Peter Bailey

What I do

I am working on the WAR project as a researcher. We are doing lots of fun and exciting things at the nexus between information retrieval and Web search. I used to work in this area while trying to avoid doing my Phd a few years ago. Now I get to do it fulltime.

The other people in the project are Dave Hawking, Nick Craswell, and Francis Crimmins.

Over the last year, we've done a number of things, including but not limited to the following:

Developed a demonstration of our research work in Web search technology. We are currently providing search services for the whole of the ANU. Check this out at http://search.anu.edu.au/. The current name is P@NOPTIC search, though it used to be called S@NITY search. My contributions have been mainly on the front end UI, the crawler (a parallel version of GNU's wget), and integration with the PADRE indexer/query processor. Since the IP is all owned by ACSys, there are attempts to commercialise the software. We are soon to be providing Web search services for CSIRO's CMIS.
Organised the Web track for TREC, which we are doing again this year. We also participate in TREC each year, both in the Web track and others.
Wrote several papers, some of which have been published, and others have been submitted to conferences.
Supervised an Honours student investigating dynamic spidering techniques
Run the Document Technologies course for the Honours students. As a project, we got them to build various different components for an information retrieval system, the goal being to be able to plug and play different document scanners, indexers and query processors.

This year there is more of the same, and some new stuff as well. I'm particularly keen to do some solid time software engineering a new, robust and configurable Web crawler. We'd like to do this in Java, both to learn the language and because it should allow us to handle distribution and platform-independence in a nice fashion.

We are also committed to building a new Web test collection of 10gB of Web data for this year's Web track in TREC-9. There are some interesting challenges in constructing this, and it has to be done by March 2000.

Prior to working with the WAR project, I spent 18 months doing some very real-world programming. In search of industry experience after years at ANU, in 1997 I went and worked as a software engineer in Sydney for Object Technology International up until September 1998. Our team of four produced a product called Server Smalltalk, which comes as part of IBM's Visual Age for Smalltalk v4.5. I worked a lot on the distributed garbage collection and CORBA/IIOP components.

Research interests ...

1998-present

My main research interests presently are in building better retrieval systems for real world data. I also like to think about the nature of the Web in general, and how we can make more use of its structure to do a better job of retrieval. We've written a paper examining issues around what doesn't get found on the Web. Currently we are thinking about where to submit it to. We've also submitted papers about our experiences with building P@NOPTIC and some of the implications to SIGIR'2000 and Advances in Digital Libraries. Some of Nick's particular interests in server selection formed the basis of a paper submitted to the Digital Libraries conference as well.

1997-1998

I learned a whole lot about object oriented computing while at OTI, but there was little time to pursue individual research interests. I also developed a healthy interest in software engineering practices, rapid application development, guerrilla programming techniques, and distributed object technologies.

pre-1997

I finished my PhD in the Department of Computer Science, The Australian National University, in January 1997. I worked on the CAP project at the same time as doing my PhD just to keep busy.

My PhD research interests were in the development of an extension of ML called paraML, which explored parallel programming languages and models. The work involved looking at formal semantics for the extensions, building implementations of data parallelism, algorithmic skeletons, and object stores to illustrate the efficacy of paraML, and keeping up with SML/NJ compiler releases and developments with message passing systems such as MPI for the backend. I retain an interest in modern programming languages and semantics, but more from a practical viewpoint these days.

I spent happy times working with Dave Hawking on older versions of PADRE as well in these years. During those years, we were very keen on having entire text bases in main memory, since the AP1000 had 2gB of RAM. Nowadays of course, you can have that in your desktop PC, and we are trying to search hundreds of gigabytes of data.

A not completely up to date list of publications is available.

Administrivia: Contact Details

My official DCS home page

http://cs.anu.edu.au/personnel/staffDisplayNames.html?lastName=bailey

Email

Peter.Bailey@cs.anu.edu.au

Airmail, Voicemail, Faxmail

: Peter Bailey
: Department of Computer Science
: Australian National University
: Canberra, ACT, 0200, AUSTRALIA
: Phone: +61 6 249 3460
: Fax: +61 6 249 0010

Last modified: Sun Feb 13 16:10:29 EST 2000