Search the FAQ Archives

3 - A - B - C - D - E - F - G - H - I - J - K - L - M
N - O - P - Q - R - S - T - U - V - W - X - Y - Z
faqs.org - Internet FAQ Archives

ISO 8859-1 National Character Set FAQ


[ Usenet FAQs | Web FAQs | Documents | RFC Index | Forum ]
Archive-name: internationalization/iso-8859-1-charset
Posting-Frequency: monthly
Version: 2.9889

See reader questions & answers on this topic! - Help others by sharing your knowledge
		  ISO 8859-1  National Character Set FAQ

			  Michael K. Gschwind 

                      <mike@vlsivie.tuwien.ac.at>


DISCLAIMER: THE AUTHOR MAKES NO WARRANTY OF ANY KIND WITH REGARD TO
THIS MATERIAL, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE.

Note: Most of this was tested on a Sun SPARCstation 10, running SunOS
4.1.* - other systems might differ slightly

This FAQ discusses topics related to the use of ISO 8859-1 based 8 bit
character sets. It discusses how to use European (Latin American)
national character sets on UNIX-based systems and the Internet.

If you need to use a character set other than ISO 8859-1, much of
what is described here will be of interest to you.  However, you will
need to find appropriate fonts for your character set (see section 17)
and input mechanisms adapted to you language.



1. Which coding should I use for accented characters?  
Use the internationally standardized ISO-8859-1 character set to type
accented characters. This character set contains all characters
necessary to type all major (West) European languages.  This encoding
is also the preferred encoding on the Internet.  ISO 8859-X character
sets use the characters 0xa0 through 0xff to represent national
characters, while the characters in the 0x20-0x7f range are those used
in the US-ASCII (ISO 646) character set.  Thus, ASCII text is a proper
subset of all ISO 8859-X character sets.

The characters 0x80 through 0x9f are earmarked as extended control
chracters, and are not used for encoding characters.  These characters
are not currently used to specify anything.  A practical reason for
this is interoperability with 7 bit devices (or when the 8th bit gets
stripped by faulty software).  Devices would then interpret the character
as some control character and put the device in an undefined state.
(When the 8th bit gets stripped from the characters at 0xa0 to 0xff, a
wrong character is represented, but this cannot change the state of a
terminal or other device.)

This character set is also used by AmigaDOS, MS-Windows, VMS (DEC MCS
is practically equivalent to ISO 8859-1) and (practically all) UNIX
implementations.  MS-DOS normally uses a different character set and
is not compatible with this character set. (It can, however, be
translated to this format with various tools. See section 5.)

Footnote: Supposedly, IBM code page 819 is fully ISO 8859-1 compliant.


ISO 8859-1 supports the following languages:
Afrikaans, Basque, Catalan, Danish, Dutch, English, Faeroese, Finnish,
French, Galician, German, Icelandic, Irish, Italian, Norwegian,
Portuguese, Spanish and Swedish.

(Reportedly, Welsh cannot be handled due to missing \^{w} and \^{y}.)

(It has been called to my attention that Albanian can be written with
ISO 8859-1 also.  However, from a standards point of view, ISO 8859-2
is the appropriate character set for Balkan countries.)

ISO 8859-1 is just one part of the ISO-8859 standard, which specifies
several character sets:
8859-1	Europe, Latin America, Caribbean, Canada, Africa
8859-2  Eastern Europe
8859-3  SE Europe/miscellaneous (Esperanto, Maltese, etc.) 
8859-4  Scandinavia/Baltic (mostly covered by 8859-1 also)
8859-5  Cyrillic
8859-6  Arabic
8859-7  Greek   (the same as ELOT928 ???)
8859-8  Hebrew
8859-9  Latin5, same as 8859-1 except for Turkish instead of Icelandic
8859-10 Latin6, for Lappish/Nordic/Eskimo languages

Unicode is advantageous because one character set suffices to encode
all the world's languages, however very few programs (and even fewer
operating systems) support wide characters. Thus, only 8 bit wide
character sets (such as the ISO 8859-X) can be used with these
systems.  Unfortunately, some programmers still insist on using the
`spare' eigth bit for clever tricks, crippling these programs such
that they can process only US-ASCII characters.


Footnote: Some people have complained about missing characters,
          e.g. French users about a missing 'oe'.  Note that oe is 
          not a character, but a typographical ligature (a combination of two
          characters for typographical purposes).  Ligatures are not 
          part of the ISO 8859-X standard.  (Although 'oe' used to 
          be in the draft 8859-1 standard before it was unmasked as
          `mere' ligature.)

          Two stories exist for the removal of the oe:
          (1) argues that in the final session, the French admitted
              that oe was only a ligature.  This prompted the
              committee to remove it.
          (2) argues that the French member missed the session and the
              members from the other countries simply decided to
              remove it.  (If this is true, where were the Swiss and 
              Belgians?)

          Note that the oe ligature is different from the 'historical
          ligature'  which is now considered a letter in Nordic
          countries and cannot be replaced by the the latters 'ae'.

A semi-official statement about the missing oe:
4.  The present part 1 reflects the position of AFNOR of 1987. It may be
    that this is regretted now, but no action can be taken before AFNOR
    makes clear what it wants now. Canada may try to convince AFNOR
    that something should be done, but as far I know the SC2-FRANCE is
    no longer active. They do not respond to letter ballots, nor to
    E-mail.


2. Getting your terminal to handle ISO characters.
Terminal drivers normally do not pass 8 bit characters. To enable
proper handling of ISO characters, add the following lines to your
.cshrc:
----------------------------------
tty -s
if ($status == 0) stty cs8 -istrip -parenb 
----------------------------------
If you don't use csh, add equivalent code to your shell's start up
file.  

