Review of Fulib by A. Idani

[meeduse]

Positive points :

1. Good performance
2. All generated cases are covered

Negative points :

1. Seems like a Java-based implementation of the transformation. The author mentions that expandLink, addColumn, and addRow are appropriate model transformations. However, they are also java methods. Note that I didn’t find the implementation of expandLink. My main question is: is the author evaluating the performances of a Java-based transformation ?
2. Validation is done only via the validator jar file (no defined testing strategy)
3. FulibSolution.java : more than 300 lines of java code with a poor documentation.
4. What about truth tables with non-filled cells like that in the call for solutions document? How this solution could be evolved in order to cover this case?