Helge Hafting wrote: > the index is normally _not_ for documenting > every occurence of some word.
Amen. How about creating a real indexing feature for LyX that would allow users to index a document in the way Helge has described: > I did it the other way around, I > read the manuscript and indexed every important concept > I came across. Sorry if this sounds like a plug--I really mean it as a suggested addition to LyX--but you might look at my indexing add-in for MS Word as a model for what could be done in LyX: http://www.editorium.com/dexter.htm Indexing is done, by hand, in a sortable table. After the entries are finished, they're embedded in the document. For LyX, it might be possible to do this even more elegantly, by showing the generated index every time the document is generated with LaTeX, or, better still, by showing it (updated) in the LyX UI every time an index entry is added to or modified in the table. It's important for the indexer to be able to see and edit all index entries in the same place and at the same time rather than just have them spread througout the document--hence, the table. Wow, this would be a *cool* feature for LyX. What do you think? Best wishes, Jack M. Lyon ___________________________________________________ The EDITORIUM Microsoft Word Add-Ins for Publishing Professionals http://www.editorium.com ___________________________________________________ > -----Original Message----- > From: Helge Hafting [mailto:[EMAIL PROTECTED] > Sent: Thursday, February 16, 2006 1:19 AM > To: John O'Gorman > Cc: lyx > Subject: Re: Creating an Index > > > John O'Gorman wrote: > > >I am embarking on creating an index for a manual of about 200 pages > >(written in LyX of course). > > > >I studied up all the advice I could find, and as a result set out to > >produce a list of all the unique words in the book. > > > > > There are plenty of word you don't want to index. "The" should be > an obvious one. I.e. don't index anything that you really expect the > reader to know already. > > Seeing that it is a manual, limit your indexing further to only > words and concepts relevant to whatever the manual is about. > > >Alas it came to over 3000 words. Even going through and removing the > >little works like and, the, but, etc would be an enormous and tedious > >task. > > > > > Sure, making an index is a big job. Expect to spend some time > if you want it to be good. I did it the other way around, I > read the manuscript and indexed every important concept > I came across. > > Also, whenever you put something in the index, think: > 1. Is this word/concept something readers of this manual actually > might want to look up? You don't have to index something just > because you wrote about it. Only index it if it is likely to > be looked for. > Would you look for that word? > > 2. Assuming the answer to the previous question is yes, don't > fall in the > trap of indexing every occurence of that word. Sure - a computer can > easily do just that for you - and the approach is so wrong. When you > look for something, do you want to see: > > concept, 1,3,5,9,12-47,54,66-97,180,182,190-196 > > It is not very likely that you're going to look up all that, is it? > My publisher told me to only index the most important places > a word is used, because that is where people need to be directed. > Only one to three indices for any given word, they said. A manual > might be different from a textbook of course. But still, make sure > you point them to useful places. Such as the definition > of/main chapter on > "concept" and such. If "concept" is also mentioned briefly > in sentences > such as "also see concept in chapter xx" then you definitely > don't want > an index entry to point there. > Also, if there is a chapter or big section on "concept", just > index the > start > of it. "concept" may be mentioned several times in that text > over several > pages - indexing all of it does not make sense. People will want to > be directed to the start - the index is normally _not_ for documenting > every occurence of some word. > > Helge Hafting > > > > > > > >
