From sriram.sankar@metamata.com Thu May 14 23:28:11 1998 Date: Thu, 14 May 1998 09:45:53 -0700 (PDT) From: Sriram Sankar To: taokr@cs.purdue.edu, palsberg@cs.purdue.edu Subject: recommendation letter To: Whomsoever this may concern It's a pleasure to write a testimonial on the research being conducted at Purdue University on tree builders for language processors. I am the principle developer and manager of the team that built JavaCC (Java Compiler Compiler) which is currently the most popular parser generator in use by Java developers. JavaCC is in use by literally thousands of users for a diverse range of applications. The technology on which JavaCC is based was inspired by an earlier research activity at Purdue University - PCCTS (Purdue Compiler Construction Tool Suite). I believe that the PCCTS work is a great example of how academic research eventually transitions into industry and facilitates changes for the better. PCCTS and JavaCC have enabled a significant shift in the methodology of building parser generator in industry from earlier methodologies that dated back to the early 70's. I must emphasize that with the advent of PCCTS and JavaCC, the use of parser generators has become more mainstream - previously, users of parser generators needed some formal education on compiler technology. Now this is no longer true. The rapid adoption of the new PCCTS/JavaCC methodology in industry have also created new demands for technologies that further support this methodology. Some of these technologies are simply reengineering of existing ideas, whereas there are other areas where there is need for academic research. The ongoing research on JTB (Java Tree Builder) at Purdue Univerity in Prof. Jens Palsberg's research group is one such activity. The goal of this research is to build technology that can be used in conjunction with tools such as JavaCC and PCCTS to automatically build parse trees during the parsing process. Automatic generation of parse trees during the parsing process has been studied for over 15 years now and many ideas have emerged. The time is now ripe for these ideas and many others to be collected together and applied in appropriate ways to provide maximal support for the new PCCTS/JavaCC style of parser generation. JTB has clearly shown its impact on the user community, however, there is more work that needs to be done to improve its design at the academic research level. I therefore strongly recommend funding for this project and look forward to when the JTB work will have the same impact as its predecessor PCCTS. Please feel free to contact me if you have any questions. Sriram Sankar President & CEO Metamata, Inc. -- Metamata, Inc. http://www.metamata.com Tools for serious Java developers. +1 510 796 0915