indxbib
, lookbib
—
build inverted index for a bibliography, find references in a
bibliography
indxbib
makes an inverted index to the named
databases (or files) for use by
lookbib(1)
and
refer(1).
These files contain bibliographic references (or other kinds of information)
separated by blank lines.
A bibliographic reference is a set of lines, constituting fields
of bibliographic information. Each field starts on a line beginning with a
``%'', followed by a key-letter, then a blank, and finally the contents of
the field, which may continue until the next line starting with ``%''.
indxbib
is a shell script that calls
/usr/ucblib/reftools/mkey and
/usr/ucblib/reftools/inv. The first program,
mkey
, truncates words to 6 characters, and maps
upper case to lower case. It also discards words shorter than 3 characters,
words among the 100 most common English words, and numbers (dates) < 1900
or > 2000. These parameters can be changed; see page 4 of the
Refer
document by Mike Lesk. The second program,
inv,
creates an entry file (.ia), a posting file
(.ib), and a tag file (.ic), all in the working directory.
lookbib
uses an inverted index made by
indxbib
to find sets of bibliographic references. It
reads keywords typed after the ``>'' prompt on the terminal, and
retrieves records containing all these keywords. If nothing matches, nothing
is returned except another ``>'' prompt.
lookbib
will ask if you need instructions,
and will print some brief information if you reply ``y''.
It is possible to search multiple databases, as long as they have
a common index made by indxbib
. In that case, only
the first argument given to indxbib
is specified to
lookbib
.
If lookbib
does not find the index files
(the .i[abc] files), it looks for a reference file with the same name as the
argument, without the suffixes. It creates a file with a `.ig' suffix,
suitable for use with fgrep
. It then uses this fgrep
file to find references. This method is simpler to use, but the .ig file is
slower to use than the .i[abc] files, and does not allow the use of multiple
reference files.
x.ia, x.ib,
x.ic, where x is the first argument,
or if these are not present, then x.ig,
x
Probably all dates should be indexed, since many disciplines refer to literature
written in the 1800s or earlier.