bug#16190: 23.4; M-x captialize-word works incorrectly

[Caleb Wakeman]

(paraphrased quotes)
>>> That's the intended behavior
>> Then its docstring is incorrect
> Fair enough, fixed

Erm, well...

I've had several days to think it over, and I don't agree [that it's 
fixed].  Yes, the docstring is (was) incorrect; and if it hadn't been 
incorrect, I wouldn't have reported it as a bug.

However, the function would still be doing 'The Wrong Thing' (for me, I 
can imagine that cases exist where this isn't), and if I ever used it 
again, it would only be because I forgot that it didn't do what I expected.

I think this is also the point where if I want something more done, I 
really should do it myself.  I'd take it on for sure if it were in 
elisp.  I dunno if I could do it in C.

This seems reasonable as a 'workaround' fix, though.

[I'd agree that "Severity: minor" is correct.]

(I can re-create my mental state leading up to actually sending the bug 
report, or share my mental image of what the function does, if either of 
those would be helpful.)