Oren Patashnik
February 8, 1988
[This document will be expanded when BibTeX version 1.00 comes out. Please report typos, omissions, inaccuracies, and especially unclear explanations to me (patashnik@SCORE.STANFORD.EDU). Suggestions for improvements are wanted and welcome.]
This documentation, for BibTeX version 0.99b, is meant for general BibTeX users; bibliography-style designers should read this document and then read ``Designing BibTeX Styles'' [3], which is meant for just them.
This document has three parts: Section 2 describes the differences between versions 0.98i and 0.99b of BibTeX and between the corresponding versions of the standard styles; Section 3 updates Appendix B.2 of the LATEX book [2]; and Section 4 gives some general and specific tips that aren't documented elsewhere. It's assumed throughout that you're familiar with the relevant sections of the LATEX book.
This documentation also serves as sample input to help BibTeX implementors get it running. For most documents, this one included, you produce the reference list by: running LATEX on the document (to produce the aux file(s)), then running BibTeX (to produce the bbl file), then LATEX twice more (first to find the information in the bbl file and then to get the forward references correct). In very rare circumstances you may need an extra BibTeX/LATEX run.
BibTeX version 0.99b should be used with LATEX version 2.09, for which the closed bibliography format is the default; to get the open format, use the optional document style openbib (in an open format there's a line break between major blocks of a reference-list entry; in a closed format the blocks run together).]
Note: BibTeX 0.99b is not compatible with the old style files; nor is BibTeX 0.98i compatible with the new ones (the new BibTeX, however, is compatible with old database files).
Note for implementors: BibTeX provides logical-area names TEXINPUTS: for bibliography-style files and TEXBIB: for database files it can't otherwise find.
This section describes the differences between BibTeX versions 0.98i and 0.99b, and also between the corresponding standard styles. There were a lot of differences; there will be a lot fewer between 0.99 and 1.00.
The following list explains BibTeX's new features and how to use them.
\nocite{*}
'
you can now include in the reference list
every entry in the database files, without having to explicitly
\cite
or \nocite
each entry.
Giving this command, in essence,
\nocite
s
all the enties in the database, in database order,
at the very spot in your document
where you give the command.
@STRING( WGA = " World Gnus Almanac" )then it's easy to produce nearly-identical title fields for different entries:
@BOOK(almanac-66, title = 1966 # WGA, . . . @BOOK(almanac-67, title = 1967 # WGA,and so on. Or, you could have a field like
month = "1~" # jan,which would come out something like `
1~January
' or `1~Jan.
' in the bbl file,
depending on how your bibliography style defines
the jan abbreviation.
You may concatenate as many strings as you like
(except that there's a limit to the overall length
of the resulting field);
just be sure to put the concatenation character `#',
surrounded by optional spaces or newlines,
between each successive pair of strings.
\cite{no-gnats}
in your document,
and suppose you have these two entries in your database file:
@INPROCEEDINGS(no-gnats, crossref = "gg-proceedings", author = "Rocky Gneisser", title = "No Gnats Are Taken for Granite", pages = "133-139") . . . @PROCEEDINGS(gg-proceedings, editor = "Gerald Ford and Jimmy Carter", title = "The Gnats and Gnus 1988 Proceedings", booktitle = "The Gnats and Gnus 1988 Proceedings")Two things happen. First, the special crossref field tells BibTeX that the no-gnats entry should inherit any fields it's missing from the entry it cross references, gg-proceedings. In this case it in inherits the two fields editor and booktitle. Note that, in the standard styles at least, the booktitle field is irrelevant for the PROCEEDINGS entry type. The booktitle field appears here in the gg-proceedings entry only so that the entries that cross reference it may inherit the field. No matter how many papers from this meeting exist in the database, this booktitle field need only appear once.
The second thing that happens:
BibTeX automatically puts the entry gg-proceedings
into the reference list if it's cross
referenced by two or more entries that you
\cite
or \nocite
,
even if you don't \cite
or \nocite
the gg-proceedings entry itself.
So gg-proceedings will automatically appear
on the reference list if one other entry
besides no-gnats cross references it.
To guarantee that this scheme works, however, a cross-referenced entry must occur later in the database files than every entry that cross-references it. Thus, putting all cross-referenced entries at the end makes sense. (Moreover, you may not reliably nest cross references; that is, a cross-referenced entry may not itself reliably cross reference an entry. This is almost certainly not something you'd want to do, though.)