Note that it is necessary to check whether your standard I/O streams
are connected to a terminal. Only then should you reconfigure the
terminal driver.  Note that tty checks stdin, but stty changes stdout.
This is OK in normal code, but if the .cshrc is executed in a pipe,
you may get spurious warnings :-(

If you use the Bourne Shell or descendants (sh, ksh, bash,
zsh), use this code in your startup (e.g. .profile) file:
----------------------------------
tty -s
if [ $? = 0 ]; then
        stty cs8 -istrip -parenb >&0
fi
----------------------------------

Footnote: In the /bin/sh version, we redirect stdout to stdin, so both
tty and stty operate on stdin.  This resolves the problem discussed in
the /bin/csh script version.  A possible workaround is to use the
following code in .cshrc, which spawns a Bourne shell (/bin/sh) to
handle the redirection:
----------------------------------
tty -s
if ($status == 0) sh -c "stty cs8 -istrip -parenb >&0"
----------------------------------



3. Getting the locale setting right.
For the ctype macros (and by extension, applications you are running
on your system) to correctly identify accented characters, you
may have to set the ctype locale to an ISO 8859-1 conforming
configuration. On SunOS, this may be done by placing
------------------------------------
setenv LANG C
setenv LC_CTYPE iso_8859_1
------------------------------------
in your .login script (if you use the csh). An equivalent statement
will adjust the ctype locale for non-csh users.

The process is the same for other operating systems, e.g. on HP/UX use
'setenv LANG german.iso88591'; on IRIX 5.2 use 'setenv LANG de'; on Ultrix 4.3
use 'setenv LANG GER_DE.8859' and on OSF/1 use 'setenv LANG
de_DE.88591'.  The examples given here are for German.  Other
languages work too, depending on your operating system.  Check out
'man setlocale' on your system for more information.

*****If you can confirm or deny this, please let me know.*****
Currently, each system vendor has his own set of locale names, which
makes portability a bit problematic.  Supposedly there is some X/Open
document specifying a

	<language>_<country>.<character_encoding>

syntax for environment variables specifying a locale, but I'm unable
to confirm this.  

While many vendors know use the <language>_<country> encoding, there
are many different encodings for languages and countries. 

Many vendors seem to use some derivative of this encoding:
It looks as if <language> is the two-letter code for the language from
ISO 639, and <country> is the two-letter code for the country from ISO
3166, but I don't know of any standard specifying <character_encoding>.

An appropriate name source for the <character_encoding> part of the
locale name would be to use the character set names specified in RFC
1345 which contains names for all standardized character sets.
(Preferably, the canonical name and all aliases should be accepted,
with the canonical name being the first choice.)  Using this
well-known character set repository as name source would bring an end
to conflicting names, without the need to introduce yet another
character set directory with the inherent dangers of inconsistency and
duplicated effort.
*****If you can confirm or deny this, please let me know.*****



Footnote on HP/UX systems:
As of 10.0, you can use either german.iso88591 or de_DE.iso88591 (a
name more in line with other vendors and developing standards for
locale names).  For a complete listing of locale names, see the text
file /usr/lib/nls/config.  Or, on HP-UX 10.0, execute locale -a . This
command will list all locales currently installed on your system.



4. Selecting the right font under X11 for xterm (and other applications)
To actually display accented characters, you need to select a font
which does contains bit maps for ISO 8859-1 characters in the
correct character positions. The names of these fonts normally
have the suffix "iso8859-1". Use the command
# xlsfonts 
to list the fonts available on your system. You can preview a
particular font with the 
# xfd -fn <fontname>
command.

Add the appropriate font selection to your ~/.Xdefaults file, e.g.:
----------------------------------------------------------------------------
XTerm*Font: -adobe-courier-medium-r-normal--18-180-75-75-m-110-iso8859-1
Mosaic*XmLabel*fontList: -*-helvetica-bold-r-normal-*-14-*-*-*-*-*-iso8859-1
----------------------------------------------------------------------------

While X11 is farther than most system software when it comes to
internationalization, it still contains many bugs.  A number of bug
fixes can be found at URL http://www.dtek.chalmers.se:80/~maf/i18n/.

Footnote: The X11R5 distribution has some fonts which are labeled as
ISO fonts, but which contain only the US-ASCII characters. 



5. Translating between different international character sets.
While ISO 8859-1 is an international standard, not everybody uses this
encoding. Many computers use their own, vendor-specific character sets
(most notably Microsoft for MS-DOS).  If you want to edit or view files
written in different encoding, you will have to translate them to an
ISO 8859-1 based representation. 

There are several PD/free character set translators available on the
Internet, the most notable being 'recode'.  recode is available from
URL ftp://prep.ai.mit.edu/u2/emacs.  recode is covered by FSF
copyright and is freely redistributable.  

The general format of the program call is one of:

recode [OPTION]... [BEFORE]:[AFTER] [FILE]

The second form is the common case.  Each FILE will be read assuming
it is coded with charset BEFORE, it will be recoded over itself so to
use the charset AFTER.  If there is no such FILE, the program rather
acts as a filter and recode standard input to standard output.

Some recodings are not reversible, so after you have converted the
file (recode overwrites the original file with the new version!), you
may never be able to recontruct the original file.  A safer way of
changing the encoing of a file is to use the filter mechanism of
recode and invoke it as follows:

recode [OPTION]... [BEFORE]:[AFTER] <[OLDFILE] >[NEWFILE]

Under SunOS, the dos2unix and unix2dos programs (distributed with
SunOS) will translate between MS-DOS and ISO 8859-1 formats.

