Hello, I am not currently using R but I have been using S+ for sevaral years. I think that I fit in the category of power users, never using the GUI prefering the flexibility of a command line interface (CLI). In a short while I will try to move to R, that is if we can convince our IT people that installing a freeware package on a Unix server will not damaged the companie's network! The main reasons for which I want to move to R, at least give it a try, are 1) David Smith of Insightful annonced us a couple months ago that there is not plan to further develop the 6.0.4 Release 1 for Sun SPARC, SunOS 5.8, 64-bit version of S+ and as we deal with very large problems, addressing well over 4GB of memory, a 64 bit version is required and we bought that Sun machine for this sole purpose. 2) From what I read and heard R manages its memory a little better than S+ and is a little faster, non negligeable aspects when dealing with large problems. All I said so far has little to do with the title of this email. I am getting to it. While I am pretty happy with a CLI I was asked to evaluate a few statistical packages to be used by casual users in my group. One of the main requirements is ease of use, hence the GUI, furthermore, it will be a lot easier for me to support them in their analyses if we use the same software. I spent the two last days reading the R-FAQ and searching the R groups mail archives, including the R-GUI mail group and as a consequence of all this reading I am a bit confused. At first I gathered that there was no GUI for R, that R was strickly a CLI language. What brought me to that conclusion is the fact that on the R-project pages they refer the readers to external builders of GUIs like the R-Commander of John Fox, Brodgar, TeXmacs, SciViews etc.. On the other hand searching the mail archives for the word "R-GUI" yielded well over 500 hits. After reading a few tens of those I realized that there existed, after building R, an executable called "Rgui.exe", what is that? is that a GUI version of R or not? can someone set my bearings straight on that issue? Thanks in advance for your valuable comments, G?rald Jean Analyste-conseil (statistiques), Actuariat t?lephone : (418) 835-4900 poste (7639) t?lecopieur : (418) 835-6657 courrier ?lectronique: gerald.jean at spgdag.ca "In God we trust all others must bring data" W. Edwards Deming
On Fri, 5 Dec 2003 11:32:30 -0500, Gerald.Jean at spgdag.ca wrote :>After reading a few tens of those I >realized that there existed, after building R, an executable called >"Rgui.exe", what is that? is that a GUI version of R or not? can someone >set my bearings straight on that issue?That's currently only on Windows. It's a GUI, but not in the sense you're looking for: it still needs external add-ons to get buttons, etc. to do most statistics. Duncan Murdoch
On Fri, 2003-12-05 at 10:32, Gerald.Jean at spgdag.ca wrote:> Hello, > > I am not currently using R but I have been using S+ for sevaral years. I > think that I fit in the category of power users, never using the GUI > prefering the flexibility of a command line interface (CLI). In a short > while I will try to move to R, that is if we can convince our IT people > that installing a freeware package on a Unix server will not damaged the > companie's network! The main reasons for which I want to move to R, at > least give it a try, are 1) David Smith of Insightful annonced us a couple > months ago that there is not plan to further develop the 6.0.4 Release 1 > for Sun SPARC, SunOS 5.8, 64-bit version of S+ and as we deal with very > large problems, addressing well over 4GB of memory, a 64 bit version is > required and we bought that Sun machine for this sole purpose. 2) From what > I read and heard R manages its memory a little better than S+ and is a > little faster, non negligeable aspects when dealing with large problems. > > All I said so far has little to do with the title of this email. I am > getting to it. While I am pretty happy with a CLI I was asked to evaluate > a few statistical packages to be used by casual users in my group. One of > the main requirements is ease of use, hence the GUI, furthermore, it will > be a lot easier for me to support them in their analyses if we use the same > software. > > I spent the two last days reading the R-FAQ and searching the R groups mail > archives, including the R-GUI mail group and as a consequence of all this > reading I am a bit confused. At first I gathered that there was no GUI for > R, that R was strickly a CLI language. What brought me to that conclusion > is the fact that on the R-project pages they refer the readers to external > builders of GUIs like the R-Commander of John Fox, Brodgar, TeXmacs, > SciViews etc.. On the other hand searching the mail archives for the word > "R-GUI" yielded well over 500 hits. After reading a few tens of those I > realized that there existed, after building R, an executable called > "Rgui.exe", what is that? is that a GUI version of R or not? can someone > set my bearings straight on that issue? > > Thanks in advance for your valuable comments,I will defer to others with more experience on your particular platform queries. Rgui.exe is the Windows 'front end' environment for R that provides a command line console for entering commands to the R interpreter, seeing the textual output of those commands and for displaying the output of plots. It supports a MDI/SDI type of interface. It does have some menus for simple operations (like installing and updating packages) but not for performing analyses. It is Windows OS only. The default R installation is a CLI interface. If you wish to add a GUI for performing "point and click" analyses, the sources that you list would be appropriate for consideration. If you require a cross-platform GUI for a multi-OS environment, John Fox's RCmdr would be a very good choice since it is built with Tcl/Tk and is not OS specific. HTH, Marc Schwartz
