- it only relies on the Glamour mechanisms without state in the implementation code. This was more of a validation that the engine is expressive enough

- it has a slightly different layout

- the Mondrian map can be used as a navigation, so clicking on a node will focus the browser on that production.

- it uses a tree widget for the debugger so we can scale for large debugging traces

What is still to do:

- install the associated instance variable for each newly created production.

- introduce parser renaming refactoring. This would entail triggering both a method and an instance var rename.

- add the possibility to remove a production. This would entail removing a method and the corresponding instance var.
- make the morph graph view interactive.

- add a new grammar class.

- integrate tests for each production. This would involve being able to browse tests associated to a production, run, add and remove them.

If these would work, we would have a true IDE and for most parts we would not need the Smalltalk editor for building parsers.

But, I could use some help on the todos left, especially on the refactorings.

Cheers,

Doru