It is somewhat more difficult to convert German, `Duden'-conformant
Ersatzdarstellung ( = ae,  = ss or sz etc.)  into the ISO 8859-1
character set.  The German dictionary available as URL
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/dicts/ also
contains a UNIX shell script which can handle all conversions except
ones involving  (German scharfes-s), as for `ss' this change is more
complicated.

A more sophisticated program to translate Duden Ersatzdarstellung to
ISO 8859-1 is Gustaf Neumann's diac program (version 1.3 or later)
which can translate all ASCII sequences to their respective ISO 8859-1
character set representation.  'diac' is available as URL
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/.

Translating ISO 8859-1 to ASCII can be performed with a little sed
script according to your needs.  But be aware that
* No one-to-one mapping between Latin 1 and ASCII strings is possible.
* Text layout may be destroyed by multi-character substitutions,
  especially in tables.
* Different replacements may be in use for different languages,
  so no single standard replacement table will make everyone happy.
* Truncation or line wrapping might be necessary to fit textual data
  into fields of fixed width.
* Reversing this translation may be difficult or impossible.
* You may be introducing ambiguities into your data.



6. Printing accented characters.

6.1 PostScript printers
If you want to print accented characters on a postscript printer, you
may need a PS filter which can handle ISO characters. 

Our Postscript filter of choice is a2ps, the more recent version of
which can handle ISO 8859-1 characters with the -8 option.  a2ps V4.3
is available as URL ftp://imag.imag.fr/archive/postscript/a2ps.V4.3.tar.gz.

If you use the pps postscript filter, use the 'pps -ISO' option for
pps to handle ISO 8859-1 characters properly.


6.2 Other (non-PS) printers:
If you want to print to non-PS printers, your success rate depends on
the encoding the printer uses. Several alternatives are possible:

* Your printer accepts ISO 8859-1:
  You're lucky. No conversion is needed, just send your files to the
  printer. 


* You printer supports a PC-compatible font:
  You can use the recode tool to translate from ISO 8859-1 to this
  encoding. (If you are using a SunOS based computer, you can also use
  the unix2dos utility which is part of the standard distribution.)
  Just add the appropriate invocation as a built-in filter to your
  printer driver.   

  At our site, we use the following configuration to print ISO 8859-1
  characters on an IBM Proprinter XL :

  /etc/printcap

lp|isolp|Line Printer with ISO-8859-1:\
        :lp=/dev/null:\
        :sd=/usr/spool/lpd/lp:mx#0:if=/usr/spool/lpd/iso2dos.sh:rs:
rawlp|Lineprinter:\
	:lp=:rm=lphost.vlsivie.tuwien.ac.at:rp=lp:sd=/usr/spool/lpd/rawlp:rs:

  /usr/spool/lpd/iso2dos.sh

#!/bin/sh
if /usr/local/gnu/bin/recode latin-1:ibm-pc | /usr/ucb/lpr -Prawlp
then
  exit 0
else
  exit -1
fi

  
* Your printer uses a national ISO 646 variant (7 bit ASCII
  with some special characters replaced by national characters):
  You will have to use a translation tool; this tool would
  then be installed in the printer driver and translate character
  conventions before sending a file to the printer.  The recode
  program supports many national ISO 646 norms.  (If you add do
  this, please submit it to the maintainers of recode, so that it can
  benefit everybody.)
  
  Unfortunately, you will not be able to display all characters with 
  the built-in characters set. Most printers have user-definable
  bit-map characters, which you can use to print all ISO characters. 
  You just have to generate a pix-map for any particular character and
  send this bitmap to the printer.  The syntax for these characters
  varies, but a few conventions have gained universal acceptance
  (e.g., many printers can process Epson-compatible escape sequences).


* Your printer supports a strange format:
  If your printer supports some other strange format (e.g. HP Roman8,
  DEC MCS, Atari, NeXTStep, EBCDIC or what have you), you have to add a
  filter which will translate ISO 8859-1 to this encoding before
  sending your data to the printer.  'recode' supports many of these
  character sets already.  If you have to write your own conversion 
  tool, consider this as a good starting base. (If you add support for
  any new character sets, please submit your code changes to the
  maintainers of recode).

  If your printer supports DEC MCS, this is nearly equivalent to ISO 
  8859-1 (actually, it is a former ISO 8859-1 draft standard. The only
  characters which are missing are the Icelandic characters (eth and
  thorn) at locations 0xD0, 0xF0, 0xDE and 0xFE) - the difference is
  only a few characters.  You could probably get by with just sending
  ISO 8859-1 to the printer. 


* Your printer supports ASCII only:
  You have several options:
  + If your printer supports user-defined characters, you can print all
    ISO characters not supported by ASCII by sending the appropriate
    bitmaps.  You will need a filter to convert ISO 8859-1 characters
    to the appropriate bitmaps.  (A good starting point would be recode.)
  + Add a filter to the printer driver which will strip the accent
    characters and just print the unaccented characters. (This
    character set is supported by recode under the name `flat' ASCII.)
  + Add a filter which will generate escape sequences (such as 
    " <BACKSPACE> a for Umlaut-a (), etc.) to be printed.  Recode 
    supports this encoding under the name `ascii-bs'.

Footnote: For more information on character translation and the
'recode' tool, see section 5.



7. TeX and ISO 8859-1
If you want to write TeX without having to type {\"a}-style escape
sequences, you can either get a TeX versions configured to read 8-bit
ISO characters, or you can translate between ISO and TeX codings.

The latter is arduous if done by hand, but can be automated if you use
emacs. Simply add the following line
to your .emacs startup file. This mode will perform the necessary
translations for you automatically:
------------------
(require 'iso-cvt)
------------------

If you want to configure TeX to read 8 bit characters, check out the
configuration files available in URL
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit.

