Inhaltsverzeichnis

Literature

nanos gigantum umeris insidentes Bernhard of Chartres

If I have seen further it is by standing on ye shoulders of giants. Isaac Newton

At least every scientific work needs proper dealing with its sources and therefore a bibliography1).

From the above, a number of fundamental questions arise about how to organise the informations necessary for a bibliography, and, if applicable, the literature itself2):

All these aspects will be dealt with in more details in the following.

Literature database

This topic can be divided in two aspects: Formats and their associated interfaces.

Formats

Although there are many more, two formats are of particular interest here:

While EndNote is a commercial (and binary) format and therefore its user depends on a (only) commercially available (and quite expensive) software, BibTeX is a complete free and open-source software and a pure text format (that can be read and modified very easily).

Those who use Word (or other word processing software) will usually use EndNote or one of its relatives (EndNote Web).

Those who write their (scientific) publications and texts with LaTeX (highly recommended!) normally tend to use BibTeX.

Note: It is possible to convert between both formats, EndNote and BibTeX. But be aware of the fact that such conversions are normally never completely perfect and without any errors.

Many websites (journals, literature search portals such as “Web of Science” and “PubMed”) allow to export the records in both (and even more) formats.

Interfaces

Whereas EndNote comes with its own interface, in case of BibTeX, the interface is - in the easiest case - a standard text editor capable of displaying and saving pure (ASCII) text3).

Besides this “classic” way to use a text editor of your choice to directly edit your BibTeX files, nowadays there are a number of programs that allow for a more comfortable editing of BibTeX databases and provide further functionality (such as searching, etc.).

One program that is especially useful (given its platform independence) and easy to use should be highlighted here and is recommended for use for all those who dislike the idea of having to use a text editor and manually edit their BibTeX files:

Nevertheless, especially when using BibteX there are good reasons for the old-fashioned way of editing the literature database using a text editor. For details, see below.4)

Organising your (PDF) files

Especially in science most of the references are articles from scientific journals. Nowadays, probably the vast majority of these articles are accessible online in electronic form (as PDF files).

Why then bother to save these files locally, and how to organise the files?

Why saving literature locally?

In former times, it was the filing cabinet, nowadays it is the hierarchy of folders on your file system of your local hard drive. Even though most of the articles can be accessed easily and comfortably using internet and search engines (given that you're currently connected to a university network), there are good reasons for saving the literature relevant to you locally on your hard drive:

How to organise the files?

Idea: Even without using the literature database, access to the actual literature (i.e., PDF file) should be as easy as possible.

One, if not the most prominent access to (scientific) literature is via the name of the first author of a publication. This is the normal way of referencing in a text, if you don't use numeric references there.

For a complete description of an article published in a journal (the common case for literature that you have access to in electronic form), you need four7) pieces of information:

Accordingly, the file name should contain these four pieces of information.8) A scheme for file names that is consistent with the criteria named here will be introduced later.

A few hints for how to organise your files:

Of course, all those hints are not obligatory, but only stem from personal experience (although experience of quite some time).

Connecting literature database and file system

How to connect your literature list/database with your (electronically available) literature?

A few of the interfaces between database and user (EndNote, JabRef) allow you to directly set links to files within the respective program. But beware of the obvious disadvantages that might have10).

A far more robust way would be to create and use a generic scheme for naming your files and then add these file names (without extension) as “keys” to the respective records. In case that your generic scheme allows furthermore to directly extract the full citation (author, journal name, volumne, first page, see above) from the key/file name, than it will be very easy for you to look for a specific reference in your files starting out from a citation in an article (or elsewhere)11).

Naming scheme for files and keys

What follows is a scheme for naming files (and corresponding (BibTeX) keys) that has been used for quite some time and proved to be useful. It follows all criteria named above.

Key of the concept is to name the (PDF) files in the exact same way as the (BibTeX) keys. Additionally, from the file/key names, you can extract all necessary information for a full reference and vice versa.

A (bibtex) key for an article in a (scientific) journal consists of the following four parts, each followed by a short explanation:

First author's name
The first four letters of the first author's name
Journal
A unique abbreviation of the journal.
If you use BibTeX, you could/should create files that contain assignments of these abbreviations to either the full or the abbreviated name of the journal. The idea: Easy switching between full and abbreviated journal names in your (LaTeX) bibliographies without having to alter the actual BibTeX records. Besides that, this scheme helps you to keep your journal names consistent.
Volumne
The volume of the respective journal the article appeared in.
Page
The first page of the article.

