|
contact
|
|
location: London, UK
louisiana at acm dot org
m +44 (0)7810 260 637
|
|
|
Articles
Indexes - not just for the 'back
of the book'
by Louise Ferguson, 1999
Indexes are not just for the 'back of
the book'. On the web, many sites provide menus and search facilities,
but to get an overall picture of a website sitemaps and indexes can be
invaluable.
The trouble with 'search' is that it
assumes the user knows what is on the site already - generally a completely
unwarranted assumption. Moreover, search will drag up every occurrence
of a word - not just the signitificant ones. If I type the word 'usability'
or 'PDA' into the HP (www.hp.com) search engine, for example, the site's
search engine comes back with hundreds of seemingly undifferentiated results.
This is not to denigrate search
- it is useful as part of a set of access mechanisms to a website, but
should not be the only way in for all visitors.
Menus also have their issues, but by
their nature they are very limited and broad-brush, and assume users know
how you have organised information on your site - often another completely
unwarranted assumption.
Can computers index? The short answer is no. I quote from
the American Society of Indexers:
"Computers can easily construct a concordance
(a list of words or phrases and where they appear), but this is not
an index, and is not very useful to someone looking for information.
The so-called automatic indexing software programs now appearing on
the market are simply not up to the task..."
[...] "indexing involves a little bit of manipulating
words appearing in a text, which computers can do, and a lot of understanding
and organizing the ideas and information in the text, which computers
cannot do and will not do for many years to come."
For example, a text on protective gloves might include
a discussion of "surgical gloves, how they get punctured and how
they are tested for integrity, but might never use the word 'holes'".
But a reader would expect to find this word in the index and be directed
to the appropriate section. An indexer handles many such issues in every
text.
Collecting and selecting index entries
- What level of detail?
- What to put in and what to leave out?
This depends on the type of user you expect (e.g. specialist
or general consumer), but there is often a compromise to be achieved between
being too sparse and being cumbersome.
When selecting entries, bear in mind:
- The needs of users - who are they likely to be?
- Common sense - is coverage of different topics at about
the same level, for example
Then
- Think about meanings, not words - a good index includes
not just terms, events and people mentioned in the text, but relevant
concepts that are not included there (see the example above about 'gloves'
and 'holes')
- Don't record every mention - only the significant ones,
otherwise you end up with long undifferentiated lists, just as with
search
- Don't be concerned with the final structure at the beginning
- keep an open mind and be prepared to change your groupings
Structuring entries
By now, there may already be a pattern of main and sub-entries,
but bear in mind the type of index you are designing:
- analytic (many short individual entries), or
- synthetic (groupings chosen to show associations)
This will help to ensure consistency of approach.
Wording entries
- Try to think like a user: what are your users looking
for and how will they look for it? A specialist audience will look for
different terms to the general user. A prospective customer will have
a different viewpoint and vocabulary to a supplier and to an exisitng
customer.
- Headings should begin with specific concrete words rather
than abstract general ones. Prefer nouns to adjectives e.g. avoid 'abnormal
psychology' - instead use 'psychology, abnormal'
- Try using plurals - they often work better
- Sub-headings are the best place for adjectives
- Don't split entries: e.g. don't have separate entries
for 'textile' and 'fabric' or for 'vegetation' and 'flora'. Put all
references under a single heading, and direct the user to that heading
if necessary. This will help to reduce confusion
Cross-references
These are useful but should be used sparingly:
- 'See' - use where the relevant references are
listed under a different entry
- 'See also' - use where some additional entries
are also highly relevant (a related aspect of the same subject, for
example)
Order of Entries
- Alphabetical order is the general rule, but should this
be letter-by-letter or word-by-word? Whichever you choose, be consistent.
- Gives names of people in full, and order by surname.
There are special rules for Arabic, Chinese and other non-Western names,
and elements such 'von'.
- Titles of publications should be listed according to
the first significant word - in the case of The Guardian, this is 'Guardian',
for example
- Put any cross-references at the end of an entry - all
words that are not index headings should be in italics
Example indexes
American
Society of Indexers - an index constructed according to the rules.
The Skeptic's
Dictionary - Not exaxtly by the book, but fun nevertheless.
Read more
articles on this website
Check out shorter pieces on the blog
|
|

|