Soren Sandmann Pedersen writes: > According to The Companion, indices should be generated with the MakeIndex > program. This is fine as long as one is writing English (and German?) texts; > for other languages, however, the sorting mechanism does not work > properly, since it does not take into account the language-specifik > characters. even for english and german you will find deficiencies with the approach MakeIndex takes, namely to use a larger number of characters as control characters. Having produced three books using that program I can confirm that it takes considerable hacking to resolve some of the glitches. (in some cases i had to use additional unix tools to resolve some of the remaining problems) On the other hand for those two languages especially for medium sized indexes makeindex offers a lot of flexibility to produce a reasonable result fast. > This made me wonder if the index generation could be generated by (La)TeX > itself. Are there any reasons why it couldn't? i don't see a reason by it could not, although i think it should or has to be a standalone program to avoid cluttering memory of the main job perhaps it should be pointed out that there is/was a project understaken or supervised by Joachim Schrod to implement a truely international index program. In any case researching the requirements for multi-lingual generic mono-lingual index processing would be worth the effort. whether a final program would/should be in written in TeX is a different matter, however as a prototype it could certainly serve. the major advantage of using TeX as the index generation machine as i see it whould be that this would allow to produce indexes on every TeX installation without the installation of additional software. > I would like to hear some opinions on this matter, as I might start > writing an experimental package. long time ago we proposed several topics that we would think being worth doing research in, requirements for indexing commands was one of them (vt08 i think). Unfortunately that one was never taken up by any volunteer (let alone a volunteer group). Clearly the work you suggested would lead to identify those requirements (or some of them at least), for example, what type of language information would need to be passed from the LaTeX class to the index processing program (be it external or TeX or whatever) so from that point alone i would welcome any research and i think it would be beneficial. however i think you should try finding about about the status of other projects first to avoid unnecessary double work frank