Alan L Tyree wrote:
This is really just the LaTeX embedded indexing. One of Jon's earlier
posts explained that this is not what "real" indexers do.
I both agree and disagree with Jon. Real indexers do not use embedded
systems. Unfortunately, in real life, the "real" indexer is the author,
and most authors use embedded indexing. It is one of the reasons why
most indexes are so bad. Embedded indexing is very hard to keep
consistent, and most authors know SFA about indexing.
I have fooled around a bit with "semi-automatic" indexing. Instead of
trying to do it all automatically, start out with index entries that
seem suitable for your book (in other words, steal an index from a
similar book). Apply these entries through some form of automatic
indexing.
The main objection to this is that it is just indexing words. This is
true. But, if you look at most "real" indexes, probably 90% of the the
entries _are_ indexing words.
Alan
You want the idea of a paragraph, and the most relevant paragraphs to
your word. Their needs to be a brain in there somewhere, and the author
is the best place to start.
Support tools and index word lists from a similar book are useful but
dont supply a complete solution.
I use indexes of large books that I never read cover to cover, that I
consult for information that I need now, the usefulness of the book is
as dependent on the quality of the index as the information in the text.
Medical texts have good indexes, that have had much human input, often
over many editions.
Ken
--
SLUG - Sydney Linux User's Group Mailing List - http://slug.org.au/
Subscription info and FAQs: http://slug.org.au/faq/mailinglists.html
