On 06/29/2014 02:21 PM, john Culleton wrote: > For my contribution to the Scribus indexing specification I have now > completed phase 1 (tyro.tcl) and phase2. Please visit: > > http://wexfordpress.com/phase2.pdf > > for the latter item. > > I will hold off on phase 3 (refinements using xindy) until phase 2 is > approved by the programmers. > > Comments and criticisms of the above referenced phase 2 document are > of course requested.
Thank you John, for producing this specification. That's not the first time you help other people (including me) on this list with indexing problems. I happen to be working on a book using Scribus, and I want to make a good index to it. I took the trouble and brushed up my knowledge on LaTeX, makeindex and xindy. So I cannot say I am an expert, but I definitely pay attention to details that English speakers tend to forget: i12n. To be short: xindy is the only acceptable solution right now. As far as I know even xindy doesn't support far Eastern languages at the moment. So xindy is not perfect, but makeindex is simply inadequate even for Hungarian (my mother tongue). Xindy has also other advantages over makindex: it is a lot more flexible in almost all aspects. 1. So please if anyone works on this project, use xindy for the external indexer, not makeindex. The only disadvantage of using xindy, that it seems to be a very niche product, it is usually not included in major linux distros, and there seems to be no windows or osx port yet. And their website (http://xindy.sourceforge.net) is actually misleading you if you want to download it, you should go here: https://github.com/jschrod/xindy.ctan A few more notes to the above document: phase2.pdf 2. I don't think it is a good idea to keep an .idx file open all the time. The index items have to be stored in the scribus document file itself anyway, so the .idx file can be generated on demand any time it becomes necessary. It means it can be a temporary file. That also means the developers can change the .idx file format if they want to switch to another indexer. Or even better, you could support more than one indexer program. 3. I also don't think the output of the indexer program should be as plain as it is suggested. The indexer program (xindy) can distinguish between different classes of index items (e.g. it is customary to use boldface for those page numbers where the topic is described in detail in a book). So the indexer can indicate which page numbers should be bold, which shouldn't. Since the output of xindy is extremely customisable, you could use e.g. xml structures to pass this info to scribus rather than just the plain strings. Or you could use some other markup. Yours: Laszlo
