[ Jocelyn Ireson-Paine's Home Page | Invited talks | OxTALENT Events | OxTALENT Home Page]


Economics on the Web
Artificial Intelligence on the Web
A new Web authoring tool
Jocelyn Ireson-Paine

This page was prepared for the OxTALENT Open Morning on January 14th 1998, a demonstration of information technology in teaching at Oxford. I have updated it since, to take account of new things that I've done.


These programs are the result of a project carried out with the Institute for Fiscal Studies, to develop methods for connecting economic simulation programs to the Web. They allow Web users to change policy variables such as tax rates, and see the effect on UK households. The IFS was set up to improve public understanding of fiscal policy, and has had a Web site, with copies of all its research papers, since late 1994. Connecting these models was an obvious next step in improving public access. All the programs below are based on the IFS TaxBen microsimulation model for investigating the distributional effects of tax changes.

Working programs

Be Your Own Chancellor
The user changes tax rates and thresholds, VAT duties, and so on, and sees the effect on government revenue and on a range of sample families.

Budget 98
The user submits their own financial details, and sees how the last Budget will affect their income and expenditure. We put this and BYOC up at the BBC.

Articles and presentations

TOW: a system for teaching economics on the Web (with Graham Stark, Institute for Fiscal Studies).
In Proceedings of CALECO97, Bristol 25-26 September 1997. Downloadable as Zip file containing article in Word format.

TOW: a system for teaching economics on the Web (with Graham Stark).
HTML slide-style presentation on the above.

Virtual Economy: an integrated economic teaching aid (with Graham Stark).
HTML slide-style presentation on a successor project. Presented at CALECO98, Bristol 10-11 September 1998.

How to connect existing educational programs to the Web: a simple guide.
In Computers in Higher Education Economics Review for September 1997.

Economics comes on-line.
In Economic Review for September 1997.

Economics on the Internet.
In Guardian On-Line magazine of Thursday 23rd November 1995.

Artificial Intelligence

This work arises from the AI practical course at the Department of Experimental Psychology, which introduces students to classic symbolic AI and the history of AI on the one hand, and to nouvelle AI on the other. The software includes an agent-building kit written in Poplog, allowing students to construct agents to inhabit a simple microworld. The old version had a grid-world VT100 interface, but I'm experimenting with a Java replacement to make it more appealing visually. This will also let students use it from a Web browser on almost any computer, solving the problems caused by every college and departmental computer having different terminal emulation. With the old VT100 interface, it was impossible to give students instructions they could take away and guarantee to use in getting the emulation to work on any computer their college might provide, no matter what. The University ought to take some blame here, since it never appears to have tried making things easier by enforcing standards.

Working programs

The front-end doesn't yet work as an applet, because of the security restrictions applets impose in accessing remote files - ask to see me demonstrate it separately. Overcoming this requires writing a file-server via which the applet can take student programs written locally and save them remotely. This must also plug into a Java editor that the student uses locally, and must be able to send code to the Prolog back-end. It is almost complete.
I'm experimenting with Mark Tacchi's Gamelet toolkit as the alternative environment for the agents. The above is a simple shoot-'em-up game he wrote as a demo.

Another Gamelet demo - some bouncing balls.


Using Java and the Web as a front-end to an agent-based Artificial Intelligence course.

Course notes

Course notes written for the pre-Java version. They lead a simple classical AI program through its tricks, showing how it builds up and analyses different representations.

Production Systems (1)
An introduction to production systems and some ideas of nouvelle AI.

Production Systems (2)
More on production systems.

About the pre-Java microworld, called Eden. The original documentation, for Pop users, is here.

Student notes
A page containing links to all the practical notes, as well as my lectures and tutorials.

Downloadable code

Gamelet can be downloaded from here. Unfortunately, there doesn't seem to be any archive that's committed to keeping or maintaining it, so this link may disappear.

Web authoring tool

Web-O-Matic/Java, or WOM for short, is a tool for writing server-side interactive Web pages. It avoids the Web page author needing to know about CGI protocols and how servers work; instead, pages can be written in an extended dialect of HTML, with Java code inserts to control interaction. You can easily write session-based applications such as shopping services, and forms that save data to a file for a program to read. WOM is written in Java and generates Java code, and comes with its own Web server. It should run on any machine with a Java implementation. It is based on a new programming method, System Limit Programming, involving some fairly advanced mathematics (category theory and sheaves) and a new way of looking at the concept of objects.


Web-O-Matic/Rexx: a tool for building complex interactive Web applications by compiling HTML to Object Rexx .
In Proceedings of 8TH REXX SYMPOSIUM, Heidelberg 22-24 April 1997.

Web-O-Matic: using System Limit Programming in a declarative object-oriented language for building complex interactive Web applications .
In Proceedings of DOCUMENT PROCESSING AND SGML, Workshop of project DAVID - Algebraic Document Processing Braga 2-4 September 1996.

Downloadable code

Web-O-Matic is free for academic use, and can be downloaded via my Web-O-Matic page.

[ Jocelyn Ireson-Paine's Home Page | Invited talks | OxTALENT Events | OxTALENT Home Page]