Re: How can I avoid unicode and use Latin1? (Was: Wrong characters with jEdit)

[stk]

On Sat, 3 Sep 2005, Han-Wen Nienhuys wrote:

> A. LilyPond actually _does_ support the Latin1 character set, as Latin1
> and Unicode coincide on the first 256 codepoints.

I don't quite see that.  If I put an e-acute (a byte of decimal value
#233) in a LilyPond file, it is skipped -- it does not appear in the PDF
output.  I have to put in the unicode equivalent, which is the two bytes
#195 #169 (where 169 = 233 - 64) in order for LilyPond to give me an
e-acute.  USASCII and unicode coincide on the first 128 codepoints, but
from what I can see, Latin1 and unicode do not correspond on byte values
#128 to #255.

> B. LilyPond does not support Latin1 encoding. This is because

>   1. It's not possible to detect the encoding of a file. Supporting
> alternate encodings implies that users have to specify the encoding via
> the command line. This is error-prone, and leads to confusion for newbies.

Not via the command line.  Via a command at the top of the LilyPond file
such as \unicode or \latin1, or some Scheme command (or even #!latin1 or
%!latin1 in a time-honoured tradition).  Having to insert a command for
including english.ly and for setting paper size to letter is error-prone
and confusing to newbies -- I know, I'm a newbie.  A command for
identifying a file's character-encoding type is no worse.

>   2.  If we do latin1, why should we not do latin2. And if we do latin1
> and 2, why not Big5?  EBCDIC?  UTF-16? tibetan-iso-8bit?  Where does it
> stop?

Well, I have to admit it's hard to argue with that.  Despite the fact that
I think that a lot of North Americans would like to have the direct Latin1
availability to which they have become accustomed, I know that at the
least, Eastern Europeans would also want Latin2 and Latin4.

Unicode only provides a way of specifying character codes for a wide
variety of symbols in the interior of a text file.  But without font files
containing the order of 64K symbols, the current fragmented font-file
situation will continue to limit what can easily be output to a screen
or a printer.  It is difficult for me to share your optimism.

> C. Unicode, not Latin1, is the future.

Maybe, but not in my lifetime.

-- Tom