Menu
▾
▴
[CIShell-SVN] SF.net SVN: cishell: [353] trunk/core/org.cishell.docs.intro/src
|
From: <bh...@us...> - 2006-12-18 14:31:50
|
Revision: 353
http://svn.sourceforge.net/cishell/?rev=353&view=rev
Author: bh2
Date: 2006-12-18 06:31:14 -0800 (Mon, 18 Dec 2006)
Log Message:
-----------
Updated the introductory paper per katy's requests
Modified Paths:
--------------
trunk/core/org.cishell.docs.intro/src/cishell+nwb.tex
trunk/core/org.cishell.docs.intro/src/cishell-short-text.tex
trunk/core/org.cishell.docs.intro/src/nwb-short.tex
trunk/core/org.cishell.docs.intro/src/research-team.tex
Modified: trunk/core/org.cishell.docs.intro/src/cishell+nwb.tex
===================================================================
--- trunk/core/org.cishell.docs.intro/src/cishell+nwb.tex 2006-12-17 17:18:02 UTC (rev 352)
+++ trunk/core/org.cishell.docs.intro/src/cishell+nwb.tex 2006-12-18 14:31:14 UTC (rev 353)
@@ -7,7 +7,7 @@
%\title{\includegraphics[width=160mm]{nwb-graphics/cishellnwb.png} \\
%\textit{Presentation Notes}}
\title{An Introduction to the Network Workbench and Cyberinfrastructure Shell \\
-\textit{Handout}}
+\textit{Introductory Paper}}
\author{Bruce Herr (bh...@bh...) & Weixia Huang (hu...@in...) \\
\small{School of Library and Information Science, Indiana University,
Bloomington, IN, USA}}
Modified: trunk/core/org.cishell.docs.intro/src/cishell-short-text.tex
===================================================================
--- trunk/core/org.cishell.docs.intro/src/cishell-short-text.tex 2006-12-17 17:18:02 UTC (rev 352)
+++ trunk/core/org.cishell.docs.intro/src/cishell-short-text.tex 2006-12-18 14:31:14 UTC (rev 353)
@@ -3,13 +3,13 @@
% \noindent
% \includegraphics[width=90mm]{graphics/cishellLogo.png}
-The Cyberinfrastructure Shell (CIShell) \cite{cishell} is an open source,
+The Cyberinfrastructure Shell (CIShell) \cite{cishell} is an open source
software specification for the integration and utilization of datasets,
-algorithms, tools, and computing resources. The Cyberinfrastructure Shell (1)
-supports algorithm writers to write and disseminate their algorithms in their
-favorite programming language while retaining their intellectual rights after
-distribution; (2) data holders to easily disseminate their data for use by
-others; (3) application writers to design applications from custom sets of
+algorithms, tools, and computing resources. The Cyberinfrastructure Shell
+supports (1) algorithm writers to write and disseminate their algorithms in
+their favorite programming language while retaining their intellectual rights
+after distribution; (2) data holders to easily disseminate their data for use
+by others; (3) application writers to design applications from custom sets of
algorithms and datasets that interoperate seamlessly; and finally (4)
researchers, educators, and practitioners to use existing datasets and
algorithms to further science.
@@ -33,55 +33,54 @@
%\end{wrapfigure}
Using CIShell, datasets and algorithms are bundled as CIShell-defined
-algorithms (see figure above). An algorithm is a black box that takes in
-zero or more datasets, some defined user parameters, and a so called CIShell
+algorithms, see figure above. An algorithm is a black box that takes in zero
+or more datasets, some defined user parameters, and a so called CIShell
Context. It then processes the inputs and returns an output in the form of zero
or more datasets. The CIShell Context provides access to services offered by
CIShell such as logging, preferences, GUI creation, and data conversion. The
-metadata dictionary associated with each algorithm can be used by applications
-to search with (see section 2.4). The metadata also comprises the author,
-citations, links to homepages, documentation, run-time complexity, etc. that
-ensures proper usage and citation.
+metadata dictionary associated with each algorithm can be used by applications,
+see section 2.4. The metadata also comprises the author, citations, links to
+homepages, documentation, run-time complexity, etc. that ensures proper usage
+and citation.
To give an example, a modeling algorithm takes in no data, some user-entered
parameters (e.g., the number of nodes to create a random graph) and returns a
-single dataset (e.g., a randomly generated graph). An analysis algorithm takes
-in some data and possibly some user-entered parameters, analyses the graph,
-then returns a new graph or simply prints out analysis results on the console.
-A visualization algorithm takes in some data, opens a new window visualizing
-the data, and typically returns no data. Each dataset is bundled as a ‘dataset
-provider’, i.e., an algorithm that takes in no data or parameters, but returns
-a dataset that is then loaded into the CIShell system.
