RE: scan-info-node-length.el

[Drew Adams]

>     > (657 "Sequence Functions")
>     > (481 "GNU Free Documentation License")
>     > (372 "pcase Macro")
>     > (363 "Special Properties")
>     > (329 "Frame Layout")
>     > (304 "Garbage Collection")
>     > (303 "Low-Level Font")
>     > (301 "Changing Files")
> 
> These modes really ought to be subdiided.

Sequence Functions and pcase Macro are recent additions.
Perhaps there is a tendency toward long nodes in an
initial attempt to doc something, with such long nodes
being refactored later.

BTW, `count-lines' counts blank lines too.  Maybe
`scan-info-node-length' or similar could offer counts
of more than just lines (e.g., using optional args).

`count-words-region' offers lines, words, and chars
(interactively, at least).  Each of those gives some
(and a different) indication of the complexity of doc.
 
> Splitting a node well is not a mechanical task -- it calls for real
> writing.

Indeed.  And thank you to TTN for recently trying to
improve the `pcase' doc (bug #31311).