Type-preserving compilation for large-scale optimizing object-oriented compilers

摘要

The Program Committee and I are pleased to present the proceedings of the ACM SIGPLAN 2008 Conference on Programming Language Design and Implementation. This volume contains 34 papers selected out of 184 submissions. We believe this volume demonstrates that our field is vibrant, influential and expanding. We hope you find these papers innovative, informative, and exciting. >In consultation with the program committee, I also selected two additional external reviewers. All together 837 reviews were collected from the PC and 371 external reviewers. In order to maintain the quality of this premier conference, it is important to get detailed and thoughtful reviews from the experts in the community. While arranging all these reviews was a nightmare, I am extremely appreciative of all of you who took on the review task and provided the PC with many important insights. Your reviews made a big difference and were influential in most decisions during the PC meeting. >We used a double-blind review process to reduce the reviewer bias against authors or institutions. However, we recognized that when authors build upon their previous work, double blind reviews can leave the reviewer without sufficient information to do a credible review. Thus, we asked the authors to submit a separate list of URLs for their own related work and indicate in the submission that the citations were available from the PC chair. In the instructions to the reviewers, I asked them to do a first read and then contact me if they felt the need to look at the omitted material. About 10 reviewers requested this information. >Another issue that came up during the process was double submission. On more than one occasion we found that the same or a similar paper was submitted to a different conference with an overlapping review period. In all these cases, in accordance with the SIGPLAN policy, we informed the other PC chair and summarily rejected the paper. One exception was the co-located International Symposium on Memory Management (ISMM), whose review period overlapped with PLDI. After a long discussion, we allowed co submission to both PLDI and ISMM with the explicit rule that if the paper is accepted to PLDI it will be automatically rejected from ISMM. We also required the authors to inform both PC chairs that they were also submitting to the other conference. >At the end of the review period, the authors were given three days to provide a short rebuttal. Then the program committee met in Cambridge, Massachusetts on January 25th and 26th and after much discussion and debate selected this program. We hope you enjoy the program and the presentations. We are also recording the presentations and are planning to make them available to the community via YouTube.