When you are doing research, you tend to collect a lot of papers. I remember that at the end of m PhD, when I moved to another continent to do a postdoc, I dumped a huge box of photocopies in my parents’ basement. A few years ago, I had collected two cupboards full of photocopies. It was getting seriously out of hand. Then, of course, journals started putting everything online as PDFs and the same process started all over again but this time filling up hard disk folders instead. I used to have subject-based folders, which sort of worked until something fit within 2 or 3 or 4 of my subjects. Searching for some old paper you had read a few years back became more and more nightmarish. Then somebody showed me Papers.

Every scientist has to cope with the problem of managing references (or citations, or notes, or literature, or whatever you call it.)  When writing his second paper he discovers that he has to type a number of references that he already typed in when preparing his first paper. This repetitive action calls for a repository of references. In an ideal world many group members submit their references to this repository and after some time a very efficient storage medium has been created.

Pitfalls
Alas. The real world is never like this. And for many reasons. Typos in entries will live for ever, or will give rise to duplicate entries. Incomplete entries will downgrade the usefulness of the database. Inconsistent use of case (uppercase, lowercase, title case) is causing a mess. Different spelling of names will lead to duplicate entries, or to angry readers when they see their name misspelled in a list of references in an article in a high-impact journal. Many programs (or ‘wizards’) that import references cannot deal with extended characters (leave alone Unicode).  Names with diacritics (like umlauts) are dealt with either inconsistently or wrongly.  Partitioning of names into initials, first names and last names is full of traps and many import filters fall in those traps. In this respect the following error in the book Latex by Leslies Lamport (an excellent book and excellent macro package, of course) is typical: on page 141 (Chapter on “The Bibliography Database”) Lamport discusses “von Beethoven, Ludwig”. The name of course is Ludwig van Beethoven, as the name is of Flemish origin. And indeed “Van” is not his middle name.