Re: Apropos commands and regexps

[Eli Zaretskii]

> From: address@hidden
> Date: Thu, 16 May 2002 13:04:13 +0200
> 
> The user enters a query.  The system searches items of documentation
> and computes a score for each one.  Then the items with the highest
> score come out first.

There is an alternative approach:

The user enters a query.  The system does the search and presents a
menu of possible refinements of the original search spec.  The user
chooses one of the possibilities, and the process repeats, until the
list of possible hits is shorter than some predefined value; when
that happens, the list of hits is displayed.

The advantage of this method is twofold:

  - You don't need to invent a good scoring system.

  - The user never needs to wade through gobs of hits, trying to
    figure out which one is relevant to his/her query.

(If this description doesn't explain the suggestion, I can craft a
ficticious example that might help.)

I generally find scoring a poor means for me to decide whether the hit
is relevant.  When I google, for example, I find myself examining the
search words shown with surrounding text much more than looking at the
scores.  But if the number of hits shown is large, my method is not
very efficient, and can be even frustrating; thus the suggestion for
interactively refining the search before showing the hits.

> When searching in info files, some subdivision makes sense I think.
> Should each node be considered a retrievable item, or is another
> subdivision more sensible?

I'd say, as the first approximation, search node names, chapter/section
names, index entries, and glossary items.