One final note: This cross-referencing feature is completely unrelated to the old BibTeX's cross referencing, which is still allowed. Thus, having a field like
note = "Jones \cite{jones-proof} improves the result"is not affected by the new feature.
author = "Kurt G{\"o}del", year = 1931,and if you're using the alpha bibliography style, then BibTeX will construct the label [Göd31] for this entry, which is what you'd want. To get this feature to work you must place the entire accented character in braces; in this case either
{\"o}
or {\"{o}}
will do.
Furthermore these braces must not themselves be
enclosed in braces (other than the ones that might delimit
the entire field or the entire entry);
and there must be a backslash
as the very first character inside the braces.
Thus neither {G{\"{o}}del}
nor {G\"{o}del}
will work for this example.
This feature handles all the accented characters and all but the nonbackslashed foreign symbols found in Tables 3.1 and 3.2 of the LATEX book. This feature behaves similarly for ``accents'' you might define; we'll see an example shortly. For the purposes of counting letters in labels, BibTeX considers everything contained inside the braces as a single letter.
author = "Jean-Paul Sartre",and if you're using the abbrv style, then the result is `J.-P. Sartre'.
@PREAMBLE
command
for the database files.
This command's syntax is just like @STRING
's,
except that there is no name or equals-sign, just the string.
Here's an example:
@PREAMBLE{ "\newcommand{\noopsort}[1]{} " # "\newcommand{\singleletter}[1]{#1} " }(note the use of concatenation here, too). The standard styles output whatever information you give this command (LATEX macros most likely) directly to the bbl file. We'll look at one possible use of this command, based on the
\noopsort
command just defined.
The issue here is sorting (alphabetizing). BibTeX does a pretty good job, but occasionally weird circumstances conspire to confuse BibTeX: Suppose that you have entries in your database for the two books in a two-volume set by the same author, and that you'd like volume 1 to appear just before volume 2 in your reference list. Further suppose that there's now a second edition of volume 1, which came out in 1973, say, but that there's still just one edition of volume 2, which came out in 1971. Since the plain standard style sorts by author and then year, it will place volume 2 first (because its edition came out two years earlier) unless you help BibTeX. You can do this by using the year fields below for the two volumes:
year = "{\noopsort{a}}1973" . . . year = "{\noopsort{b}}1971"According to the definition of
\noopsort
,
LATEX will print nothing but the true year for these fields.
But BibTeX will be perfectly happy pretending that
\noopsort
specifies some fancy accent
that's supposed to adorn the `a' and the `b';
thus when BibTeX sorts it will pretend that
`a1973' and `b1971' are the real years,
and since `a' comes before `b', it will place volume 1 before volume 2,
just what you wanted.
By the way, if this author has any other works included
in your database, you'd probably want to use instead something like
{\noopsort{1968a}}1973
and
{\noopsort{1968b}}1971
,
so that these two books would come out in a reasonable spot
relative to the author's other works
(this assumes that 1968 results in a reasonable spot,
say because that's when the first edition of volume 1 appeared).
There is a limit to the number of @PREAMBLE
commands
you may use, but you'll never exceed this limit if
you restrict yourself to one per database file;
this is not a serious restriction,
given the concatenation feature (item 2).
This section describes changes to the standard styles (plain, unsrt, alpha, abbrv) that affect ordinary users. Changes that affect style designers appear in the document ``Designing BibTeX Styles'' [3].
~
) have been removed.
LATEX thus will produce slightly fewer
`Underfull \hbox
' messages
when it's formatting the reference list.
{\em ...}
)
has replaced italicizing ({\it ...}
).
This will almost never result in a difference
between the old output and the new.
\etalchar
command
that the alpha style writes onto the bbl file
(just preceding the \thebibliography
environment);
use LATEX's \renewcommand
inside
a database @PREAMBLE command,
described in the previous subsection's item 6.
\cite
of the cross-referenced entry and by
omitting from the cross-referencing entry
(most of the) information that appears
in the cross-referenced entry.
These styles do this when
a titled thing (the cross-referencing entry)
is part of a larger titled thing (the cross-referenced entry).
