Sadly, the bibliographic data provided in many technical publications is sparse, inaccurate, and hard to use as a resource in itself. Some of this can be attributed to the pain that authors and typists suffer in preparing bibliographic data for publication, and some is also certainly due to human egos: most authors would rather write about what they have done, than about what others have produced.
The availability of large bibliographic databases, such as those in the TEX User Groupbibliography archive at the author's FTP site, =ftp://ftp.math.utah.edu/pub/tex/bib=, and the huge Computer Science archive at =ftp://ftp.ira.uka.de/pub/bibliography= in Karlsruhe, Germany, offers interesting possibilities for substantially improving the traditional citation bibliographies in technical documents.
First of all, when literate references are readily available in marked-up form that can be converted automatically to the style required by the publisher or organization, much of the tedium of citing references in documents is eliminated. For example, with my favorite bibliographic system, [6, pp. 155-164], all that I require in the running text of this article for the citation in this sentence is the text =[6, pp. 155-164]=, plus two simple \bibliography and \bibliographystyle declarations near the end of this document, and an invocation of the program between LATEX runs.
Second, when a fast and powerful search engine is used to access a large bibliographic archive, such as the BIBSEARCH [12] system at my site (a simple interface to the MG database system [12]), authors can readily locate many references that they might otherwise have overlooked, and they are therefore much more likely to provide more extensive bibliographies in their documents.
Third, when bibliographic data access is easy, and the collection is large, authors are likely to turn up many interesting references that they had not previously encountered, thereby enriching their own ideas, and promoting intellectual progress by building on the work of others.