Re: [Lynx-dev] wide-char support with different builds

[shalasz]

2025/01/18 10:15 ... Mouse:
> There is no one-size-fits-all character set, nor encoding, nor even
> serialization, no matter what the priests of the UTF-8 religion would
> have you believe.
>
> If you want to argue that UTF-8 is the best default, that at least is
> worth discussing.  But maintaining that there is any single "the right"
> character set, encoding, or serialization is...nonsense.  There is, at
> most, "right" for a particular use case, or set of use cases.

ASCII is ugly, Latin-1 is ugly (at the defining meeting, the member from 
France, no printer, no linguist, no typographer, against his country s 
tradition repudiated the letter OE. Of course, another member jumped at 
that and proposed multiplication and division. There is now a hole in 
the added letters), Unicode is ugly. But UTF-8 is particularly ugly. It 
has 5 message bits and 3 overhead bits, and writers in Devanagari, ... 
Malayalam, ... Hangul, ... hiragana & katakana, ... above all in 
Chinese, find that their text files are huge, bigger than if entered in 
a 16-bit code (24-bit code, anyone?) with all its surrogates.
In bijective binary ("Bijective numeration" in Wikipedia) using seven 
bits and one marker-bit, one can get up to only three bytes, and 
inherent uniqueness. Anyone interested in losing UTF-8 s deliberate 
redundancies?