[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [docbook] DocBook Technical Committee Meeting Minutes: 24 Sep2003
Norman Walsh wrote: > 8b. Topic Index > > NW: Explains the problem after more discussion with the submitter: the > need is really for general index terms of different types. > > MS: TEI has this feature. And if you generate an index just from other > markup, then you don't get the richness of secondary, tertiary, etc. > If you really wanted to make a sophisticated index with different > types, you'd need to have this kind of markup. > > NW: What are the semantics of this attribute? > > Proposal: add a 'type' attribute to 'indexterm' and 'index' to support > this markup. > > Accepted. > > BS: How tightly we need to define the processing expectations? > > Proposed semantics: > > Indexterms of type 'x' go in index of type 'x'. An index with no type > gets all of the index terms regardless of their type. > > In 4.3? > > Yes. That's awesome ! I'm now looking into patches for the xsl stylesheets to make that work. Looking into xhtml/index.xsl, I find the following comment, which I find slightly confusing, and may be even contradictory to the above processing expectations: <!-- some implementations use completely empty index tags to indicate --> <!-- where an automatically generated index should be inserted. so --> <!-- if the index is completely empty, skip it. Unless generate.index --> <!-- is non-zero, in which case, this is where the automatically --> <!-- generated index should go. --> So what should I do ? Should I (for the time being) simply assume the 'generate.index' parameter is non-zero ? Regards, Stefan
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]