In LaTeX 2.09 (or earlier), use the isolatin or isolatin1 styles to
include support for ISO latin1 characters.  Use the following
documentstyle definition: 
\documentstyle[isolatin]{article}

isolatin.sty and isolatin1 are available from all CTAN servers and
from URL ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit. (The isolatin1
version on vlsivie is more complete than the one on CTAN servers.)

There are several possibilities in LaTeX 2e to provide comprehensive
support for 8 bit characters:

The preferred method is to use the inputenc package with the latin1
option.  Use the following package invocation to achieve this:
\usepackage[latin1]{inputenc}

The inputenc package should be the first package to be included in the
document.  For a more detailed discussion, check out URL
ftp://ftp.vlsivie/tuwien.ac.at/pub/8bit/latex2e.ps (in German).

Alternatively, the styles used for earlier LaTeX versions (see above)
can also be used with 2e.  To do this, use the commands:
\documentclass{article} 
\usepackage{isolatin}


You can also get the latex-mode to handle opening and closing quotes
correctly for your language.  This can be achieved by defining the
emacs variables 'tex-open-quote' and 'tex-closing-quote'.  You can
either set these varaibles in your ~/.emacs startup file or as a
buffer-local variable in your TeX file if you want to define quotes on
a per-file basis.

For German TeX quotes, use:
-----------
(setq tex-open-quote "\"`")
(setq tex-closing-quote "'\"")
-----------

If you want to use French quotes (guillemets), use:
-----------
(setq tex-open-quote "")
(setq tex-closing-quote "")
-----------

Bibtex has some problems with 8 bit characters, esp. when they are
used as keys.  BibTeX 1.0, when it eventually comes out (most likely
some time in 1996), will support 8-bit characters.



8. ISO 8859-1 and emacs
Emacs 19 (as opposed to Emacs 18) can automatically handle 8 bit
characters. (If you have a choice, upgrade to Emacs version 19.23,
which has the most complete ISO support.)  Emacs 19 has extensive
support for ISO 8859-1. If your display supports ISO 8859-1 encoded
characters, add the following line to your .emacs startup file:
-----------------------------
(standard-display-european t)
-----------------------------

If want to display ISO-8859-1 encoded files by using TeX-like escape
sequences (e.g. if your terminal supports only ASCII characters), you
should  add the following line to your .emacs file (DON'T DO THIS IF
YOUR TERMINAL SUPPORTS ISO OR SOME OTHER ENCODING OF NATIONAL
CHARACTERS): 
--------------------
(require 'iso-ascii)
--------------------

If your terminal supports a non-ISO 8859-1 encoding of national
characters (e.g. 7 bit national variant ISO 646 character sets,
aka. `national ASCII' variants), you should configure your own display
table.  The standard emacs distribution contains a configuration
(iso-swed.el) for terminals which have ASCII in the G0 set and a
Swedish/Finnish version of ISO 646 in the G1 set.  If you want to
create your own display table configuration, take a look at this
sample configuration and at disp-table.el for available support
functions.


Emacs can also accept 8 bit ISO 8859-1 characters as input. These
character codes might either come from a national keyboard (and
driver) which generates ISO-compliant codes, or may have been entered
by use  of a COMPOSE-character mechanism. 
If you use such an input format, execute the following expression in
your .emacs startup file to enable Emacs to understand them:
-------------------------------------------------
     (set-input-mode (car (current-input-mode))
                     (nth 1 (current-input-mode))
                     0)
-------------------------------------------------

In order to configure emacs to handle commands operating on words
properly (such as 'Beginning of word', etc.), you should also add the
following line to your .emacs startup file:
-------------------------------
(require 'iso-syntax)
-------------------------------

This lisp script will change character attributes such that ISO 8859-1
characters are recognized as such by emacs.

The GNU Emacs package iso-cvt+ supports reading and writing character
sets in various character set encodings (such as ISO Latin1, HTML,
TeX, IBM PC,...).  To do so, iso-cvt+ install two 'File' menu items,
'Load As' and 'Write As' which allow you to select the desired
character set encoding.  iso-cvt+ is available at
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/


For further information on using ISO 8859-1 with emacs, also see the
Emacs manual section on "European Display" (available as hypertext
document by typing C-h i in emacs or as a printed version).


If you need to edit text in a non-European language (Arabic, Chinese,
Cyrillic-based languages, Ethiopic, Korean, Thai, Vietnamese, etc.),
MULE (URL ftp://etlport.etl.go.jp/pub/mule) is a Multilingual
Enhancement to GNU Emacs which supports these languages.



9. Typing ISO with US-style keyboards.  
Many computer users use US-ASCII keyboards, which do not have keys for
national characters.  You can use escape sequences to enter these
characters.  For ASCII terminals (or PCs), check the documentation of
your terminal for particulars.


9.1 US-keyboards under X11 
Under X Windows, the COMPOSE multi-language support key can be used to
enter accented characters.  Thus, when running X11 on a SunOS-based
computer (or any other X11R4 or X11R5 server supporting COMPOSE
characters), you can type three character sequences such as
COMPOSE " a -> 
COMPOSE s s -> 
COMPOSE ` e -> 
to type accented characters.

Note that this COMPOSE capability has been removed as of X11R6,
because it does not adequately support all the languages in the world.
Instead, compose processing is supposed to be performed in the client
using an `input method', a mechanism which has been available since
X11R5.  (In the short term, this is a step backward for European
users, as few clients support this type of processing at the moment.
It is unfortunate that the X Consortium did not implement a mechanism
which allows for a smoother transition.  Even the xterm terminal
emulator supplied by the X Consortium itself does not yet support this
mechanism!)

Input methods are controlled by the locale environment variables (LANG
and LC_xxx).  The values for these variables are (or at least, should be
made equivalent by any sane vendor) equivalent to those expected by
the ANSI/POSIX locale library.  For a list of possible settings see
section 3. 



