Fix Bibliographic summarizer to support BibTeX format.

On-the-fly processing, classify and summarize must be able to handle
(filename + file's contents) as well as only filename.  Might need
to use UNIX processes.

Summary vs. full-text WAIS retrieval.

Record-level vs. file-level indexing support.

Post-processing for summary files that adds bundle tags and/or combines
summary files.  bundle rules:  if directory and bundle_type then process
directory as a single file and prune the tree here; if file and 
bundle_type then expand and process as a single file.

Fix detex so that it never hangs.  Should have a timeout for all execution
of external programs.

Better FAQ summarizer that can identify the table of contents, questions,
and answers.

Port to another OS; provide sample .cf files for other platforms.

Less I/O intensive unnesting implementation; for example, use GNU tar
to extract files to standard output rather than extracting to disk.

Add full support for freeWAIS.

waisindex is space-inefficient (50% null characters in the index); maybe
add support for a simple, space-efficient inverted indexer.

Better statistics or tools to uncover the characteristics of the file
system data processed and file summaries generated.
