Last class we ploughed into eGanges in greater depth, with Ms Gray leading us through an example of building an eGanges Map using International Accounting Standards (IAS) 24 which concerned related party disclosures. This exercise really illustrated the fact that a map is dependent on how the user interprets the text and how they choose to build it. The river paradigm means that all the nodes (tests at each node) forms a part of the flow. Thus it is possible that if one node was incorrect (either due to incorrect interpretation or typographical error, eg typing “disclosed” instead of “undisclosed”), the outcome is likely to be incorrect or at least the quality of the outcome would be adversely affected.
I think using a colour scheme in the river hierarchy such that the colours represented the depth of the map is a good idea so users can tell the different branches of the map apart. Once a user knows the colour scheme, they will also know which branch is the primary flow and also how deep in the map they are. eGanges also has extensive functionalities for glosses and notes. They can be in the form of positive, negative and uncertain lists, or they may be a link to another node, a website, a database or other software.
We only covered a small rule in building the map in class. I can only imagine what a task it would be to create a map for a full piece of complex legislature. There will be a lot of effort involved, and it may require a few runovers to check its accuracy. Modelling case law would be an even bigger and more problematic task as the “rules” (if you can call them rules) from cases will not be so clear cut and nor will you necessarily be able to provide it with a straight answer of yes/no.
One area I think needs improvement is the user interface. I didn’t find it very attractive nor aesthetically pleasing, a view shared by fellow classmates Sachin and Lanny. There are quite a lot of buttons crammed into the bottom lefthand corner, and I remember at one stage during the demonstration of building a map, Ms Gray’s assistant tried several times to give input to a node but the system did not react. The reason was because he was still in ‘build’ mode. Perhaps a tool-bar at the top of the screen with all the commonly used features would make buttons easier to find, and a status box showing what mode the system is in might help make the system clearer.
Overall, I feel like I have a better understanding of what eGanges seeks to achieve and how it works after seeing it in action. I see that it can model rules or processes and act as a way of checklisting all relevant steps are satisfied, as well as providing a possible conclusion/outcome. I guess now we’ll just have to wait and see if such expert system shells will take off amongst the legal profession.