web
publishing
Why DOIs Rule
DOIs (digital object identifiers) are much more than a unique id for scientific papers.
Making the style for bibliographies consistent probably ranks among the
least favorite tasks of researchers who would like to disseminate their
findings.
Thanks to LaTeX and BibTeX, the task of citing other research
is mostly reduced to curating a high-quality bib
-file of references.
But that shifts the problem to getting high-quality bib entries! My experiences with publisher-provided bib-entries and services like Google Scholar were very mixed – most required manual tweaking (and double checking!) of the entries.
This should be much easier. All it needs is a well-curated data base of metadata for scientific research (maintained by those who care for consistency: the publishers!), but with a machine-readable well-defined interface to be used by some tool created by someone who understands BibTeX well (it seems, this is rather not the publishers strength …).
Luckily, both exists! DOIs (digital object identifiers) are not just an id for papers, they also serve as keys in exactly such a data base. And with doi2bib.org, there is a service that produces high-quality bib-entries from a doi.
This brings us one step closer to a system, in which
the TeX source would only give the DOI and everything else is taken care of
automatically
(retaining the option for manual tweaking as with the bbl
files of BibTeX).
Update: Nothing is infallible
I found a case where I was not happy with the result of doi2bib: For papers in Springer journals that first appear online and later in the printed journal, doi2bib mixes the two entries. Month and year of publication are set to the first online version, but volume and issue number are also filled in, so that the resulting bib entry looks as if the printed issue appeared earlier. This is confusing, I would rather use only the final printed information, and ended up manually adapting the bib files.