[rkward-cvs] SF.net SVN: rkward:[3283] branches/jss_dec_10
sjar at users.sourceforge.net
sjar at users.sourceforge.net
Mon Dec 20 22:55:41 UTC 2010
Revision: 3283
http://rkward.svn.sourceforge.net/rkward/?rev=3283&view=rev
Author: sjar
Date: 2010-12-20 22:55:41 +0000 (Mon, 20 Dec 2010)
Log Message:
-----------
* minor corrections
* some references
* Conlusion ... expanded
Modified Paths:
--------------
branches/jss_dec_10/FINAL_JSS_TEX/GUI_elements.tex
branches/jss_dec_10/FINAL_JSS_TEX/background.tex
branches/jss_dec_10/FINAL_JSS_TEX/conclusion_summary_acknowledgments.tex
branches/jss_dec_10/FINAL_JSS_TEX/example_session.tex
branches/jss_dec_10/FINAL_JSS_TEX/sources.bib
branches/jss_dec_10/commented_versions/rkward_merge.odt
Modified: branches/jss_dec_10/FINAL_JSS_TEX/GUI_elements.tex
===================================================================
--- branches/jss_dec_10/FINAL_JSS_TEX/GUI_elements.tex 2010-12-20 19:30:40 UTC (rev 3282)
+++ branches/jss_dec_10/FINAL_JSS_TEX/GUI_elements.tex 2010-12-20 22:55:41 UTC (rev 3283)
@@ -384,8 +384,10 @@
capturing and documenting \proglang{R} output can also
be used in RKWard, RKWard provides a dedicated output file and output
window for the documentation of results. All GUI-driven data handling
-functions (see Section~\ref{sec:analyzing_data}) write their output to this file. It is also
-possible to write to the output from \proglang{R}
+functions (see Section~\ref{sec:analyzing_data}) write their output to this file.
+The output is presented as journal view. Therefore all results are presented
+sequentially with the most current performed task on the bottom.
+It is also possible to write to the output from \proglang{R}
scripts by using a number of dedicated \proglang{R}
functions. For the GUI-driven data handling functions, the output is
standardized to include the name of the feature, the date and time of
Modified: branches/jss_dec_10/FINAL_JSS_TEX/background.tex
===================================================================
--- branches/jss_dec_10/FINAL_JSS_TEX/background.tex 2010-12-20 19:30:40 UTC (rev 3282)
+++ branches/jss_dec_10/FINAL_JSS_TEX/background.tex 2010-12-20 22:55:41 UTC (rev 3283)
@@ -3,8 +3,8 @@
In mid 1993 Ihaka and Gentleman published initial efforts on the computing
language and programming environment \proglang{R} on the s-news mailing list. Ambitions for
this project aimed to develop an \proglang{S}-like language but without inheriting memory
-and performance issues. The source code of R was finally released in 1995 and
-development has since evolved under the umbrella of the R Development Core Team
+and performance issues. The source code of \proglang{R} was finally released in 1995 and
+development has since evolved under the umbrella of the \proglang{R} Development Core Team
since mid 1997 \citep{RDCT2001, RDCT2010, Ihaka_Gentlemen_1993}.
\proglang{R} does not include an advanced cross-platform GUI (graphical user interface) as known from other
statistical software packages. However, \proglang{R} includes tools for building GUIs
@@ -43,7 +43,7 @@
specifically aims to hide the complexities of the \proglang{R} programming language by using concept of visual data-flow
programming \citep{Sutherland1966}.
}. In doing so RKWard accepts relatively verbose
-generated code, deliberately. In particular, RKWard limits itself to generate R
+generated code, deliberately. In particular, RKWard limits itself to generate \proglang{R}
code from GUI settings. It avoids wrapping complex sequences of data
manipulation or analysis into custom high-level \proglang{R} functions. The task of
providing high-level functions is logically independent of the development the
@@ -55,7 +55,7 @@
While RKWard tries to support users wishing to learn \proglang{R}, it is specifically not
designed as a teaching tool (such as \pkg{Rcmdr} or \pkg{TeachingDemos}), but as
a productive tool. Dialogs for statistical procedures in RKWard do not
-necessarily show a 1:1 correspondence to the underlying steps in R, but are
+necessarily show a 1:1 correspondence to the underlying steps in \proglang{R}, but are
rather oriented at statistical tasks. Furthermore, RKWard does not impose
artificial limitations on how users can work with the application. For example,
the user is not limited to using only one \code{data.frame} or one model at a
Modified: branches/jss_dec_10/FINAL_JSS_TEX/conclusion_summary_acknowledgments.tex
===================================================================
--- branches/jss_dec_10/FINAL_JSS_TEX/conclusion_summary_acknowledgments.tex 2010-12-20 19:30:40 UTC (rev 3282)
+++ branches/jss_dec_10/FINAL_JSS_TEX/conclusion_summary_acknowledgments.tex 2010-12-20 22:55:41 UTC (rev 3283)
@@ -3,7 +3,25 @@
In this article we have introduced the RKWard GUI to \proglang{R}. RKWard provides features ranging
from easy to use dialogs for common statistical procedures, targetted at \proglang{R} novices, to advanced
IDE features targetted at \proglang{R} experts.
-
+The RKWard development
+does not focus on \proglang{R} package development, except those internally
+required for RKWard, but keeps it at the \proglang{R} community. This design brings the intrinsic
+benefit of highly accurate results since calculations entirely rely on \proglang{R} code.
+Comparison of the commonly used spreadsheet applications
+regarding estimation, random number generation and statistical distributions revealed serious
+limitations. \proglang{R} in contrast was found to be a reliable and accurate statistical
+software package \citep{Almiron2009, Almiron2010}.