There are five such situations:
when (1) an INPROCEEDINGS
(or CONFERENCE, which is the same)
cross references a PROCEEDINGS;
when (2) a BOOK, (3) an INBOOK,
or (4) an INCOLLECTION
cross references a BOOK
(in these cases, the cross-referencing entry is a single
volume in a multi-volume work);
and when (5) an ARTICLE
cross references an ARTICLE
(in this case, the cross-referenced entry is really a journal,
but there's no JOURNAL entry type;
this will result in warning messages about
an empty author and title for the journal--you should just ignore these warnings).
type = "{Ph.D.} dissertation"in your database entry.
chapter = "1.2", type = "Section"in your database entry.
This section is simply a corrected version of Appendix B.2 of the LATEX book [2], © 1986, by Addison-Wesley. The basic scheme is the same, only a few details have changed.
When entering a reference in the database, the first thing to decide is what type of entry it is. No fixed classification scheme can be complete, but BibTeX provides enough entry types to handle almost any reference reasonably well.
References to different types of publications contain different information; a reference to a journal article might include the volume and number of the journal, which is usually not meaningful for a book. Therefore, database entries of different types have different fields. For each entry type, the fields are divided into three classes:
The following are the standard entry types, along with their required and optional fields, that are used by the standard bibliography styles. The fields within each class (required or optional) are listed in order of occurrence in the output, except that a few entry types may perturb the order slightly, depending on what fields are missing. These entry types are similar to those adapted by Brian Reid from the classification scheme of van Leunen [4] for use in the Scribe system. The meanings of the individual fields are explained in the next section. Some nonstandard bibliography styles may ignore some optional fields in creating the reference. Remember that, when used in the bib file, the entry-type name is preceded by an @ character.
In addition to the fields listed above, each entry type also has an
optional key field, used in some styles
for alphabetizing, for cross referencing,
or for forming a \bibitem
label.
You should include a key field for any entry whose
``author'' information is missing;
the ``author'' information is usually the author field,
but for some entry types it can be the editor
or even the organization field
(Section 4 describes this in more detail).
Do not confuse the key field with the key that appears in the
\cite
command and at the beginning of the database entry;
this field is named ``key'' only for compatibility with Scribe.
Below is a description of all fields recognized by the standard bibliography styles. An entry can also contain other fields, which are ignored by those styles.
\cite
command and at the beginning of the database entry.
This section gives some random tips that aren't documented elsewhere, at least not in this detail. They are, roughly, in order of least esoteric to most. First, however, a brief spiel.
I understand that there's often little choice in choosing a bibliography style--journal says you must use style and that's that. If you have a choice, however, I strongly recommend that you choose something like the plain standard style. Such a style, van Leunen [4] argues convincingly, encourages better writing than the alternatives--more concrete, more vivid.
The Chicago Manual of Style [1], on the other hand, espouse the author-date system, in which the citation might appear in the text as `(Jones, 1986)'. I argue that this system, besides cluttering up the text with information that may or may not be relevant, encourages the passive voice and vague writing. Furthermore the strongest arguments for using the author-date system--like ``it's the most practical''--fall flat on their face with the advent of computer-typesetting technology. For instance the Chicago Manual contains, right in the middle of page 401, this anachronism: ``The chief disadvantage of [a style like plain] is that additions or deletions cannot be made after the manuscript is typed without changing numbers in both text references and list.'' LATEX, obviously, sidesteps the disadvantage.
Finally, the logical deficiencies of the author-date style are quite evident once you've written a program to implement it. For example, in a large bibliography, using the standard alphabetizing scheme, the entry for `(Aho et al., 1983b)' might be half a page later than the one for `(Aho et al., 1983a)'. Fixing this problem results in even worse ones. What a mess. (I have, unfortunately, programmed such a style, and if you're saddled with an unenlightened publisher or if you don't buy my propaganda, it's available from the Rochester style collection.)
Ok, so the spiel wasn't very brief; but it made me feel better, and now my blood pressure is back to normal. Here are the tips for using BibTeX with the standard styles (although many of them hold for nonstandard styles, too).
\bibliography
command
(but you should list this argument before the ones that
specify real database entries).
month = jul # "~4,"will probably produce just what you want.
\nocite{*}
feature
(all entries in the database are included),
the placement of the \nocite{*}
command
within your document file will determine the reference order.