9.2 US-keyboards and emacs
9.2.1 Using ALT for composing national characters 
There are several modes to enter Umlaut characters under emacs when
using a US-style keyboard.  One such mode is iso-transl, which is
distributed with the standard emacs distribution.  This mode uses the
Alt-key for entering diacritical marks (accents et al.).  

To activate iso-transl mode, add the following line to your .emacs
setup file:
(require 'iso-transl)

As of emacs 19.29, Alt-sequences optimized for a particular language
are available.  Use the following call in .emacs to select your
favorite keybindings:
(iso-transl-set-language "German")

If you do not have an Alt-key on your keyboard, you can use the C-x 8
prefix to access the same capabilities.

For pre-19.29 versions, similar functionality is availble as extended
iso-transl mode (iso-transl+) which allows the definition of language
specific short cuts is available as URL
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/.  This file
also includes sample configurations for the German and Spanish
languages.


9.2.2 Electric Accents
An alternative to using Alt-sequences for entering diacritical marks
is the use of `electric accents', such as used on old type writers or
under many MS Windows programs.  With this method, typing an accent
character will place this accent on the next character entered.  One
mode which supports this entry method is the iso-acc minor mode which
comes with the standard emacs distribution. Just add 
------------------
(require 'iso-acc)
------------------
to your emacs startup script, and you can turn the '`~/^" keys into
electric accents by typing 'M-x iso-accents-mode' in a specific
buffer.  To type the  (c with cedilla) and  (German scharfes s)
characters, type ~c and "s, respectively.

Footnote: When starting up under X11, Emacs looks for a Meta key and
if it finds no Meta key, it will use the Alt key instead.  The way to
solve this problem, is to define a Meta key using the xmodmap utility
which comes with X11.



10. File names with ISO characters
If your OS is 8 bit clean, you can use ISO characters in file names.
(This is possible under SunOS.)



11. Command names with ISO 8859-1 
If your OS supports file names with ISO characters, and your shell is
8 bit clean, you can use command names containing ISO characters. If
your shell does not handle ISO characters correctly, use one of the
many PD shells which do (e.g. tcsh, an extended csh). These are
available from a multitude of ftp sites around the world.

See section 14 on application specific information for a discussion of
various shells.



12. Spell checking
Ispell 3.1 has by far the best understanding of non-English
languages and can be configured to handle 8-bit characters
(Thus, it can handle ISO-8859-1 encoded files).

Ispell 3.1 now comes with hash tables for several languages (English,
German, French,...).  It is available via URL ftp://ftp.cs.ucla.edu/pub.  
Ispell also contains a list of international dictionaries and about
their availability in the file ispell/languages/Where.

To choose a dictionary for ispell, use the `-d <dictionary>'
option.  The `-T <input-encoding>' option should be set set to `-T
latin1' if you want to use ISO 8859-1 as input encoding.

If you use ispell inside emacs (using the ispell.el mode) to spell
check a buffer, you can choose language and input encoding either
using the `M-x ispell-change-dictionary' function, or by choosing the
`Spell' item in the `Edit' pull-down menu.  This will present you with
a choice of dictionaries (cum input encodings): all languages are
listed twice, such as in `Deutsch' and `Deutsch8'.  `Deutsch8' is the
setting which will use the German dictionary and the 8 bit ISO 8859-1
input encoding.  

Alternatively, ispell.el lets you specify the dictionary to use for a
particular file at the end of of that file by adding a line such as
----
Local IspellDict: castellano8
----

The following sites also have dictionaries for ispell available via
anonymous ftp:
language	site			 file name
French		ireq-robot.hydro.qc.ca	 /pub/ispell
French		ftp.inria.fr		 /INRIA/Projects/algo/INDEX/iepelle
French		ftp.inria.fr		 /gnu/ispell3.0-french.tar.gz
German		ftp.vlsivie.tuwien.ac.at /pub/8bit/dicts/deutsch.tar.gz
Spanish		ftp.eunet.es		 /pub/unix/text/TeX/spanish/ispell
Portuguese      http://www.di.uminho.pt/~jj/pln/pln.html

Some spell checkers use strange encodings for accented characters.  If
you have to use one of these spell checkers, you may have to run
recode before invoking the spell checker to generate a file using your
spell checker's coding conventions.  After running the spell checker,
you have to translate the file back to ISO with recode.

Of course, this can be automated with a shell script:
---------------------
recode <options to generate spell checker encoding from ISO> $i tmp.file 
spell_check tmp.file
recode <options to generate ISO from spell checker encoding> tmp.file $i 
---------------------

Footnote: Ispell 4.* is not a superset of ispell 3.*. Ispell 4.* was
	  developed independently from a common ancestor, but DOES NOT
	  support any internationalization, but is restricted to the
	  English language.



13. TCP and ISO 8859-1
TCP was specified by US-Americans, for US-Americans. TCP still carries
this heritage: while TCP/IP protocol itself *is* 8 bit clean, no
effort was made to support the transfer of non-English characters in
many application level protocols (mail, news, etc.).  Some of these
protocols still only specify the transfer of 7-bit data, leaving
anything else implementation dependent.

Since the TCP/IP protocol itself transfers 8 bit data correctly,
writing applications based on TCP/IP does not lead to any loss of
encoding information.


13.1 FTP and ISO 8859-1
Transmitting data via FTP is an interesting issue, depending on what
system you use, how the relevant RFCs are interpreted, and what is
actually implemented. 

If you transfer data between two hosts using the same ISO 8859-1
representation (such as two Unix hosts), the safest solution is to
specify 'binary' transmission mode.

