On Thu, Nov 07, 2024 at 07:33:32PM +0000, Werner LEMBERG wrote: > > >> I ask to implement the same for the PDF format (and probably for > >> Info, too) to help with very large indices – for example, > >> LilyPond's main index has a length of almost 30 pages. > > > > [...] PDF readers also have search commands, so I'm not sure what > > exactly is missing here. > > Well, a big problem is that normal PDF readers don't have regex > search. Additionally, the index is also used by people who don't know > exactly what they are looking for. As an example, in music there are > appoggiaturas and accacciaturas, which are very similar but not > identical things – (a) you have to know these words, and (b) you have > to know how to spell them correctly. However, the iser might remember > that the word start with letter 'a', so it's quite natural IMHO to > look up this letter in the index. > > A similar situation is to look up keywords, functions, variables, > properties, etc., that are very similar, for example, `foo`, `\foo`, > `\Foo`, `\foobar`, `\fooBar`. Within an index you can identify them > easily, however, it's not so easy to search a single one of them > without hitting zillions of unwanted matches.
It could be possible to put letter sections for indices in the table of contents but this would require outputting extra information in the .toc file output in a TeX run, so would have to be implemented carefully. If it turns out it is simple to implement in texinfo.tex it could be controlled with a @set flag for texinfo.tex only, as we shouldn't have too many options specific to one output format only. We would have to consider how this feature would work for printed output as well, so would have to show page numbers for the letter references.
