Re: [gnu.org #1363250] ASCII maintain.txt is no longer ASCII

[Antonio Diaz Diaz]

Alfred M. Szmidt wrote:
> My machine is over 10 years old.  I can't even remeber when you
> _couldn't_ modify the locale on a GNU system.  If your system is that
> old then you have more problems than viewing a text file! :-)

This machine is 20 years old. It is powered by a Slackware from 2007 
because nothing more modern seems to be installable on it. Years ago I 
tried to install on it the version of Slackware following the one it has 
installed, but the desktop didn't start. There is possibly not even one 
piece of software in this machine that supports UTF-8 properly. For 
example, this is how maintain.info looks on an UTF-8 X terminal (Konsole 
1.6.6 from KDE 3.5.7) viewed with info (GNU texinfo) 4.13:

-----------------------------------------------------------------------
   This document uses the gender-neutral third-person pronouns “person”
(which can be shortened to “perse”), “per”, “pers” and “perself.�\
�� These
pronouns (aside from “perse”) were promoted, and perhaps invented, by
Marge Piercy in ‘Woman on the Edge of Time’.  They are used just like
“she”, “her”, “hers” and “herself”, except that they apply rega\
rdless of
gender.  For example, “Person placed per new program under the GNU GPL,
to maintain freedom for all users of per work, and this way perse knows
perse has done the right thing.”
-----------------------------------------------------------------------

Note that it even breaks a 3-byte double quote into 3 invalid UTF-8 
characters. So, even if I "modify the locale", it still looks wrong.

But X is the easy part. The text console can't "modify the locale" to 
UTF-8. Possibly I could compile a kernel with UTF-8 support in this 
machine, but I do not want to do it because I use this machine to test 
an old but perfectly working system. I know there are users out there 
(not in the USA) with machines even less capable than this one. And in 
fact I test some of my programs on even older hardware with older 
software installed (80486 with Suse 6.3).


> It sounds more and more like your situation is not the norm, but
> rather a very special case.

In the first world, sure.


> Why can you not simply convert (or regenerate) the relevant files?

Because it is inconvenient and could be easily fixed upstream.


> Today, and for the last 10 years, UTF-8 has been the default and I
> cannot think of any kind of terminal in the last 30 years that is
> unable to display UTF-8 properly.

It is not (only) the terminal. It is the software. Given the lack of 
concern for backwards compatibility among software developers, it seems 
from very difficult to impossible to install UTF-8 capable software (or 
anything else) in old machines.


>     IMHO, the problem is clear. Using multibyte UTF-8 characters where ASCII
>     suffices, inconveniences all users lacking an UTF-8 capable machine or
>     software.
>
> Maybe we should have a poll, on how many 7-bit ASCII users there are
> versus UTF-8?  I don't see that the problem is clear at all.  And I am
> guessing that 7-bit terminals is not the norm at all.

Who said anything about 7-bit terminals? There are lots of 
ASCII-compatible encodings that are not 7-bit ASCII.


>     Moreover, what is the advantage of multibyte UTF-8 quotes for users with
>     UTF-8 capable screens? They make the text easier to understand or something?
>
> If we are talking about Info, then using UTF-8 produces a much nicer
> output in X11 -- which is what the majority of users would be using
> today.

http://www.gnu.org/prep/maintain/html_node/Web-Pages.html

"The overall goals are to support a wide variety of browsers, to focus 
on information rather than visual adornments".

FWIW, I find the vertical ASCII quotes much nicer than those horrible 
slanted multibyte UTF-8 quotes. IMHO, "much nicer output" does not 
justify inconveniencing people until they throw away their old but 
working machines, creating electronic waste in the process. For most 
people most of the time UTF-8 is a waste of resources. I find 
ISO-8859-15 superior to UTF-8 in every respect where both can be used 
(English or Spanish texts, for example). Supporting a significant 
fraction of Unicode is orders of magnitude more expensive than 
supporting an 8-bit encoding, and is one of the causes why modern 
software can't be installed (or runs horribly slow) on old machines, 
forcing the premature obsolescence of perfectly working machines. I 
consider the use of multibyte UTF-8 characters where ASCII (or 
ISO-8859-1) suffices, the use of javascript or complicated CSS in web 
pages, etc, as forms of planned obsolescence that should be avoided for 
the good of society.


> GNU always prided it self in moving forward, not backward.

I'm fine with moving forward, but I can't imagine in what way wasting 
two bytes per quote is "moving forward".


Best regards,
Antonio.