Skip to content
Topos Uranos
English

The bibliographic database

Summary

This lesson teaches how to build the bibliographic database of a piece of work: the .bib file from which biblatex and biber take the data of each work to typeset the citations and the list of references. It explains why the data of the works should be kept apart from the style in which they are cited, and relates that separation to logical markup; it describes the structure of an entry, its most common types, the choice of citation keys and the fields each type needs; it compares the field names of biblatex with those of classic BibTeX; it shows how the names of people, compound surnames, particles and corporate authors must be written; it explains when the capital letters of a title must be protected; it shows that biber reads accented letters directly and why the bibtex program sorts them wrongly; it presents exporting from Zotero; and it ends with the most frequent errors and how to recognize them.

Learning objectives

  1. Explain why a single bibliographic database, separate from the citation style, makes it possible to change the appearance of every reference without touching the data.
  2. Write correct entries of the most common types, with their key and with the fields each type requires.
  3. Recognize the biblatex field names and their classic equivalents, and decide which to use according to the program that will process the database.
  4. Write the names of people and institutions so that the surname, the given name and the particle are interpreted correctly, predicting how each is sorted and cited.
  5. Diagnose, from the output and from the messages of biber and LuaLaTeX, a missing comma, a repeated key, an unescaped & sign and a work that does not appear in the list.

Why a bibliographic database

A bibliographic reference brings together two things of different nature. One is the data of the work: who wrote it, what it is called, where and when it was published, on which pages of which journal it appears. The other is the style: the order in which those data are written, the typeface of each, the punctuation that separates them, the form of the citation in the text. The data of a work do not change; the style changes with the discipline, with the journal, with the university. Whoever types by hand “Darwin, C. (1859). On the Origin of Species. London: John Murray.” has fused both things into one line, and if the journal requires another style, every reference has to be redone.

It is the same distinction that the lesson on formatting with meaning established for the text: logical markup says what each passage is and lets a decision taken elsewhere fix how it looks. Here, the .bib file says what each work is; the biblatex package, with the style chosen in the preamble, decides how it is written; and the biber program, which runs between two passes of LuaLaTeX, reads the database, finds the cited works, sorts them and hands biblatex the list ready for use.

Three advantages follow from this. The first is that the data of each work are written only once, even if it is cited in ten different papers: the database is the single source of truth, and an error corrected in it is corrected in all of them. The second is that the style is changed with an option, without touching the database. The third is that the list of references contains exactly the cited works, not one more and not one less, sorted by the program and not by the author's memory.