathcad. A
notebook is a kind of rich-text document with embedded graphs and code. The
code seemlessly updates items in the text and in the embedded graphs when
the user request it. The 'spreadsheet' is another concept and it is
notorious for its inadequacy in serious data analysis. The 'object explorer'
is yet another concept. It can be easily married with a CLI because it is
complementary: it just helps managing variables in your workspace. An
excellent example of an object-explorer oriented GUI is Matlab (the latest
version 6.x). Several users noticed that the only part of the Splus GUI they
use is the object explorer. So, wouldn't an R object explorer be a good
addition to the current interface?

If the CLI is not suitable for all users and the Splus menus/dialog boxes
GUI is not the best concept, then, there is here an interesting field of
investigation for researchers, with experimentation of course. And by the
way, this research passes by the implementation of various GUIs. So, let's
do it for R, an excellent and free data analysis software. The various R GUI
projects, present and future, are certainly the concretisation of this


Philippe Grosjean

-----Message d'origine-----
De : owner-r-help at stat.math.ethz.ch
[mailto:owner-r-help at stat.math.ethz.ch]De la part de Frank E Harrell Jr
Envoye : jeudi 11 juillet 2002 04:20
A : Zed A. Shaw
Cc : r-help at stat.math.ethz.ch
Objet : Re: [R] R GUI For Which User?

On Wed, 10 Jul 2002 13:30:06 -0700
"Zed A. Shaw" <zedshaw at zedshaw.com> wrote:

> Hi All,
> I've been staying out of this discussion as I'm too busy actually
> implementing the GUI to argue over toolkits and languages or whether
> things easy for people ruins their education.  As a *student* I can tell
> that you are dead wrong if you think making something difficult makes it
> easier to learn.  The statement itself is just absurd.  As a researcher in
> Human Computer Interfaces and Educational Technology, I can tell you that
> you are wrong if you think making interfaces harder makes people smarter.
> What I find strange is how everyone on this list is most likely using R to
> do empirical research, and not single person has offered any empirical
> evidence to support their claims as to what makes a usable interface for
> students.  Practice what you preach!  If you claim that "students do not
> learn with a GUI", then back it up with some research.
> I guarantee that if you actually did some surveys and experimentation,
> find that you really don't understand student's needs.  Based on my
> experiences as a student, I can tell you point blank that the worst
> professors are the ones who know nothing about their students.  Likewise,
> the worst computer programs are the ones designed by programmers who know
> nothing about their users.  I intend to build a GUI that focuses on what
> students need, not on what professors think students need.
> Now, with that said, I'm going back to work.  Hopefully, with all these
> great *researchers* in here, I'll start seeing some *evidence* supporting
> what works in an R user interface.
> ----
> Zed A. Shaw
> University of British Columbia

With all due respect I feel that your opinions are somewhat naive.  I have
been doing and teaching data analysis for many years and spend a lot of time
watching students.  Students do not learn good habits by using GUIs and they
certainly do not learn how to do reproducible research that way.  You are
wrong in assuming that professors try to make something difficult and even
more wrong in assuming that students should avoid difficulties.  I shudder
to think what a GUI designed by a student for students will look like based
on what you've said.

Frank Harrell
> -.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.
> r-help mailing list -- Read
> Send "info", "help", or "[un]subscribe"
> (in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

Frank E Harrell Jr              Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine  http://hesweb1.med.virginia.edu/biostat
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch

More information about the R-help mailing list