Re: wishlist: implicit anchor for @deffn etc

[Eli Zaretskii]

> Cc: address@hidden
> From: Per Bothner <address@hidden>
> Date: Tue, 31 Oct 2017 21:46:10 -0700
> 
> >> How about @dref ?
> > 
> > I'm not good with names, but @dref sounds like a command from the *ref
> > family (@xref, @pxref, etc.), which it isn't.
> 
> Yes and no. Like @uref and the *ref family, it is a link (reference) to
> another location.  I.e. a reference to a definition.
> 
> So @dref works quite well.

Perhaps there's a misunderstanding.  I was talking about a command
which would _establish_ an anchor, and be used instead of @defun.
@dref is a fine name for a command that _goes_ to that anchor, which
should be placed where you want a link to where the function was
described, and where the command I was talking about was supposed to
be used.

> > I don't see what does this have to do with reliability.  Index-search
> > is a very basic and important feature of Info.  If you are arguing for
> > having each indexed phrase that is also literally in the text
> > (something that frequently doesn't happen), then I think I'm in
> > favor.  But that's a separate discussion, I think.
> 
> Just that an index may have multiple entries for a given NAME.
> It may be a command or a function.  An index could also usefully have entries
> to background information or usages examples.  When using a command or a 
> function
> is a discussion or an example, it can be helpful to link to the specific 
> definition
> of the name as used in the discussion or example.

Good indexing is a non-trivial job.  Duplication of iundex entries
should be avoided, or they should be qualified by some context, to
make them unique.  Discussions about the function that are not its
detailed description should be either not indexed or indexed with
appropriately qualified entries.  E.g., the place where the function
is described should have just

  @findex FUNCTION

and the place where it is mention in relation to some other TOPIC
should have

  @findex FUNCTION, and TOPIC

> Furthermore, there is familiarity of links:  Even if the 'i' keystroke is 
> implemented
> in a webpage (using JavaScript) it is useful to have an actual link to the 
> definition,
> because everybody knows what a link is, and most people may not know the 'i' 
> shortcut.

Users of Info should know.