From: CSMR09 Date: Mon, Dec 8, 2008 at 5:31 PM Subject: CSMR09 notification To: Ralf Laemmel Dear Author, Thank you for your submission to the 13th European Conference on Software Maintenance and Reengineering (CSMR 2009) to be held in Kaiserslautern on March 24-27, 2009 We are pleased to inform you that your paper has been accepted as a short paper for presentation at CSMR 2009 and publication in the conference proceedings. Papers went through a rigorous reviewing process. Of 70 papers submitted, 21 were accepted as full papers and 10 more as short papers. Please find the reviews of your paper attached. You are expected to address the suggestions of the reviewers carefully, when preparing the final, camera-ready copy of the paper. Short papers are limited to 4 pages following the IEEE standards (http://www.computer.org/portal/site/cscps/menuitem.02df7cde46985ea21618fc2e6bcd45f3/index.jsp?&pName=cscps_level1&path=cscps/cps&file=cps_forms.xml&xsl=generic.xsl&). Your camera-ready paper is due on Friday, January 09, 2009. It is a requirement that at least one of the authors registers and presents the paper during the conference. Please provide a short notice to the program chairs, if you accept the invitation to present your short paper at CSMR 2009 until December 12, 2008. You will be receiving an author kit from Sheridan Printing shortly, which will provide further instructions on submission web site etc. Congratulations on having your paper accepted as short paper. We look forward to seeing you in Kaiserslautern. Sincerely, Jens Knodel (General Chair) Rudolf Ferenc, Andreas Winter (Program Chairs) --------------------------------------------- Paper: 65 Title: Java is not syntax-safe -- apparently -------------------- review 1 -------------------- PAPER: 65 TITLE: Java is not syntax-safe -- apparently Overall I foind this a poor paper. There is no motivaton as to what syntax safe means and why should the reader should care. I found the first part of the paper outlining details of extracting a Java grammar from html to be of some interest but I was not really sure if this adds to what the paper is about. Was there any place where the complete grammar of the various versions of Java were specified that did not involve interpreting what should be there in html and developing a whole set of rules to sort out an incomplete html specification. What do you mean by scraps grammars in the roadmap - section 1? In figure 6 does the number of lines refer to the number of lines in the grammar? If so then it is surprising that there is such a drop in the # lines in jls123? Is there an explanation for this? In the conclusions you talk about JLS required corrections in 21 locations. How is this related to safety? Also you talk about the size of transformation scripts. This is the first time that you mention this. Why is this an issue? -------------------- review 2 -------------------- PAPER: 65 TITLE: Java is not syntax-safe -- apparently Summary of Paper The paper describes an experiment in comparing published grammars for the Java language intended as both reference and implementation grammars, and analyzes the relationships between them, indicating the presence of a number of apparent unintended errors. The recovery of the grammars from HTML pages and their comparison using grammar convergence is described in some technical detail. Review I liked the idea of this paper, which seems to have clear practical applications and to my knowledge introduces some new ideas and techniques. However, although there is a lot of technical discussion, I was disappointed not to see concrete examples of the XBGF language/technology that was apparently used, and there is little about what exactly grammar convergence is and how it works. The related work section makes some claims about it which are unsubstantiated and difficult to believe without. Instead the paper spends a lot of time and space describing the lexical problems of recovering the specified grammars from the HTML source and analyzing the irregularities in that source. It is not clear what this contributes to the paper - these issues do not indicate differences in the specified grammars, which seems to be the point of the work. I don't think this analysis adds anything, and it delays the paper coming to the point. Many others have faced parsing HTML sources to identify or extract things. If this is an advance over their methods, please compare to them. If not, please shorten this section, there is no scientific advance in this. Some details: p.3 leafs -> leaves p.5 taken apart into several ones -> separated into its parts p.9 parsed like the “ocean” in the terminology of island grammars - I think you mean "water" or maybe "lake" - I have never seen "ocean" in island grammar terminology p.9 between sized grammars -> between large grammars The title is misleading. Something more like "Convergence of published Java grammars" might be more appropriate. Positive Points + reasonably well written + on topic + interesting new ideas with practical applications (validating reference grammars vs. implementation grammars - perhaps the paper should have explicitly said that?) Negative Points - too much emphasis on HTML page details, too little on the meat of the matter - would like to see some examples of the technology used -------------------- review 3 -------------------- PAPER: 65 TITLE: Java is not syntax-safe -- apparently This paper introduce the first case study, for the verification method named grammar convergence. My main concern with this publication is that the aforementioned method, is not validated by the scientific community. Consequently, i cannot be sure if the case study can actually be trusted or not. The original publication that describes the method is presented by the same authors in: * Ralf Lämmel and Vadim Zaytsev, An Introduction to Grammar Convergence, online at http://www.uni-koblenz.de/~laemmel/convergence/. which is submitted to a conference (could not find which one), and it is still (?) under review. I feel that the authors' decision to submit this paper to CSMR was premature, because their first paper was not accepted yet. For this sole purpose, I propose to reject the current publication. The paper itself, if one accepts the authors' method as valid, can be less than convincing. The authors write that they took many versions of the JLS found as HTML and PDF files, and resolved and captured all differences between them. Some details of the description like "tag elimination", which for example describes how the authors parsed the actual grammar rules from the HTML code, are not scientifically interesting. A large portion of the paper is covered with technical descriptions of trivial implementation details. The paper's most interesting concept is that the authors actually found some errors in the grammar productions. -------------------- review 4 -------------------- PAPER: 65 TITLE: Java is not syntax-safe -- apparently some further comments from the discussions on this paper: - please focus in your short paper version more on convergence (2nd half of the paper)instead of grammar recovery. - also reflect the title of your paper, which was viewed as misleading.