Bit vector indexing for IDSgrep
It's in place and basically working. Still to do:
* unordered match operator * associative match operator * enhancements to "not" operator (at present, it necessarily becomes a short circuit) * research questions
Basically done except for the research paper and the babes.
Implement bit vector indexing (enhanced Bloom filters) for IDSgrep. Generate bit vector indices, search them automatically if they are found and useful, features to ignore the indices if they're not wanted, etc. Instrument this system so it can be optimized. Write a paper about the results, gain publicity, impress the babes.
See Skala et al. (ACL 2010) http://www.aclweb.org/anthology/P10-1153 and Skala and Penn (MOL 2012) http://ansuz.sooke.bc.ca/professional/mol12-approximate.pdf for some background