If you look on the R site, under "Related project" in the left frame near the bottom, you will find a list of various attempts at GUIs. (It seems to be down right now, but usually it works.) I've tried Rcmdr and Rweb. There are others in various states of development, some apparently not being developed. It is my impression from reading the mailing list that many of these are useful for what might be called canned analyses. That is, you want to make a data set available to lots of people (e.g., students, co-workers) who do not know much about R and do not want to learn, but they do want to do certain standard things like analysis of variance, and they want some flexibility in what variables they use, etc. etc. But I think the level of development is nowhere near what you would find in Splus or (what I used to use) Systat, where you could use the GUI for a lot (but not everything). FWIW, I find Rweb to be very useful. I use it with my students. You can execute actual R commands, or even a batch file of them, on an arbitrary data set (in the right format). It runs in any web browser. Although a Google search shows that a few others are using it, it is not being developed. The person who wrote it has not answered my email. (His web page says he travels a lot.) It is a bit limited because of its security features. Your sysadmin will like that, but others will want to disable them. I think Rcmdr has great potential too, but I gave up on it because its behavior is a bit dependent on the OS and monitor settings, so I felt I could not be as helpful to my students (being myself an oddball in both categories). BUT THIS IS OPEN SOURCE. If you like it and want to make it better, fix it and share what you do. I will do that with Rweb when I get some time. (Ha.) Jon -- Jonathan Baron, Professor of Psychology, University of Pennsylvania Home page: http://www.sas.upenn.edu/~baron
Dear Gerald, Although R doesn't come with a GUI, in the same sense as S-PLUS does, the tcltk package, which is part of the standard R distribution, provides facilities for building GUIs. The Rcmdr package is an example. Although it's not as extensive as the S-PLUS GUI, the Rcmdr package might satisfy the needs of the kinds of users who require a GUI (and it is extensible). I'm close to a new release of the package, which is described at <http://socserv.socsci.mcmaster.ca/jfox/Misc/Rcmdr/index.html>, where you can also find the development version. I hope that this helps, John At 10:47 AM 12/5/2003 -0600, Marc Schwartz wrote:>On Fri, 2003-12-05 at 10:32, Gerald.Jean at spgdag.ca wrote: > > Hello, > > > > I am not currently using R but I have been using S+ for sevaral years. I > > think that I fit in the category of power users, never using the GUI > > prefering the flexibility of a command line interface (CLI). In a short > > while I will try to move to R, that is if we can convince our IT people > > that installing a freeware package on a Unix server will not damaged the > > companie's network! The main reasons for which I want to move to R, at > > least give it a try, are 1) David Smith of Insightful annonced us a couple > > months ago that there is not plan to further develop the 6.0.4 Release 1 > > for Sun SPARC, SunOS 5.8, 64-bit version of S+ and as we deal with very > > large problems, addressing well over 4GB of memory, a 64 bit version is > > required and we bought that Sun machine for this sole purpose. 2) From what > > I read and heard R manages its memory a little better than S+ and is a > > little faster, non negligeable aspects when dealing with large problems. > > > > All I said so far has little to do with the title of this email. I am > > getting to it. While I am pretty happy with a CLI I was asked to evaluate > > a few statistical packages to be used by casual users in my group. One of > > the main requirements is ease of use, hence the GUI, furthermore, it will > > be a lot easier for me to support them in their analyses if we use the same > > software. > > > > I spent the two last days reading the R-FAQ and searching the R groups mail > > archives, including the R-GUI mail group and as a consequence of all this > > reading I am a bit confused. At first I gathered that there was no GUI for > > R, that R was strickly a CLI language. What brought me to that conclusion > > is the fact that on the R-project pages they refer the readers to external > > builders of GUIs like the R-Commander of John Fox, Brodgar, TeXmacs, > > SciViews etc.. On the other hand searching the mail archives for the word > > "R-GUI" yielded well over 500 hits. After reading a few tens of those I > > realized that there existed, after building R, an executable called > > "Rgui.exe", what is that? is that a GUI version of R or not? can someone > > set my bearings straight on that issue? > > > > Thanks in advance for your valuable comments, > >I will defer to others with more experience on your particular platform >queries. > >Rgui.exe is the Windows 'front end' environment for R that provides a >command line console for entering commands to the R interpreter, seeing >the textual output of those commands and for displaying the output of >plots. It supports a MDI/SDI type of interface. It does have some menus >for simple operations (like installing and updating packages) but not >for performing analyses. It is Windows OS only. > >The default R installation is a CLI interface. If you wish to add a GUI >for performing "point and click" analyses, the sources that you list >would be appropriate for consideration. If you require a cross-platform >GUI for a multi-OS environment, John Fox's RCmdr would be a very good >choice since it is built with Tcl/Tk and is not OS specific.----------------------------------------------------- John Fox Department of Sociology McMaster University Hamilton, Ontario, Canada L8S 4M4 email: jfox at mcmaster.ca phone: 905-525-9140x23604 web: www.socsci.mcmaster.ca/jfox
Gerald.Jean at spgdag.ca wrote:>Hello, > >... > >All I said so far has little to do with the title of this email. I am >getting to it. While I am pretty happy with a CLI I was asked to evaluate >a few statistical packages to be used by casual users in my group. One of >the main requirements is ease of use, hence the GUI, furthermore, it will >be a lot easier for me to support them in their analyses if we use the same >software. > > >One idea to archive easy use is to construct an editor from which you can start R-commands. If you have saved frequently used R-expressions in a text file you can load this file into such an editor, locate the correct formula and eval it. For this task I am developing an editor that uses the tcltk package of R: rwined(). Here is a screen shot: http://www.wiwi.uni-bielefeld.de/~wolf/software/revweb/rwinedwin.jpg For experimentation you can download the code (window and linux) from http://www.wiwi.uni-bielefeld.de/~wolf/software/revweb/rtrevive.exe after installation the package, you have to type: > library(rtrevive) > rwined() Peter Wolf