bug#1144: closed by Chong Yidong <address@hidden> (Re: 23.0.60; No doc string for `symbol-at-point' and other, similar functions)

[Drew Adams]

> >> Must I send a separate bug for each such function? I thought 
> >> when your complete response was "Done" that you had taken 
> >> care of all such functions.
> >> 
> >> The following functions are all in the same boat. They are 
> >> all functions for which users will want to see the doc.
> >> 
> >> beginning-of-thing
> >> end-of-thing
> >> in-string-p
> >> end-of-sexp
> >> beginning-of-sexp
> >> thing-at-point-bounds-of-url-at-point
> >> forward-whitespace
> >> forward-symbol
> >> forward-same-syntax
> >> word-at-point
> >> sentence-at-point
> >> form-at-point
> 
> No, Chong added documentation to the autoloaded, "external" functions,
> but not the purely internal thingatpt functions.  And I think 
> that's the right decision.

No, it's not the right thing.  Emacs users include Emacs-Lisp users.  Users of
thingatpt.el, in particular, are mainly Emacs-Lisp users.

Whether to document something has nothing to do with whether it is a command or
is autoloaded.  That is, those predicates do not at all suffice to determine
whether something should be documented.

And for Emacs there really is no such thing as an "internal" function.  Yes, we
need not document every function or variable, and yes, some functions and
variables we generally do not expect users to fiddle with.  But Emacs is 100%
open, and it should document itself, as originally intended.