Specification

The specification for this project is very broad as the topic is research based. I am not going to attempt to pin down precisely what the project should achieve, but instead give an insight into the directions this project may take.

Aim

To investigate, understand and develop the ideas of Agent-oriented Parsing in the context of Empirical Modelling.

Possible Directions

Comprehension of the parser

• Improving documentation.
  • A technical document detailing the precise behaviour of the parser.
  • Write a tutorial for the parser using examples.
• Add visualisation + debugging features.
  • Optional debug information (in the same style as other tkeden debugging).
  • A visual deconstruction of the input string showing which agents are invoked and the parameters that are passed between them.
  • Animating the behaviour of the parser - A way to show the EM concepts behind the parser.
• Show how parsing process depends on the parsing rules (notation template).
  • An interface that allows direct manipulation of a notation.
• Introducing dependencies to notation templates.
  • For example, Chris Brown's palindrome parser.

Applications

• Reimplementing existing parsers.
  • Demonstrate benefits of implementing donald/scout/sasami using AOP.
• Implement new notations
  • Can complicated languages be implemented using our notation? For example, an XML parser. Maybe write a universal Turing machine to prove power of parser/eden.
  • Special purpose notations, for example a notation for railway building, that could be integrated with existing models.

Modelling applications

• Parser being used to build scripts (Higher order definitions…)
• Develop the parser as a multi-agent execution environment.
  • Visualisation used to demonstrate dependencies between agents.

Comparative study

• How does the agent-oriented parser relate to conventional parsing?