+single dataset (e.g., a random graph). An analysis algorithm takes in some data
+and possibly some user-entered parameters, analyses the graph, then returns a
+new graph or simply prints out analysis results on the console. A visualization
+algorithm takes in some data, opens a new window visualizing the data, and
+typically returns no data. Each dataset is bundled as a ‘dataset provider’,
+i.e., an algorithm that takes in no data or parameters, but returns a dataset
+that is available to the CIShell system.
\subsection{Algorithm \& Dataset Integration Templates}
To ease the integration of algorithms and datasets, wizard-driven templates are
provided that acquire information from the algorithm writer and then generate
the appropriate files and resources. Templates are available to integrate
-arbitrary file-based datasets, compiled executable code, Java libraries, and
-Java-based algorithms. In the case of the Java-based template, after running
-the wizard, only one method needs to be filled in. This is the execute method for
-the actual algorithm. Integration of executable code typically does not require
-writing even one line of new code.
+arbitrary file-based datasets, compiled executable code, Java code, and Java
+libraries. In the case of the Java code template, after running the wizard,
+only one method, the execute method for the actual algorithm, needs to be
+filled in. Typically, the integration of executable code does not require
+writing one line of new code.
\subsection{Algorithm \& Dataset Distribution}
-Distribution of algorithms and datasets is made easy by using
-Eclipse \cite{eclipse} update sites and the NWB Community Wiki. Eclipse update
-sites allow an algorithm writer to package up their algorithms as features and
-place them on a webserver. Any algorithm writer can create his or her own
-private or public update site. Algorithm users can download any subset of
-datasets and algorithms from any number of update sites. He or she simply needs
-to select Help/Update from the menu and search for new features or updates of
-the currently installed features.
+Distribution of datasets and algorithms is made easy by using Eclipse
+\cite{eclipse} update sites. Eclipse update sites allow an algorithm writer to
+package up their algorithms and place them on a webserver. Any algorithm writer
+can create his or her own private or public update site. Algorithm users can
+download any subset of datasets and algorithms from any number of update sites.
+He or she simply needs to select Help/Update from the menu and search for new
+features or updates of the currently installed features.
The NWB Community Wiki provides an easy interface for algorithm writers and data
holders to post links to their algorithms and datasets, explain how to use
-them, and advertise their update sites (see section 4).
+them, and advertise their update sites, see section 4.
\subsection{Creating A Pool}
-Algorithms defined in section 2.1 can be located and run by using a
+CIShell Algorithms defined in section 2.1 can be located and run by using a
service registry that is provided by the underlying Open Services Gateway
Initiative (OSGi) technology \cite{osgi}. This service registry defines a pool
where services can be registered, searched for, and retrieved for use.
@@ -97,9 +96,8 @@
Once the algorithms are in the service pool (see figure above), CIShell
applications can easily search for algorithms based on their metadata. A
querying mechanism can be used by applications to find subsets of algorithms
-like all algorithms that belong in the visualization menu, all converter
-algorithms, etc. This is important for automatically converting data (see
-section 2.7).
+like all algorithms that belong in the visualization menu, all conversion
+algorithms, see also section 2.7.
\subsection{Creating An Application}
@@ -117,7 +115,7 @@
fashion. This inter-pool communication mechanism (see figure below) will also
support client-server applications in which computationally demanding
algorithms are run on servers and multiple CIShell clients can connect to it to
-use datasets, algorithms, and computing power.
+access datasets, algorithms, and computing power.
\begin{center}
\includegraphics[width=70mm]{graphics/connectingPools.png}
@@ -137,8 +135,8 @@
The service searches through the service registry for converter algorithms
(type=conversion in the algorithm's metadata) and builds a directed graph based
-on them in which edges are converters and nodes are file formats (see figure to
-the right). Edges can be weighted according to the 'losslessness',
+on them. Edges are converters and nodes are file formats (see figure
+to the right). Edges can be weighted according to the 'losslessness',
'complexity', or other features as reported in a converters' metadata.
When asked to find a chain of converters from format A to format G, it searches
@@ -151,15 +149,15 @@
\subsection{Brandable CIShell Reference GUI}
-The CIShell project includes a reference GUI that can be used directly or
-customized by application writers. The GUI (see figure below) is a menu-driven
-interface where all algorithms appears as menu items. Datasets can be loaded or
-simulated and are listed in the right-hand side data manager. When a user
-selects a dataset, the GUI determines if it can be used by any algorithms
-(either directly or through a series of converters). Only algorithms that can
-process the selected dataset are clickable, all others are grayed out. A
+The CIShell project includes a reference GUI that can be used directly or can
+be customized by application writers. The GUI (see figure below) is a
+menu-driven interface where all algorithms appear as menu items. Datasets can
+be loaded or simulated and are listed in the right-hand side data manager. When
+a user selects a dataset, the GUI determines what algorithms, either directly
+or through a series of converters, can be applied. Only algorithms that can
+process the selected dataset are selectable, all others are grayed out. A
console is provided to give information about the selected algorithms (authors,
-implementers, citations, etc.) and any output the algorithm logs when run.
