Preamble: 

The paper was rejected quite consistently. It seems ICMT wasn't
exactly the right conference for grammar-related stuff. That is, it
looks like that the mere relevance flop would have lead to rejection
-- regardless of any quality/clarity issues that are also raised by
the reviews.

The reviews were quite critical in that they convinced us that we had
failed on two major fronts: (i) motivation of this work: why would
someone want a consistent Java standard. While it may have seemed
obvious to us, not just one reviewer did not confirm the significance
of this work. (ii) the reproducibility of the grammar transformations
wasn't described convincingly. How would we be able to achieve
objective results with regard to reported issues to be fixed by
transformations?

In addition to these major issues, there was some recurrency of
complaints about covering the grammar extraction part at all.
Further, we were asked to better relate our method (and results) to
other uses of the JLS standard. Also, some of the numbers that we were
using in describing the transformation effort (e.g., LOC or "number of
issues") were considered problematic (which relates to the
objectiveness critique mentioned above).

Conclusion: we decided to get as much as possible out of the ICMT
reviews and find a venue that is more interested in programming
languages, grammars, language standards.

Ralf Laemmel
6.5.2009

------------------------------------------------------------

Changes compared to the ICMT 2009 submission.

- The title has been changed to clarify the result.
- The abstract has been replaced.
- The introduction has been replaced. 
  * Formal style is removed.
  * Contributions have been clarified.
  * Significance has been described in some detail.
  * Mentioning of existing bug lists.
- The grammar extraction is gone (except for a short paragraph).
- Content on grammar transformation has been colleced in a designated section.
- Removed subsection: workflow of grammar convergence (not reproducible).
- New section: refined process of grammar convergence
  * Identified phases (all transformation scripts were refactored accordingly)
  * Present the idea of "difference reduction" as driving element
- Effort metrics for transformation: changes to the classification scheme.