Re: Upcoming loss of usability of Emacs source files and Emacs.

[Matthew Carter]

Marcin Borkowski <address@hidden> writes:

> On 2015-06-23, at 21:21, Eli Zaretskii <address@hidden> wrote:
>
>> They shouldn't be expected to be able to distinguish between
>> confusingly similar, but subtly different characters.  Let me give you
>> a few examples:
>>
>>  "
>>  “
>>  ”
>>  ‟
>>  〞
>>  〟
>>  “
>>  „
>
> 1. My personal opinion is: I prefer `...', but I could live with ‘...’
> provided that the input and search problems will be solved in
> a satisfactory way and that I could find a font which renders them
> better than the default font in my Emacs (Ubuntu Mono), where the
> Unicode quotes look terrible.  (I guess that when I finally get rid of
> this pile of crap which Ubuntu has become, it will get better
> automatically...)
>
> 2. I would like to point out to all Unicode fanboys (no offence, please,
> I'm also a fanboy, though definitely not of Unicode) that AFAIK (though
> I can't find any source now) there is a bug in Unicode: U+201C should
> really be split into two characters, "LEFT DOUBLE QUOTATION MARK" and
> "GERMAN RIGHT DOUBLE QUOTATION MARK" (see
> e.g. 
> https://en.wikipedia.org/wiki/Quotation_mark#German_.28Germany_and_Austria.29).
> While their shape is identical, their bounding boxes should be mirror
> images of each other.
>
> So please don't tell me that Unicode "solves the problem of nice
> quotes".  It does, but at the same time it introduces problems of its
> own.  (A funny one: see http://unicodelookup.com/#math and try to guess
> what happened to 0x1D455, or "mathematical italic small h".  A stupid
> one: U+FE18, with a typo in the name.)  In fact, Unicode seems to be
> fundamentally broken by design, since it identifies characters by
> numbers instead of names.  This basically excludes any language with an
> infinite[1] alphabet.
>
> [1] "infinite" in a practical sense, not a theoretical one, of course.
> It is closer to "potential infinity".  I mean here a language where it
> is legal to create new characters on the fly, when needed, provided they
> are combined from some basic shapes according to the rules.
>
> Best,

Having done a lot of web development work (where unicode and different
character sets between databases, web clients etc.) have caused many
headaches, I would hate to see something that should be kept to display
formats (rendered markdown, LaTeX, .odt files) in my primarily ASCII 
based terminal (emacs -nw mode).

It increases the difficulty in working with the characters (I'd always
prefer to type `...' than whatever is required to insert curlies) and it
will break some web based tutorials/guides where the writer/blogger
copies and pastes the unicode quotes out of some part of emacs and the
web portion fails on it (similar to users copying and pasting curly
quotes from Microsoft Word into various WYSIWYG editors).

In my Common Lisp personal projects I love taking advantage of the
non-ASCII character set that Common Lisp can use (ƒ, λ, α, ψ etc.), but
I would never expect someone else to maintain or touch said codebase.

-- 
Matthew Carter (address@hidden)
http://ahungry.com