A difficult(?) question about keywords and indexing

[brunos]

Hi all,

I had recently some doubts about keywords in online help, regarding in particular whether to repeat or not the chapter related keywords in child topics.

Let's assume I have a chapter named "Printing pages and page layouts". The chapter contains 2 child-chapters, and about 20-25 topics, e.g.:

Code: Select all

Printing pages and page layouts [chapter]
  Overview [1st level topic]
  Procedures of printing pages and page layouts [child chapter]
     Print pages with a predefined configuration   [topics in the child chapter]
     Print pages with an ad-hoc setting
     Print color separation of pages
     Print with auto-fit to media size
     Font downloading options

and so on.

The chapter topic contains keywords such as:

Code: Select all

pages,
pages, printing
page layouts,
page layouts, printing
printing,
printing, pages
printing, page layouts

but not the keywords, related to all child topics, e.g. it doesn't mention
"overview", nor "fonts downloading".

Now, my question regards the keywords for child topics in the mentioned
chapter. Of course, one would surely enter the topic related keywords, e.g. for the Overview topic, one would add :

Code: Select all

pages,
pages, printing overview
overview of pages printing

but my question is whether all the keywords used for the chapter topic are to be repeated in this topic too?

I had a long discussion about advantages and drawbacks of each approach. If all chapter level keywords are repeated in each child topic, the TOC is practically replicated in the index. Since it make the document heavier, and it's painful to achieve, does it makes sense? Does it help users to find information?

If the chapter level keywords are not repeated in child topics, the users
who search in the index for:

Code: Select all

printing, pages

and double-click it, will never pick, for example, the "Overview" topic nor
the "Font downloading options" topic.

What would be the best practice (if there's one)?

Of course, the above example is not the worst possible; there could exist
chapters with 4 or more levels of child-chapters and 50 or more topics.

Thank you for your comments on this, and all the recommendations for the resources which treat this matter.

Regards,
Bruno

[Tim Green]

Bruno:

Basically, it depends on the information in the individual topics. It's completely impossible to make a standard rule for this; you can't say "always put the keywords there if your topics are structured like this". It depends on the content, not on the structure.

The question I always ask is, will a link from this keyword to this topic be helpful or not? Too many keywords with too many irrelevant links are just as bad as, and possibly worse than, too few, because they make the index useless.

I always try to keep the number of keywords to an absolute minimum, whilst also attempting to think of all the important and relevant variants the user might try to enter. I also try to make sub-entries for different themes, e.g.

Code: Select all

Rabbits
Rabbits, white
Rabbits, black
Rabbits, breeding
Rabbits, cooking and recipies

so that clicking on any entry always produces links to the same theme or group of themes, rather than a mish-mash that is confusing rather than helpful.

To summarize: I really think it's always the topic content that decides this question, not the structrure.

[brunos]

Hi Tim,

I agree it depends on information in topics, but I disagree that it's completely impossible to make a rule. For almost anything (including even rather unpredictable human behaviour) it is possible to set rules; perhaps not 100% general, but still very useful.

What I'm looking for, is the best practice for a specific case I explained in my question; what I'm getting as answers are - 'it depends on content'.

What I'm talking about indeed are methods to validate the usability of an index. I can hardly believe it's all left to a subjective idea about 'being useful' or not? Being useful to which users? For which level of knowledge? There must be some analysis around, or some documentation where some cases are explained and validated. Indexing can't be only based on "as few as possible" - or no?

For example, you surely know the rule about reverting locators (e.g. if you use printing, articles, that you're supposed to use also articles, printing). Useful or not, respected or not, it's a rule. I believe it's even useful. When I was creating my first index, some number of years ago, I didn't know that rule, and the index was not so good as it could be - if I'd knew the rule.

I'm sure out there's a rule on grouping and repetition of keywords, and sooner or later it will pop up. The fact that you and me are not aware of it, is worse for us than for the rule, right?

Anyway, thank you for replying.

To comfort you: I posted the same question to HATT@yahoo.com groups. All very dear ladies and gentlemen over there who happily spend hours on discussing about "whether it make sense to provide help for menu items" (yes, no, maybe, depends), or "eHelp mailings look like spam" (yes, no, maybe, depends), or "RobohelpX is better than tool Y", in 5 days did produce 1 (one) reply on my question, stating a big truth that "a rule can't be set"). And that was it. Obviously, an OT question. Omg.

Beh, I will continue to hope.

Regards,
Bruno

[Tim Green]

I understand what you are saying but I still stand by my reply: I don't see any way to create hard and fast rules for this, it really depends on the content of the topics -- and also on your target audience and their level of knowledge and all sorts of other imponderables.

There's no "right" way to make an index. Perhaps the closest thing to a rule might be, "As much as necessary and as little as possible." I also don't think you can get around making human decisions for every single case, at least not if you want to make a really good index. In one case it might be useful to include the same keyword for all levels, in another it might not -- in the same help project. It simply depends.

BTW: I'm sure that people on HATT and many other general discussion groups would be willing to discuss subjects like that for hours, days or even weeks. I often wonder where the people involved in these discussions find all their time. Among other things, that's why I almost never use Usenet any more...

[Matt]

It is probably optimistic to expect a response to a reply to posting that is over two years old, but I live in hope...

My reply is really a question. Did you ever find any material, Bruno, or Tim, or anyone else who may see this, that defines the best practice for indexing? Is there a (reasonably brief) guide, book, manual, paper or online entry out there that tackles this thorny topic?

My team has just entered the final phase of the first part of our help file and are tackling the indexing/ keywords rubric and none of us have any significant prior indexing experience. We could really use an acknowledged authority on indexing to refer to.

Also, if you know of any other interesting discussions/threads on this forum...

Regards,

Matt

[fiona]

Two links that may be useful to you:

http://www.asindexing.org/site/index_usability.pdf

(this is a summary of research conducted into what end users expect of an index)

and

http://www.backwordsindexing.com/Novice/Process.html

(a number of other useful links are included here too).

As guidelines, both sites make interesting reading.

Fiona