Quick-Search Database Systems Research Bibliography

Before your access to this page, queries have been answered by this search engine.

Please provide the following details:
(Refer to the hints page for tips on how to quickly retrieve what you are looking for)

Maximum Number of Entries to be returned:
Proximity of Keywords:
Keyword1: AND
Keyword2: AND
Case Sensitive Search: Partial Match: <--(keyword is part of a string)
Check out the high performance version of the search engine under development.

Details of the Search Engine and the Underlying Bibliography:

This search is performed using a indexing/search engine designed and developed for bibliography files by Mohan U. Kamath using a combination of Perl and C programs. The total system consists of less than 800 lines of code and uses inverted lists, multilevel indexing and other optimizations to perform competitively with other bibliography search sites equipped with text retrieval engines. By incorporating additional query primitives and a document ranking facility this search engine can be extended to a full text retrieval engine. The search engine uses close to 10 MB of database bibliography (20,000+ entries) available from Database Bibliographies. This bibliography will be enhanced gradually. Currently the bibliography is indexed only on important fields like author, title, keyword, abstract, year, annote and institution. Approximate CPU times for processing queries that match 1, 2 and 3 keywords are 0.7, 2.5 and 3.3 seconds respectively. Actual response time will vary depending upon the keyword frequencies, network/server traffic and other web overheads. Currently index files and bibliography files are read from the disk for each request. Plans are on to setup a server, in which case index files need not be read everytime, making the retrieval even faster. This is not a probabilistic system and hence precision and recall in the text retrieval sense are both 100%. The keyword(s) specified are highlighted in the entries returned. If the source bibliography files contain repeated entries, this will be reflected in your search results. If a search for multiple keywords fails, the system provides some feedback about the number of matches found for each keyword so that you can modify and resubmit your query. If you have problems/suggestions and pointers to bibliography entries that could be added to our database, please send mail to the author - thanks!

Mohan Kamath 1995
Last Update: September 12, 1995

To the Database Systems page of Univ of Massachusetts