Office:

Bell Labs, Lucent Technologies
700 Mountain Avenue, Room 2C-259
Murray Hill, NJ 07974-2070

E-Mail:

duncan@research.bell-labs.com

Phone:

office: (908)582-3217

Fax:

fax: (908)582-3340

Upcoming Events

Research

I work in the broad area of scientific computing. The primary focus is on applications to statistics, and the research is in software architecture and engineering for high-level languages and systems in this domain. As the field of statistics becomes both increasingly computational and inter-disciplinary, the need to be able to easily and flexibly

export statistical functionality to other fields,
access facilities from other fields within our programming environments, and
experiment with new algorithms, interfaces and paradigms for computing

is paramount.

My interests include high-level language design, parallel and distributed computing, meta-computing (i.e. computing on software itself), inter-system interfaces, data visualization, and software engineering.

Interactive Programming Environments

I am involved in the development of the S language and its available implementations: the S4 system underlying S-Plus and R. I work on the design, development and maintenance of the core or internals of each of these systems, and over the past several years, I and others have been experimenting with the development of a more extensible, embeddable, next-generation, interactive programming environment for statistics in the Omegahat project.

Add-on Packages (see Omegahat)

In addition to work on the core systems, I develop a collection of add-on packages or libraries that extend S-Plus, R and Omegahat. A list of most of the packages is available from the Omegahat site. These packages range in focus from inter-system interfaces which allow S and other languages such as Java, Perl, Python, XLispStat, Octave to be "seamlessly" combined in code; to S-language bindings for GUI toolkits (e.g. Gtk); to embedding S within spreadsheets relational database management servers (RDBMS) and Web browsers, distributed computing with SOAP or CORBA, (D)COM, XML and XSL facilities; documentation systems; meta-computing and automatic code generation. An underlying motivation is to explore and encourage ways in which statistics can be used by non-statisticians directly within their daily tools.

GGobi

I am one of the developers of the interactive, high-dimensional dynamically linked, visualization software GGobi. This is a highly customizable, component-based library that acts as a stand-alone application that supports plugins written in any one of several languages, and which can itself act as a plugin to other software such as S, Perl or Gnumeric.

Education: TILE and StatDocs

I also work on educational software for teaching statistics.

TILE

With Deb Nolan, I have developed TILE a Toolkit for Interactive Learning Environment that is both

a collection of interactive, thought provoking labs that teach different statistical concepts by allowing the student to explore the statistical concept by deductive reasoning; and
an extensible framework of Java classes that others can ue to build other labs.

StatDocs

Deb Nolan, myself and others are currently working on the StatDocs project which allows us to author dynamic and interactive documents both for pedagogical, and reproducible and updateable reports and research papers. This will culminate in an interactive textbook and authoring tools for creating these types of documents.

Ad Hoc Networks

I am also interested in communication networks and did some work on timescales within ad hoc networks with Raissa D'Souza and Sharad Ramanathan.

Up [ Stat. and Information Anal. Dept. | Math Center | Bell Labs | Lucent Technologies ]

Last modified: Fri Jun 13 21:45:02 EDT 2003