+\proglang{R} scripts which are routinely in use or require complex steps can be programmed as GUI
+elements in RKWard. Due to its plugin architecture RKWard provides the basis
+for a highly customizable \proglang{R}-GUI. We show that the plugin development can take place
+independently without changes in the core application or compilation of binaries.
+Current developments also targets plugin development via external canals like
+GHNS (Get Hot New Stuff)\footnote{\url{http://ghns.freedesktop.org/}} independently of the
+RKWard plugin development.
+Basically all \statistical \proglang{R} functions can be provided with GUIs.
+The exclusive use of standard-complient formats (e.g. \proglang{HTML}, PNG)
+and sole use of \proglang{R} formats (\proglang{R} Data Files \citep{RDCT2010c})
+guarantees cross-platform interoperability.
Future versions of RKWard will continue to add value for both groups of users. Planned features include
an enhanced interface for debugging \proglang{R} code, support for editing more types of data, and the
ability to connect the RKWard GUI to a remote \proglang{R} engine. Perhaps most importantly, RKWard will
Modified: branches/jss_dec_10/FINAL_JSS_TEX/example_session.tex
===================================================================
--- branches/jss_dec_10/FINAL_JSS_TEX/example_session.tex 2010-12-20 19:30:40 UTC (rev 3282)
+++ branches/jss_dec_10/FINAL_JSS_TEX/example_session.tex 2010-12-20 22:55:41 UTC (rev 3283)
@@ -18,10 +18,13 @@
example ``comma'' and ``period'' were chosen via ``Quick mode'' as field
separator character and decimal point character respectively.
+The corresponding \proglang{R} code would read as:
+
\code{read.csv(file=/media/software/experiment.txt,
na.strings = NA, nrows = -1, skip = 0,
check.names = TRUE, strip.white = FALSE, blank.lines.skip = TRUE)}
+
Checking the ``Edit Object'' box will automatically open a data editor tab
showing the imported data (Figure~\ref{fig:import_data}B).
Modified: branches/jss_dec_10/FINAL_JSS_TEX/sources.bib
===================================================================
--- branches/jss_dec_10/FINAL_JSS_TEX/sources.bib 2010-12-20 19:30:40 UTC (rev 3282)
+++ branches/jss_dec_10/FINAL_JSS_TEX/sources.bib 2010-12-20 22:55:41 UTC (rev 3283)
@@ -1,6 +1,34 @@
% This file was created with JabRef 2.6.
% Encoding: UTF8
+ at article{Almiron2009,
+author = {Marcelo G. Almiron and E. S. Almeida and M. N. Miranda},
+title = {The reliability of statistical functions in four software packages freely used in numerical computation},
+journal = {Brazilian Journal of Probability and Statistics},
+year = {2009},
+volume = {23},
+pages = {107-119},
+number = {2},
+doi = {10.1214/08-BJPS017}
+}
+
+ at article{Almiron2010,
+ author = {Marcelo G. Almiron and Bruno Lopes and Alyson L. C. Oliveira and Antonio C. Medeiros and Alejandro C. Frery},
+ title = {On the Numerical Accuracy of Spreadsheets},
+ journal = {Journal of Statistical Software},
+ volume = {34},
+ number = {4},
+ pages = {1-29},
+ day = {12},
+ month = {4},
+ year = {2010},
+ CODEN = {JSSOBK},
+ ISSN = {1548-7660},
+ bibdate = {2010-01-31},
+ URL = {http://www.jstatsoft.org/v34/i04},
+ accepted = {2010-01-31},
+}
+
@PHDTHESIS{Sutherland1966,
author = {William R. (Bert) Sutherland},
title = {The On-line Graphical Specification of Computer Procedures},
@@ -183,7 +211,7 @@
}
@Article{Gentleman2004,
-AUTHOR = {Gentleman, Robert and Carey, Vincent and Bates, Douglas and Bolstad, Ben and Dettling, Marcel and Dudoit, Sandrine and Ellis, Byron and Gautier, Laurent and Ge, Yongchao and Gentry, Jeff and Hornik, Kurt and Hothorn, Torsten and Huber, Wolfgang and Iacus, Stefano and Irizarry, Rafael and Leisch, Friedrich and Li, Cheng and Maechler, Martin and Rossini, Anthony and Sawitzki, Gunther and Smith, Colin and Smyth, Gordon and Tierney, Luke and Yang, Jean and Zhang, Jianhua},
+AUTHOR = {Robert Gentleman and Vincent Carey and Douglas Bates and Ben Bolstad and Marcel Dettling and Sandrine Dudoit and Byron Ellis and Laurent Gautier and Yongchao Ge and Jeff Gentry and Kurt Hornik and Torsten Hothorn and Wolfgang Huber and Stefano Iacus and Rafael Irizarry and Friedrich Leisch and Cheng Li and Martin Maechler and Anthony Rossini and Gunther Sawitzki and Colin Smith and Gordon Smyth and Luke Tierney and Jean Yang and Jianhua Zhang},
TITLE = {Bioconductor: open software development for computational biology and bioinformatics},
JOURNAL = {Genome Biology},
VOLUME = {5},
@@ -542,6 +570,16 @@
url = {http://www.koffice.org/kword/}
}
+ at MANUAL{RDCT2010c,
+ title = {R Data {Import/Export}},
+ author = {{R Development Core Team}},
+ organization = {R Foundation for Statistical Computing},
+ address = {Vienna, Austria},
+ year = {2010},
+ note = {{ISBN} 3-900051-10-0},
+ url = {http://www.R-project.org/}
+}
+
@comment{jabref-meta: selector_publisher:}
@comment{jabref-meta: selector_author:}
Modified: branches/jss_dec_10/commented_versions/rkward_merge.odt
===================================================================
(Binary files differ)
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
More information about the rkward-tracker
mailing list