In May of this year, the First Workshop of the Indian Database Research Community took place in Pune at the new premises of Persistent Systems Private Limited (PSPL), nestled at the base of the famous Chatushringhi hill. This event, dubbed the ``Panini DB Workshop'', went virtually unnoticed by the rest of the world, but for the budding database research community in India, it marked a watershed. For the first time, an opportunity was created wherein these researchers, who are scattered across a variety of institutions, could get together and discuss databases, the whole of databases and nothing but databases. No inaugurations, no speeches, no tamasha - it was a purely technical affair from beginning to end. This was a welcome change from the typical conference mode wherein the social and promotional activities largely overshadow the technical aspects.
The participants in the workshop, numbering over fifty, included faculty from the IITs, IISc, and Osmania, senior officials from government agencies such as NCST, and software engineers from Tessera, TCS, Versant, and, of course, PSPL, itself. The best part, however, was the enthusiastic student participation - they formed more than half the group! Since it was mandatory for all attendees to make presentations, it ensured that nobody missed the experience of being ``in the firing line''.
Although there was a good turnout, as described above, we missed the presence of quite a few database stalwarts, including Krithi Ramamritham, D. K. Subramanian, S. Seshadri, K. Padmanabhan and Sumit Ganguly, who were unable to attend due to other committments. On the positive side, however, the workshop attracted quite a few researchers for whom databases form a secondary interest - this included P. P. Chakrabarti, the legendary VLSI designer from IIT-Kharagpur and the first winner of the Swarnajayanthi fellowship in the field of computer science last year, as also Devadatt Dubhashi and Sachin Maheshwari from IIT-Delhi, both theory buffs. Their inputs went a long way in putting perspective into the deliberations.
The tone for the workshop was set in the very first presentation by Anand Deshpande (PSPL). He gave a quick summary of the workshop objectives, which included encouraging students to take up database research, to foster closer ties between database industry and research, to disseminate the state-of-the-art, and to impact the national IT effort. Anand also raised provocative questions about the lack of visibility of the Indian database research effort, its indifference to industrial needs, and its lack of presence in policy-making committees.
On the core technical front, there were six tutorials, starting with Web Databases (Soumen Chakrabarti, IIT-Mumbai) and OLAP Tools (A. Balachandran, PSPL), followed by Data Mining (Sunita Sarawagi, IIT-Mumbai) and Knowledge Management (P. Sadanandan, NCST), and winding up with Query Processing (S. Sudarshan, IIT-Mumbai) and Real-time Database Systems (Jayant Haritsa, IISc). The material for all the tutorials is available on the Workshop home-page (http://web-plus.pspl.co.in/dbworkshop).
Interspersed between the tutorials were individual presentations on the current research projects of the participants - these covered a broad spectrum of issues ranging from temporal and spatial databases to mobile transactions and directory services, apart from the obligatory OLAP and Internet-based topics. The entire set of presentations is also available on the Workshop home-page.
A particularly exciting aspect of the workshop were the panel discussions, especially those on ``Whither Indian Database Research'' and ``Why do a PhD in Databases'', which generated tremendous audience reaction. In the former, panelist recommendations included indigenous development of low-cost equivalents of foreign products, projects arising out of the ``think global, act local'' philosophy, support for Indian languages, and bringing database technology to bear on Indian data sources. There was mixed opinion on whether it was now possible to think of a big group project that would have major impact - some felt that the critical mass to undertake such an effort had been created over the last few years, spearheaded by IIT-Mumbai, while others were less sanguine, opining that the very individualistic nature of the Indian academic scene made such ``Big Bang'' projects fundamentally unworkable. There were also calls for expanding the scope of our research horizons to more than just hard-core databases and to attempt high-impact projects instead of incremental work. A synergistic effort between databases and theory was perceived to have particularly strong potential since mathematical artifacts such as lattice theory form a common thread across areas as diverse as data mining, data warehousing, query plan generation, etc. It was also considered important to form a local support community that would provide inputs for enhancing the research quality - for example, by (informally) reviewing papers prior to submission to outside fora.