Note, however, that use of the binary mode for text files will disable
translation between the line-ending conventions of different operating
systems.  You might have to provide some filter to convert between the
LF-only convention of Unix and the CR-LF convention of VMS and MS
Windows when you copy from one of these systems to another.

If the FTP server and client computers use different encoding, there
are two possible approaches:
* Transfer all data as binary data, then convert the format using a
  conversion tools such as recode to translate the tranferred data.
* Specify an ASCII connection, and have your FTP server and client
  convert the encoding automatically.

While the first approach always works, it is somewhat cumbersome if
you transmit a lot of data.  The second transfer solution is much more
comfortavle, but it depends on you client (and server) to take care of
the appropriate character translations.  Since there is no universal
standard for network characters beyond ASCII (NVT-ASCII as specified
in RFC 854), this depends on attitude of your software vendor. 

Most Apple Macintosh network software is configured to treat all
network data as having ISO 8859-1 encoding and automatically
translates from and to the internal MacOS data representation.  (This
can be problematic, if you want to send or receive text using the
Macintosh character set.  The correct solution to this problem is
to use MIME.)

MS-DOS programs are much less well-behaved, and you have to try
whether your particular FTP program performs conversion.

An additional issue with the automatic translation is how to translate
unavailable characters.  If FTP is used to store and retrieve data,
the original file should be re-constructable after conversion.  If
data is to printed or processed, different encodings (e.g. graphic
approximation of characters) may be necessary. (See the section on
character set translation for a full discussion of encoding
transformations.)

A second, optional parameter is possible for 'type ascii' commands,
which specifies whether the data is for non-printing or printing
purposes.  Ideally, FTP servers for non-8859-1 servers would use this
parameter to determine whether to use an invertible encoding or
graphical and/or logical approximation during translation.  (Although
RFC 959, section 3.1.1.5 does not require this.)


13.2 Mail and ISO 8859-1
Most Internet eMail standards come from a time when the Internet was a
mostly-US phenomenon.  Other countries did have access to the net, but
much of the communication was in English nevertheless.  With the
propagation of Internet, these standards have become a problem for
languages which cannot be represented in a 7 bit ISO 646 character
set.  

Using ISO 646, which uses a slightly different character set for each
language, also poses a problem when crossing a language barrier, as
the interpretation of characters will change.  As a result, most
countries use the ISO 646 standard commonly referred to as US-ASCII
and will use escape sequences such as 'e () or "a () to refer to
national characters.  The exception to this rule are Nordic countries
(more so in Sweden and Finland, less so in Denmark and Norway, I'm
being told), where the national ISO 646 variant has garnered a
formidable following and is a common reference point for all Nordic
users.

There are several languages, for which there are not enough
replacement characters to code all national variants (e.g. French).

Footnote:
Hence, French has not followed the nordic track.  French
net-convention is e' instead of 'e ("l''el'ephant" is strange
spelling) and many think that this is very ugly writing anyway and
drop the accents altogether but this makes text sometimes funny and
incorrect at least.


As this situation is clearly unsatisfactory, several methods of
sending mails encoded in national character sets have been developed.
We start with a discussion of the mail delivery infrastructure and
will then look at some high-level protocols which can protect mail
users and their messages from the shortcomings of the underlying mail
protocols.

Footnote: Many other email standards exist for proprietary systems.
If you use one of these mail systems, it is the responsibility of the
mail gateway to translate your messages to an appropriate Internet
mail message when you send a message to the Internet.


13.2.1 Mail Transfer Agents and the Internet Mail Infrastructure
The original sendmail protocol specification (SMTP) in RFC 821
specified the transfer of only 7 bit messages.  Many sendmail
implementations have been made 8 bit transparent (see RFC 1428), but
some SMTP handling agents are still strictly conforming to the
(somewhat outdated) RFC 821 and intentionally cut off the 8th bit.
This behavior stymies all efforts to transfer messages containing
national characters.  Thus, only if all SMTP agents between mail
originator and mail recipient are 8 bit clean, will messages be
transferred correctly.  Otherwise, accented characters are mapped to
some ASCII character (e.g. Umlaut a -> 'd'), but the rest of the
messages is still transferred correctly.

A new, enhanced (and compatible) SMTP standard, ESMTP, has been
released as RFC 1425. This standard defines and standardizes 8 bit
extensions.  This should be the mail protocol of choice for newly
shipped versions of sendmail.

Much of the European and Latin American network infrastructure
supports the transfer of 8 bit mail messages, the success rate is
somewhat lower for the US.

DEC Ultrix sendmail still implements the somewhat outdated RFC 821 to
the letter, and thus cuts off the eighth bit of all mail passing
through it.  Thus ISO encoded mail will always lose the accent marks
when transferred through a DEC host.

If your computer is running DEC Ultrix and you want it to handle 8 bit
characters properly, you can get the source for a more recent version
of sendmail via ftp (see section 14.9).  OR, you can simply
call DEC, complain that their standard mail system cannot handle
international 8 bit mail, encourage them to implement 8 bit
transparent SMTP, or (even better) ESMTP, and ask for the sendmail
patch which makes their current sendmail 8 bit transparent.
(Reportedly, such a patch is available from DEC for those who ask.)
In the meantime, an 8 bit transparent sendmail MIPS binary for Ultrix
is available as URL
ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/) 

If you want to change MTAs, the popular smail PD-MTA is also 8 bit
clean.


13.2.2 High-level protocols
In the Good Old Days, messages were 7-bit US-ASCII only. When users
wanted to transfer 8 bit data (binaries or compressed files, for
example), it was their responsibility to translate them to a 7 bit
form which could be sent.  At the other end, the recipient had to
unpack the data using the same protocol.  The commonly used encoding
mechanism used for this purpose is uuencode/uudecode.

