Re: Reasons for online

["Doug, Data Librarian at Ext 4225" <engstromdd -at- PHIBRED -dot- COM>]

Coleen Mackay, defending the honor and jobs of professional indexers
everywhere, writes:

***************************
...If the online index is created with hypertext links, it's missing a lot.
Concepts that are not spelled out explicitly in the text are not indexed.
A manual index is likely created by a human indexer, and would pick up
those missing concepts.  That's why the manual index is far superior!
****************************

Let's straighten out a terminology problem before this gets out of hand.
(Ever notice how many problems are caused by people talking past each
other using the same words?)

Anyway, the problem here is not the hypertext links, (which affect only how
the user travels from the index entry to the relevant text) but rather how
those links are created.

An "automatic indexer," (the technology Coleen is actually railing against)
typically searches the text for an indexed word and marks the index with a
page number (for paper manuals) or a hyperlink (for on-line versions).  As
Coleen points out, this approach can miss important occurences of the
concept, such as a discussion of "Bills of Lading" which does not contain
the word "shipping," a clearly related concept.  It also suffers from the
opposite problem; a word covering two concepts.  For example, a database
program dealing with crop management may use the word "row" for two
dramatically different concepts--a row of data and a row of plants.  An
automatic indexer will brainlessly combine the two in a single index entry.
Full text search, often touted as an index replacement, suffers from
exactly the same problems.

However, hyperlink indexes can be built by a human indexer, just like the
index in a paper manual.  As someone commented, there's no reason the
index for a paper document and a on-line document have to be different if
the material is the same.  (There's a whole issue here on the extent to
which the material *should* differ, but, one thing at a time.)

Myself, I often let the auto-indexer have first crack at the material, but
in most instances, I wouldn't let it build the index alone any more than
I'd let the grammar checker correct my prose unassisted.

Skoal,

Doug                            "Farm policy, although it's complex,
ENGSTROMDD -at- phibred -dot- com           can be explained. What it can't be
                                 is believed."
                                                - P.J. O'Rourke

***********************************************************************
The preceding opinions and positions are mine alone, and are only
coincidentally related to those of Pioneer Hi-Bred International, Inc.
***********************************************************************