An impassioned plea was made for integrating the activities of the DB group within the larger umbrella of IARCS, the Indian Association for Research in CS, and this suggestion was well received. Some participants recommended that there be a much closer interaction between industry and academia, as well as more emphasis on governmental and societal issues - a case in point being creating the database of all birth records for the entire country. However, there was considerable opposition to this viewpoint on the grounds that such projects were more appropriate for software houses and that the community's focus should remain with what they know best - basic research. But generally all were in agreement that that it would be preferable for research proposals to address an entire vertical segment, rather than just kernel features.
An issue on which there was absolute unanimity was that the general standards of database teaching were woefully poor, especially with regard to undergraduate education in most engineering colleges. It was strongly suggested that we as a community should try to rectify this situation by creating educational material in the form of slides, videos, Web packages, etc., that could be directly disseminated to college teachers and students. One possibility in this regard is to tie up with the Virtual Classroom project being promoted by Goa University to provide the infrastructure for content creation. A few students mentioned that inclusion of project demos by database companies was essential to bridge the gap between theory and practice.
Finally, on the funding front, it was gratifying to hear that most participants felt that this was not a major issue. At the same time, a strong view was expressed that we should lobby more forcefully with government agencies for funding and that we should, in fact, set up a task force that would form the entire community's interface to these agencies. Prof. Maheshwari (IIT-Delhi) provided the welcome news that the Naval Research Board is extremely keen on funding open-ended research and invited proposals under this channel.
In the ``Why do a PhD in database research'' discussion, the panelists were unanimous that the dissertation process had transformed their lives and technical abilities, enabling them to lead the state-of-the-art, apart from, of course, the obvious merit of giving them the freedom to be ``their own bosses''. Specifically with regard to doing a PhD in databases as opposed to other fields, it was mentioned that while the industry has always recognized that databases are the most important application of computer systems, this is being increasingly acknowledged by the rest of the computer science community as well - a case in point being that in recent times database benchmarks, rather than scientific benchmarks, are being used to evaluate processor designs. This is confirmed in the Proceedings of last year's ISCA conference (this is the premier computer architecture conference in the world) - the first four papers are all database-related!
A few panelists also felt that the database research environment in India had made great strides over the last decade and suggested that students seriously think about pursuing their PhDs here and not make the default assumption that meaningful research can only be carried out abroad. They pointed out that it is easier to focus single-mindedly on the research work here since issues such as scholarship and living arrangements are completely taken care of at the institutional level.
On the down-side, however, it was also highlighted that, with the sole exception of IIT-Mumbai in recent years, we do not have groups with the requisite critical mass to have worldwide impact. Other points of concern were the poor job prospects of PhDs and the lack of internal competition. Overall, in a holistic sense, there was general agreement that the Indian database research environment was now becoming a serious contender in the international league and that better quality research could be done here rather than going to second-tier institutions abroad.
In the final session of the workshop, Anand Deshpande summed up the proceedings and ensured that the momentum generated here was maintained by chalking out the workflow for the following year. It was determined that the chair of the next workshop would be S. Sudarshan (IIT-Mumbai) and that the venue would be Goa, probably on the campus of Goa University. An e-group called dbindia was to be set up by Anand (this is now functional). It was also confirmed that Krithi Ramamritham (IIT-Mumbai) would be the PC Chair of COMAD 2000 and that the conference would be held in Pune (the PC is currently being constituted and the call for papers should be out in a few months). A committee comprising of P. Sadanandan (NCST), D. B. Phatak (IIT-Mumbai) and Sunita Sarawagi (IIT-Mumbai) was identified to interface with CSI and ensure that the money remaining from the 1996 VLDB conference proceeds could be better utilized for supporting database research activities. Finally, Jinesh Vora (IIT-Mumbai) volunteered to be the student co-ordinator.
Overall, the workshop served to give a sense of purpose and direction to the Indian database community. Of particular note is that it helped the students, who often tend to feel disheartened that their research is an isolated individual endeavour, feel party to a vibrant group effort. In closing, we hope that future workshops will achieve the same degree of success.
This document was generated using the LaTeX2HTML translator Version 99.2beta8 (1.42)