[NTG-context] Bibliographic Databases

[Robin Kirkham]

On 20 April 2008, "George N. White III" <gnwiii at gmail.com> wrote:

> On Thu, Apr 17, 2008 at 11:19 AM, Robin Kirkham  
> <robin.kirkham at csiro.au> wrote:
>
>>  I want to set up a shared bibliographic reference database for my
>>  research group, [...]
>
> I can tell you a few things that don't work!  In our lab we have both
> TeX and Word users.  Many of them had been using a DOS package
> called papyrus, using a special markup that could be translated
> to tex (.bbl) files.  Nothing we found was really satisfactory, so
> we bought EndNote, which could import from papyrus via "refer"
> format and can export to "almost bibtex".  One problem is that
> EndNote uses unicode, so we end up with è, etc. that must
> be translated for some user's versions of bibtex.  The database
> now has a nearly infinite variety of different quote marks:
> `a`, 'a', ``a'', "a", etc. depending on how the entry was made
> (many are pasted from online or pdf sources).
>
> EndNote is really designed for individual users, although sold
> in bulk.  If 2 people open the same database on a shared
> drive they end up with a corrupt database.
>
> In my view, a bibliographic database needs to store each
> reference in the "source" or original format, whether bibtex,
> refer, or one of the newer forms, and provide translators
> and version tracking, so each file can have forks for different
> uses (e.g., ascii vs unicode char. sets) and edits can be
> preserved for the next user.  In practice, people just dump
> selected refs to a bib file, make the .bbl file, and fix problems
> there, so fixes rarely make it back to the master database.
> If they did, we would still have accents and quote marks
> being switched back and forth depending on who last used
> the entry.


Thanks George. EndNote is I believe the corporately-approved solution  
here, and similar disasters occur when people try and share its data  
files. For this reason I don't call these sorts of personal-level  
programs "databases" (any more than I'll call a .bib file a database).

The TeX folk don't fare much better. Multiple personal .bib files are  
common, often with duplicate references but different citation keys,  
leading to rather variable results depending on who run LaTeX/ConTeXt  
on the file. Inconsistency in data entry is also a problem, although  
for us, accents and quotes don't seem to too big a problem.  
(`Authorless', i.e., corporate author documents, like data sheets,  
seems to be more of an issue).

For this reason I'm looking for a proper SQL database solution like  
refdb or refbase (or maybe wikindx, thanks Andreas) with a web front- 
end that will hopefully enforce somewhat more consistent data entry,  
and maybe even auto-generate citation keys. Taco, your remarks  
regarding interchange formats are valuable (both refdb and refbase  
support MODS XML output).

Robin