RE: info-find-source

[Drew Adams]

> OK, so it is the *.info files* that one is
> disencouraged from editing? Well yeah, why
> would you do that?!

Quick correction/annotation/addition to one's
own copy of a manual.

Versus re-creating lots of stuff (after having
installed `makeinfo', if it's not installed
locally).

The existence of `makeinfo' does not obviate
the usefulness of `Info-edit'.

> The "by hand" phrasing is where the confusion
> begins because it seems to imply there is
> a better way to "edit" them. But to me it seems
> totally backward to edit the result of
> compilation

When that result is human-readable text it's
not a big deal to edit it.  Of course, if
you want the change to be reflected more
globally or to be shared etc. then you want
to only modify Texinfo source and generate
Info output.

You yourself argued for using *.info files in
plain editing mode (e.g. after `C-x n w').
Something like `Info-edit' is nowhere near as
extreme as that.  It's used for simple, quick
one-off changes or additions.

One doesn't have to argue _against_ generating
Info from Texinfo to see some utility in a 
command such as `Info-edit'.  Granted, that
utility is limited, and most people have never
even heard of it.  But that's not the same as
saying that it has no raison d'etre.
____

I mentioned that Info+ has a command,
`Info-merge-subnodes', for creating a
plain-text, prettified merge of Info nodes
(even a whole manual, but more typically a
section of a manual, however small).

Such a flat buffer can be useful sometimes
(e.g., plain-text printing, some kinds of
searching, sending excerpts), but I wouldn't
argue that `Info-merge-subnodes' is a _super_
useful command.  Some commands have limited
usefulness and use cases.  `Info-edit', like
`Info-merge-subnodes' is one such.