Re: Feature request(?): adding annotations for nodes?

[Eli Zaretskii]

Pavel Janík wrote:
> 
> My main reason for all that is that we see many valuable bug-reports
> created by M-x report-emacs-bug in GNU Emacs. But we do not have that many
> feedbacks about the quality of manuals. Only from conferences and
> mailing-lists where users report their problems. But M-x report-info-notes
> (or similar) can be valuable for manual writers (take all/some notes about
> all/some nodes and send them somewhere).

If the goal is to give us feedback about manual quality, I think we
should consider a broader range of possibilities.  For example, traces
of index searches and string searches (the `s' command) might suggest
changes in the indexing and additions to the Glossary.

As for the specific feature you suggest, don't we have a similar
facility in the bookmark package?

As an aside, it is my impression that we should add a kind of
high-level interactive guide to the Info system, which would ``take
the user by the hand'' and help them find what they are looking for by
letting user pick one of a few alternatives in several steps, with
each step progressively narrowing the set of possible topics, so that
eventually we have only one node that fits the requirement, and we then
display that node.

For exampe, you begin by asking whether a feature pertains to display,
changing text, or moving through text.  If "moving" is chosen, you then ask
if it's moving by small chunks like characters or words, or by
screen-fuls.  Etc., etc.

Internally, this could be built around some hierarchical data structure
which is either crafted by hand or built by some code based on the indices
and the chapter/secion names in the manual.

The rationale for such a feature seems to be that there's a group of users,
mostly newbies, who don't get along well with the features we currently
provide for searching the documentation.  The features we have require the
user to come up with strings or regexps which are then submitted to the
commands.  That is, a user needs to construct a search string out of thin
air, out of a virtually infinite set of possibilities.  (Of course, in
practice, that set is heavily constrained by what the user is looking for,
but users sometimes have a mental block that prevents them from narrowing
down the set of possible strings.)  When presented with a small set of
alternatives to choose from, those users will find their way more easily, I
think.

(Several Windows programs have such a feature, they call it ``a wizard''. 
Since we loathe MS, someone suggested to call this ``a saint'' instead, and
even invented a name--St. iGNUcius.)