+implementers, citations, etc.) and any output the algorithm logs when run.
\begin{center}
\includegraphics[width=2.24in]{graphics/cishell-using1.png}
Modified: trunk/core/org.cishell.docs.intro/src/nwb-short.tex
===================================================================
--- trunk/core/org.cishell.docs.intro/src/nwb-short.tex 2006-12-17 17:18:02 UTC (rev 352)
+++ trunk/core/org.cishell.docs.intro/src/nwb-short.tex 2006-12-18 14:31:14 UTC (rev 353)
@@ -3,10 +3,10 @@
% \includegraphics[width=90mm]{nwb-graphics/nwbLogo.jpg}
The Network Workbench (NWB) Tool \cite{nwb} is a large-scale network analysis,
-modeling and visualization toolkit for biomedical, social science and physics
+modeling, and visualization toolkit for biomedical, social science, and physics
research (see page 1). The NWB Tool rebrands the CIShell Reference GUI and
-provides a custom filling of datasets, algorithms, and converters for the
-network science community.
+provides a custom filling of datasets, algorithms, and converters relevant for
+the network science community.
%\\
%\\
%\noindent
@@ -18,15 +18,19 @@
%\subsection{Integrated Algorithms \& Datasets}
The current version of the NWB Tool gives easy access to 41 algorithms and a
-few sample datasets. About half of the algorithms were implemented (in FORTRAN)
+few sample datasets. About half of the algorithms were implemented in FORTRAN
and integrated by a physicist using CIShell's static executable algorithm
integration template. A listing of the currently integrated algorithms follows.
\begin{multicols}{2}{
\noindent
\begin{tiny}
-\textbf{Analysis}
+\textbf{Sampling}
\begin{list}{}{\setlength{\itemindent}{-7mm}\setlength{\itemsep}{0mm}\setlength{\topsep}{1mm}}
+\item Directory Hierarchy Reader
+\end{list}
+\textbf{ Analysis}
+\begin{list}{}{\setlength{\itemindent}{-7mm}\setlength{\itemsep}{0mm}\setlength{\topsep}{1mm}}
\item Node Degree
\item Node Indegree
\item Node Outdegree
@@ -51,10 +55,6 @@
\item CAN Search
\item Chord Search
\end{list}
-\textbf{Sampling}
-\begin{list}{}{\setlength{\itemindent}{-7mm}\setlength{\itemsep}{0mm}\setlength{\topsep}{1mm}}
-\item Directory Hierarchy Reader
-\end{list}
\textbf{Modeling}
\begin{list}{}{\setlength{\itemindent}{-7mm}\setlength{\itemsep}{0mm}\setlength{\topsep}{1mm}}
\item Barab\'{a}si-Albert Scale-Free Model
@@ -94,12 +94,12 @@
The Network Workbench Community Wiki \cite{nwbCommunityWiki} is a place for
users of the Network Workbench Tool, the Cyberinfrastructure Shell, or any
-other CIShell-based application to upload, download, and request algorithms and
-datasets. The site (see figure below) was created so that the network science
+other CIShell-based application to upload, download, and request datasets and
+algorithms. The site (see figure below) was created so that the network science
community can collaboratively create a tool which meets their needs and the
needs of the scientific community at large. Users can post want-ads for
algorithms and datasets to be integrated into CIShell/NWB, and learn how to use
-the resources available on the wiki for own their research.
+the resources for their own research.
\begin{center}
\includegraphics[width=2.85in]{nwb-graphics/nwbWiki.png}
Modified: trunk/core/org.cishell.docs.intro/src/research-team.tex
===================================================================
--- trunk/core/org.cishell.docs.intro/src/research-team.tex 2006-12-17 17:18:02 UTC (rev 352)
+++ trunk/core/org.cishell.docs.intro/src/research-team.tex 2006-12-18 14:31:14 UTC (rev 353)
@@ -8,7 +8,7 @@
\section{Overview}
The Network Workbench (NWB) project will design, evaluate, and operate a unique
-distributed, shared resources environment for for biomedical, social science,
+distributed, shared resources environment for biomedical, social science,
and physics research. Investigators are Katy B\"{o}rner, Santiago Schnell,
Alessandro Vespignani, Stanley Wasserman, Eric Wernert (Indiana University),
and Albert-L\'{a}szl\'{o} Barab\'{a}si (Notre Dame University). Major software
This was sent by the SourceForge.net collaborative development platform, the world's largest Open Source development site.
|