Today, a standard, MIME (MIME stands for Multi-purpose Internet Mail
Extensions), exists which automatically packs and unpacks data as is
required.  This standard can take advantage of different underlying
protocol capabilities and automatically transform messages to
guarantee delivery.  This standard can also be used to include
multimedia data types in your mail messages. 

The MIME standard defines a mail transfer protocol which can handle
different character sets and multimedia mail, independent of the
network infrastructure.  This protocol should eventually solve
problems with 7-bit mailers etc.  Unfortunately, no mail transfer
agents (mail routers) and few end user mail readers support this
standard.  Source for supporting MIME (the `metamail' package) in
various mail readers is available in URL
ftp://thumper.bellcore.com/pub/nsb.  MIME is specified in RFC 1521 and
RFC 1522 which are available from ftp.uu.net.  There is also a MIME
FAQ which is available as URL
ftp://ftp.ics.uci.edu/mh/contrib/multimedia/mime-faq.txt.gz.  (This
file is in compressed format. You will need the GNU gunzip program to
decompress this file.)

PS: Newer versions of sendmail support ESMTP negotiation and can pass
8 bit data.  However, they do not (yet?)  support downgrading of 8 bit
MIME messages.


13.3 News and ISO 8859-1
Much as mail, the Usenet news protocol specification is 7 bit based,
but the infrastructure has been upgraded to 8 bit service...  Thus,
accented characters are transferred correctly between much of Europe
(and Latin America).

ISO 8859-1 is _the_ standard for typing accented characters in most
newsgroups (may be different for MS-DOS centered newsgroups ;-), and
is preferred in most European news group hierarchies, such as at.* or
de.* 

For those who speak French, there is an excellent FAQ on using ISO
8859-1 coded characters on Usenet by Franois Yergeau (URL
ftp://ftp.ulaval.ca/contrib/yergeau/).  This FAQ is
regularly posted in soc.culture.french and other relevant newsgroups.


13.4 WWW (and other information servers)
The WWW protocol can transfer 8 bit data without any problems and you
can advertise ISO-8859-1 encoded data from your client.  The display
of data is dependent upon the user client.  xmosaic (freely available
from the NCSA) which is available for most UNIX platforms uses an
ISO-8859-1 compliant font by default and will display data correctly.


13.5 rlogin
For rlogin to pass 8 bit data correctly, invoke it with 'rlogin -8' or
'rlogin -L'.



14. Some applications and ISO 8859-1
14.1 bash
You need version 1.13 or higher and set the locale correctly (see
section 3).  Also, to configure the `readline' input function of bash
to handle 8 bit characters correctly, you have to set some environment
variables in the readline startup file .inputrc:
-------------------------------------------------------
set meta-flag On
set convert-meta Off
set output-meta On
-------------------------------------------------------

Before bash version 1.13, bash used the eighth bit of characters to
mark whether or not they were quoted when performing word expansions.
While this was not a problem in a 7-bit US-ASCII environment, this was
a major restriction for users working in a non-English environment.

These readline variables have the following meaning (and default
values): 
     meta-flag (Off)
          If set to On,  readline  will  enable  eight-bit  input
          (that is, it will not strip the high bit from the char-
          acters it  reads),  regardless  of  what  the  terminal
          claims it can support.
     convert-meta (On)
          If set to On, readline will convert characters with the
          eighth  bit  set  to an ASCII key sequence by stripping
          the eighth bit and prepending an escape  character  (in
          effect, using escape as the meta prefix).
     output-meta (Off)
          If set to On, readline will display characters with the
          eighth  bit set directly rather than as a meta-prefixed
          escape sequence.

Bash is available from prep.ai.mit.edu in /pub/gnu.


14.2 elm
Elm automatically supports the handling of national character sets,
provided the environment is configured correctly.  If you configure
elm without MIME support, you can receive, display, enter and send 8
bit ISO 8859-1 messages (if your environment supports this character
set).

When you compile elm with MIME support, you have two options:
* you can compile elm to use 8 bit ISO-8859-1 as transport encoding:
  If you use this encoding even people without MIME compliant mailers
  will be able to read your mail messages, if they use the same
  character set.  The eight bit may however be cut off by 7 bit MTAs
  (mail transfer agents), and mutilated mail might be received by the 
  recipient, regardless of whether she uses MIME or not.  (This
  problem should be eased when 8 bit mailers are upgraded to
  understand how to translate 8 bit mails to 7 bit encodings when they
  encounter a 7 bit mailer.) 

* you can compile elm to use 7 bit US-ASCII `quoted printable' as
  transport encoding: 
  this encoding ensures that you can transfer your mail containing 
  national characters without having to worry about 7 bit MTAs.  A
  MIME compliant mail reader at the other end will translate your
  message back to your national character set.  Recipients without
  MIME compliant mail readers will however see mutilated messages:
  national characters will have been replaced by sequences of the type
  '=FF' (with FF being the ISO code (in hexadecimal) of the national
  character being encoded). 


14.3 GNUS
GNUS is a newsreader based on emacs.  It is 8 bit transparent and
contains all national character support available in emacs 19.


14.4 less
Version 237 and later automatically displays latin1 characters, if
your locale is configured correctly.  

If your OS does not support the locale mechanism, or if you use a
version of less older than 237, set the LESSCHARSET environment
variable with 'setenv LESSCHARSET latin1'.
 

14.5 metamail
To configure the metamail package for ISO 8859-1 input/output, set the
MM_CHARSET environment variable with 'setenv MM_CHARSET ISO-8859-1'.
Also, set the MM_AUXCHARSETS variable with 'setenv MM_AUXCHARSETS
iso-8859-1'. 


14.6 nn
Add the line
-----------------
set data-bits 8
-----------------
to your ~/.nn/init (or the global configuration file) in order for nn
to be able to process 8 bit characters. 