According to the rule given in Section 2.1:
If the command is placed at the beginning of the document,
the entries will be listed in exactly the order
they occur in the database;
if it's placed at the end,
the entries that you explicitly
\cite
or \nocite
will occur in citation order,
and the remaining database entries will be in database order.
author = "Donald E. Knuth" . . . author = "D. E. Knuth"There are two possibilities. You could (1) simply leave them as is, or (2) assuming you know for sure that these authors are one and the same person, you could list both in the form that the author prefers (say, `Donald E. Knuth'). In the first case, the entries might be alphabetized incorrectly, and in the second, the slightly altered name might foul up somebody's electronic library search. But there's a third possibility, which is the one I prefer. You could convert the second journal's field to
author = "D[onald] E. Knuth"This avoids the pitfalls of the previous two solutions, since BibTeX alphabetizes this as if the brackets weren't there, and since the brackets clue the reader in that a full first name was missing from the original. Of course it introduces another pitfall--`D[onald] E. Knuth' looks ugly--but in this case I think the increase in accuracy outweighs the loss in aesthetics.
organization = "The Association for Computing Machinery", key = "ACM"Without the key field, the alpha style would make a label from the first three letters of information in the organization field; alpha knows to strip off the `The ', but it would still form a label like `[Ass86]', which, however intriguing, is uninformative. Including the key field, as above, would yield the better label `[ACM86]'.
You won't always need the key field to override the organization, though: With
organization = "Unilogic, Ltd.",for instance, the alpha style would form the perfectly reasonable label `[Uni86]'.
author = "\AA{ke} {Jos{\'{e}} {\'{E}douard} G{\"o}del"there are just two special characters, `
{\'{E}douard}
' and `{\"o}
'
(the same would be true if the pair of double quotes
delimiting the field were braces instead).
In general, BibTeX will not do any processing
of a TEX or LATEX control sequence inside a special character,
but it will process other characters.
Thus a style that converts all titles to lower case
would convert
The {\TeX BOOK\NOOP} Experienceto
The {\TeX book\NOOP} experience(the `The' is still capitalized because it's the first word of the title).
This special-character scheme is useful for handling accented characters, for getting BibTeX's alphabetizing to do what you want, and, since BibTeX counts an entire special character as just one letter, for stuffing extra characters inside labels. The file XAMPL.BIB distributed with BibTeX gives examples of all three uses.
Each name consists of four parts: First, von, Last, and Jr; each part consists of a (possibly empty) list of name-tokens. The Last part will be nonempty if any part is, so if there's just one token, it's always a Last token.
Recall that Per Brinch Hansen's name should be typed
"Brinch Hansen, Per"The First part of his name has the single token ``Per''; the Last part has two tokens, ``Brinch'' and ``Hansen''; and the von and Jr parts are empty. If you had typed
"Per Brinch Hansen"instead, BibTeX would (erroneously) think ``Brinch'' were a First-part token, just as ``Paul'' is a First-part token in ``John Paul Jones'', so this erroneous form would have two First tokens and one Last token.
Here's another example:
"Charles Louis Xavier Joseph de la Vall{\'e}e Poussin"This name has four tokens in the First part, two in the von, and two in the Last. Here BibTeX knows where one part ends and the other begins because the tokens in the von part begin with lower-case letters.
In general, it's a von token if the first letter at brace-level 0 is in lower case. Since technically everything in a ``special character'' is at brace-level 0, you can trick BibTeX into thinking that a token is or is not a von token by prepending a dummy special character whose first letter past the TEX control sequence is in the desired case, upper or lower.
To summarize, BibTeX allows three possible forms for the name:
"First von Last" "von Last, First" "von Last, Jr, First"You may almost always use the first form; you shouldn't if either there's a Jr part, or the Last part has multiple tokens but there's no von part.
この文書はLaTeX2HTML 翻訳プログラム Version 2002-2-1 (1.70)
Copyright © 1993, 1994, 1995, 1996,
Nikos Drakos,
Computer Based Learning Unit, University of Leeds,
Copyright © 1997, 1998, 1999,
Ross Moore,
Mathematics Department, Macquarie University, Sydney.
を日本語化したもの( 2002-2-1 (1.70) JA patch-2.0 版)
Copyright © 1998, 1999,
Kenshi Muto,
Debian Project.
Copyright © 2001, 2002,
Shige TAKENO,
Niigata Inst.Tech.
を用いて生成されました。
コマンド行は以下の通りでした。:
latex2html -split 0 -lcase_tags -nosubdir -nonavigation -show_section_numbers btxdoc.tex.
翻訳は Deniz Yuret によって 平成18年12月6日 に実行されました。