A Typical Index File

The FTP archive here at Indiana is a typical basic technical report index. The index file looks something like this: TR 340 Gregory J. E. Rawlins. The new publishing: Technology's impact on the publishing industry over the next decade. (Nov. 1991). 68pgs TR 346 J. Lawall & D. P. Friedman. Toward leakage containment. (Feb. 1992). 11pgs TR 347 Latha S. Colby. An algebra for list-oriented applications. (Feb. 1992). 84pgs This file can be considered as a bunch of records, each of which is separated by some standard method (a blank line, used here, is most common; other delimiters include a line which does not begin with whitespace or a line beginning with two hyphens.)

The directory contains filenames of the form:

TR340.dvi.Z TR340.ps.Z TR346.ps.Z TR347.ps.Z Since the filename extensions are highly standardized, they're easily removed. Since the base filenames are are contained (with the standard addition of a space, a minor hassle) within the records of the index file, it's reasonably easy to match files with index records. The records are folded into a normalized form with things like whitespace squeezed out, and anchors are created as appropriate.

It is important to note that the program building the index understands nothing about the index file except that it consists of records delimited by blank lines and that it contains some strings which happen to be the same as some of the filenames in that directory.

The result of the index-builder are some entries that look like this:

Indiana University CS TR340.dvi.Z(95K) TR340.ps.Z(167K)
TR 340 Gregory J. E. Rawlins. The new publishing: Technology's impact on the publishing industry over the next decade. (Nov. 1991). 68pgs
Indiana University CS TR346.ps.Z(47K)
TR 346 J. Lawall & D. P. Friedman. Toward leakage containment. (Feb. 1992). 11pgs
Indiana University CS TR347.ps.Z(305K)
TR 347 Latha S. Colby. An algebra for list-oriented applications. (Feb. 1992). 84pgs
This is a typical example of an archive, and it is perfectly adequate. However, it's not terribly impressive, and it won't allow for more sophisticated features we discussed earlier. Now, let's go on to an example of bib format...

Marc V