Re: [PATCH] docs: submitting-patches - change non-ascii character to ascii

[Frank Rowand]

On 07/21/17 10:27, Jonathan Corbet wrote:
> On Thu, 20 Jul 2017 18:30:55 -0700
> frowand.list@xxxxxxxxx wrote:
> 
>> Documentation/process/submitting-patches.rst contains a non-ascii
>> character.  Change it to the ascii equivalent.
> 
> You should know better than to tell somebody like me that a hyphen and an
> m-dash are equivalent! :)

OK, so they aren't totally equivalent, but close enough.  :)  Should I
have said analog instead of equivalent?  And would you prefer '--' to '-'?


> I don't have any real objection to this change, but I am curious: is the
> m-dash creating a problem somewhere?  We have plenty of non-ASCII
> characters in Documentation/ and beyond, why change this one?  Or to put
> it another way, do you think we should have an ASCII-only policy for
> documentation files?

Ascii is a lowest common denominator.  I can view and manipulate the file
with any common text editor and common text utilities (eg, cat, grep, etc)
on pretty much any Linux system that I walk up to.  I don't need to go
to any effort to try to figure out what a non-ascii character is (which
is exactly what prompted my patch -- I wanted to know what the character
my patch modifies is).

Yes, I can change my terminal emulator character encoding to UTF-8, and
change my LANG to en_US.UTF-8.  And now vi and cat show the correct
m-dash character.  But then how do I grep for m-dash in files?  Google
tells me I might be able to <ctrl> + <shift> + u hex_value_of_mdash
to enter an mdash, but I sure don't know what the hex value of mdash
is.  Plus I need to be observant enough to notice that the string I
am grepping for contains an m-dash instead of a dash.  And why should
I assume "en_US" as the prefix to my UTF-8 LANG?

To answer your second question, I would _prefer_ ASCII-only except for
cases where being limited to ASCII is restricting the ability to
convey information (properly).

I would reverse your question, and ask what is the added value of
non-ascii characters __in cases similar to this one__, that justifies
the negative impact?  (Please don't answer what the added value is
in cases that are not similar to this one.  I know the answer to that
question is different.)


> Thanks,
> 
> jon
>