Lars Gullik Bjønnes wrote: > I just think that we should actively > support as few bib packages as possible (preferrably only one).
Depends on the aims of the packages, but in general: yes. > If > jurabib can do the same as natbib and more: Great. Let's use that and > ditch natbib. OTOH if natbib does all the same as jurabib does... why > support jurabib? Jurabib basically includes natbib's features (almost all, except the "force upper case" and "full author list" features we have in the citation dialog). It has lots of fundamental features for human scientists that natbib doesn't have (short title citation, ibidem, fullcite, footcite) and it is the only package that supports citation for legal texts. As I wrote in my first post of this thread, the problem is that jurabib needs the styles choice in the citation dialog, but that was bound to natbib, and jurabib is not compatible to natbib. So just including jurabib via preamble is not a solution (and vice versa). OTOH natbib is the older package and it is certainly more widespread. I think we have some layouts that require natbib (i.e. some latex classes require natbib). And then, natbib has a slightly different approach. The layout of the bibliography is hardcoded in the bst files in natbib, while it is configurable via options in jurabib. This has both its pros and cons. In jurabib, you can only choose between 4 bst files that are shipped with the package. All the rest has to be done by package options (it's a little bit like the KOMA classes; this is also the reason why jurabib needs very much memory and therefor need "bibtex 8 --huge" on windows). Natbib also has some shipped bst files, but if you want your own special layout, you usually build it with custom-bib (which was written by natbib's author). This is not (yet) possible with jurabib. An example: I needed a very special style for my thesis' bibliography: the only solution was to build my own bst style and hack its source. Jurabibs options just didn't provide what I required (though they provide a lot). AND natbib also works without bibtex (with the bibliography environment), I'm not sure jurabib does this too. Conclusio: while those packages are heading in the same direction, they have both their special areas where the other packages' limits are reached (since jurabib is very actively developped, it might completely superceed natbib eventually, but ATM this is not the case). I think we should provide this bit of selection to the users. Remember that citing is the most important feature for human scientists. You wouldn't take the argument "you can both write formulas with and without AMS, so let's decide", would you? Some users depend on jurabibs extra features, while others need natbib for their style files (personally, I toggle between those two packages for different documents). UI wise, I think both packages can easily exist in parallel, the only extra gui element is a radio button "use jurabib", the handling is exactly the same (you can also switch between those two packages any time). It also does not blow the code very much, because it basically uses the natbib framework. Perhaps we can help propagating jurabib a bit with our support. Then it might well be that we can indeed ditch natbib some day. Regards, Jürgen.