An example

To demonstrate the scheme introduced above, a short example using BibTeX:

literature.bib
@STRING{LASTEDIT = {25/NOV/2011}}
 
@Article{bisk-aciee-48-404,
  author = 	 {Till Biskup AND Erik Schleicher AND Asako Okafuji AND Gerhard Link AND Kenichi Hitomi AND Elizabeth D. Getzoff AND Stefan Weber},
  title = 	 {Direct Observation of a Photoinduced Radical Pair in a Cryptochrome Blue--Light Photoreceptor},
  journal =	 ACIEE,
  year = 	 {2009},
  volume = 	 {48},
  pages = 	 {404--407}
}
fulljrnl.bib
@STRING{LASTEDIT = {25/NOV/2011}}
 
@STRING{ACIEE = {Angewandte Chemie International Edition English}}
abbrjrnl.bib
@STRING{LASTEDIT = {25/NOV/2011}}
 
@STRING{ACIEE = {Angew. Chem. Int. Ed.}}

The corresponding file name of the PDF file is therefore: bisk-aciee-48-404.pdf

If you use BibTeX together with LaTeX, include the bibliography as follows:

\bibliography{fulljrnl,literatur}

or:

\bibliography{abbrjrnl,literatur}

In both cases, it is important to load the file with the assignments of abbreviations and actual journal names before loading the bibliography itself. Otherwise, BibTeX cannot do the necessary replacement of strings.

Please note: The abbreviations given here for the journal names (i.e., the abbreviations used in the file and key names) are no “official” abbreviations, but rather user-defined short string. Therefore, it might be useful if you plan to share BibTeX and PDF files within a group to agree not only about the general naming scheme of the keys and files, but as well about the abbreviations used for the journals.

Dos and Don'ts

A well-maintained literature database is very useful for your own scientific publications, be it a thesis or any arbitrary other publication (article in a journal, chapter in a book, etc.).

As literature databases grow over time, it is important and helpful to follow a few general rules straight from the beginning.

If you apply these rules consequently and straight from the beginning, that will save you lots of tedious work later on. Furthermore, it allows you to rely on your own library/literature database without restrictions and without having to check every time.

1)
The term “bibliography” is used here in two meanings: On the one hand, it is the list of references of a given text, that is usually added at its end. On the other hand it is a list of all potentially useful literature for a given topic, and therefore at least as comprehensive as the former, but mostly larger. Ideally, when starting to deal with a topic, one immediately starts creating a bibliography of all the literature one has read already or intends to read, that is or might be relevant.
2)
Given the scientific background of the author, the following is clearly influenced by what is usual in science. Nevertheless, many aspects should be useful and applicable in a much wider scope.
3)
Using Windows, one example would be Notepad, but under no circumstances Word or simliar programs. Using Linux and Mac OS X, there are a vast number of different editors, starting with the “classic” choices such as vi and Emacs, but as well gedit and TextEdit or, in the case of Mac OS X, advanced programs such as TextWrangler/BBedit capable of much more than simple editing of text files.
4)
One reason: I still couldn't figure out how to convince JabRef using BibTeX strings for certain fields. This but is a particularly useful and easy way of switching between full and abbreviated journal names without having to edit the main bibliography database and besides that to achieve a consistent naming of the journals.
5)
To my knowledge, sharing PDFs of articles this way is perfectly legal, as long as you don't share the electronic copy via universally accessible web pages or “publish” them otherwise.
6)
One of the big advantages of printed copies of journals in libraries was that the printed copy remained with the library, even if they cancelled subscription. With the now usual electronic subscription, access to previous volumes is no longer guaranteed once the library cancels the contract with a publisher.
7)
Strictly speaking, journal name, volume, and page number are sufficient. Nevertheless, adding the first author's name has obvious practical advantages.
8)
Of course, you can put whatever you like into your file names. Nevertheless, in my experience, sticking to these four pieces of information helps greatly with organising your literature, as your file names can be created completely generic that way.
9)
Most modern file browsers allow to search for a file simply by starting to type the first characters of its name.
10)
If you link files with their full path, that will work only as long as you don't change your directory hierarchy. And it will most certainly not work when you want to transfer your database to another computer.
11)
This is particularly useful when you're reading an article and want to shortly look for a specific refrence.
12)
Yes, such things do happen, as I know from my own experience.