Indexing hypertext

[Norman Walsh <norm@xxxxxxxxxxxxx>]

christo@xxxxxxxxxxxxxxxxxx (Frank Christoph) writes:
> I would say that the notion of "page numbers" itself is the
> limitation, part of the legacy of an age when the only way to
> reference parts of a document was by numbering pages.  A

Personally, I think linear documents are just fine.  That's the
way that it seems natural to write and read.  (I suppose that's
because I've been doing it that way since I was six or
something, but it works for me).

Anyway, I have a different topic in mind.  When a printed
document is indexed, page numbers are suitable references into
the document (if a bit coarse).  And actually, the page numbers
convey more meaning that just reference points.  If you're
familiar with a book, you may know what's around page 150 so you
can tell that 153 is probably the reference you're after.
Similarly, given page numbers "3, 5, 19, 25-34, 67, 128" you can
tell at a glance that the meaty discussion is probably at 25-34.

When a document is placed online, how should the index be
represented (or should the traditional index be replaced by
something else)?

I've tried three ways, none of which is ideal:

1. Use the page numbers from the printed book.  It's a bit odd
for the online reader.  It requires that the document _have_
page numbers from a printed edition, and that you can extract
them.  On the plus side, you get the positive features of the
print index.

2. Use other marks (dots, asterisks, sequential numbers or
letters).  The marks have no relationship to one another so,
while it works, it has none of the advantages of the print index.

3. Use the title of the closest surrounding section.  It's even
coarser than page numbers (if the sections are long), but it may
help you figure out where to go first.

For the books that O'Reilly published online while I was there,
I used option 3.

Is there a better way?

--norm




 DSSSList info and archive:  http://www.mulberrytech.com/dsssl/dssslist