Summary of meeting on 9th June 2005
The main issue discussed was regarding the re-implementation of ERA using a more appropriate framework, this is due mainly to the fact that the majority of rules in the ERA system are simply control flow rather than performing any actual inference. The current plan is to reimplement in php transcoding the rules at an appropriate point.
The system will be modular with a separate inference engine and jtms system all implemented in php based on the prolog prototype. The main problem identified with this approach is that due to the relatively short time period that this project spans there is a risk that delay in one part of the system could result in not having a deliverable for the actual dissertation
This is to be partly mitigated by allowing the inference engine and jtms modules to be deliverables in their own right. Secondly a timetable is to be constructed as a "living document" in order to try and control slippage as much as possible.
Also for next meeting a document demonstrating how the jtms is relevant to this problem should be produced including examples of inference chains within the prolog prototype. This should be partly based on the existing inference in ERA but should also include as much original inference chains as possible. This document can be used as part of a chapter in the final dissertation.
The other issue discussed was getting hold of a copy of "Building Problem Solvers" trying first an email to staff@inf and students@inf and then trying an inter-library loan
Next full meeting is scheduled for 30th June at 4:30pm with an interim meeting with Stephen in the interim.