14.7 nroff
The GNU replacement for nroff, groff, has an option to generate ISO
8859-1 coded output, instead of plain ASCII.  Thus, you can preview
nroff documents with correctly displayed accented characters.  Invoke
groff with the 'groff -Tlatin1' option to achieve this. 

Groff is free software. It is available from URL
ftp://prep.ai.mit.edu/pub/ and many other GNU archives around the
world.


14.8 pgp
PGP (Phil Zimmermann's Pretty Good Privacy) uses Latin1 as canonical
form to transmit crypted data.  Your host computer's local character
set should be configured in the configuration file
${PGPPATH}/config.txt by setting the CHARSET parameter.  If you are
using ISO 8859-1 as your native character set, CHARSET should bet set
to LATIN1, on MS-DOS computers with code page 850 set 'CHARSET =
CP850'.  This will make PGP automatically translate all crypted texts
from/to the LATIN1 canonical form.  A setting of 'CHARSET = NOCONV'
can be used to inhibit all translations.  (

When PGP is used to code Cyrillic text, KOI8 is regarded as canonical
form (use 'CHARSET = KOI8').  If you use the ALT_CODES encoding for
Cyrillic (popular on PCs), set 'CHARSET = ALT_CODES' and it will
automatically be converted to KOI8.

Footnote: Note that PGP treats KOI8 as LATIN1, even though it is a
completely different character set (Russian), because trying to
convert KOI8 to either LATIN1 or CP850 would be futile anyway.


14.* samba
To make samba work with ISO 8859-1, use the following line in the
[global] section:
valid chars     = 0xa0 0xa1 0xa2 0xa3 0xa4 0xa5 0xa6 0xa7 0xa8 0xa9 0xaa 0xab 0xac 0xad 0xae 0xaf 0xb0 0xb1 0xb2 0xb3 0xb4 0xb5 0xb6 0xb7 0xb8 0xb9 0xba 0xbb 0xbc 0xbd 0xbe 0xbf 0xc0:0xe0 0xc1:0xe1 0xc2:0xe2 0xc3:0xe3 0xc4:0xe4 0xc5:0xe5 0xc6:0xe6 0xc7:0xe7 0xc8:0xe8 0xc9:0xe9 0xca:0xea 0xcb:0xeb 0xcc:0xec 0xcd:0xed 0xce:0xee 0xcf:0xef 0xd0:0xf0 0xd1:0xf1 0xd2:0xf2 0xd3:0xf3 0xd4:0xf4 0xd5:0xf5 0xd6:0xf6 0xd7 0xf7 0xd8:0xf8 0xd9:0xf9 0xda:0xfa 0xdb:0xfb 0xdc:0xfc 0xdd:0xfd 0xde:0xfe 0xdf 0xff


14.9 sendmail 
BSD Sendmail Version 8 has a flag in the configuration file set to
True or False which determines whether v8 passes any 8-bit data it
encounters, presumably to match the behavior of other 8-bit
transparent MTAs and to meet the wants of non-ASCII users, or if it
strips to 7 bits to conform to SMTP.  The source code for an 8 bit
clean sendmail is available in URL ftp://ftp.cs.berkeley.edu/ucb/sendmail.  
A pre-compiled binary for DEC MIPS systems running Ultrix is available
as URL ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/.


14.10 tcsh
You need version 6.04 or higher, and your locale has to be set
properly (see section 3).  Tcsh also needs to be compiled with the
national language support feature, see the config.h file in the tcsh
source directory.  Tcsh is an extended csh and is available in URL
ftp://ftp.deshaw.com/pub/tcsh

If tcsh has been configured correctly, it will allow national
characters in ENVIRONMENT variables, shell variables, file names, etc.

set BentigteDateien=/etc/rc
cat $BentigteDateien > /dev/null


14.11 vi
Support for 8 bit character sets depends on the OS.  It works under
SunOS 4.1.*, but on OSF/1 vi gets confused about the current cursor
position in the presence of 8 bit characters.  Some versions of vi
require an 8bit locale to work with 8 bit characters.


All major replacements for vi seem to support 8 bit characters:

14.11.1 vile  ('VI Like Emacs')
Vile (by Paul Fox) can be told that the usual range of 8th-bit
characters are printable with "set printing-low 160" and "set
printing-high 255".  By either executing these command in vile or by
placing them in ~/.exrc, vile will not use the usual octal or hex
expansion for these characters.  vile is available from
ftp://id.wing.net/pub/pgf/vile.


************************* REQUIRES A RE-WRITE ********************************
Normally, 8 bit chars are printed either in hex (the default) or octal
("set unprintable-as-octal").  they look like "\xC7" or "\307" on your
screen.

vile was the first vi rewrite to provide multi-window/multi-buffer
operation.  and since it was derived from micro-emacs, it retains
fully rebindable keys, and a built in macro language.  the ftp site is
id.wing.net:/pub/pgf/vile.  the current version is 5.2.  it's pretty
mature (5 years old).  there's an X-aware version as well, that makes
full use of the mouse, with scrollbars, etc.  (to answer your
question, initialization stuff goes in a .vilerc file.)

Do you require use of the correct locale settings?
no.  8-bit support is fairly primitive.  i'll include the
pertinent sections of the doc down below.


hope all this helps --

paul

------------------------------------
from vile's Help file:

8-Bit Operation
---------------

	vile allows input, manipulation, and display of all 256 possible
	byte-wide characters.  (Double-wide characters are not supported.)
	
	Output 
	------
	By default, characters with the high bit set (decimal value 128 or
	greater) will display as hex (or octal; see "non-printing- octal"
	above) sequences, e.g.  \xA5.  A range of characters which should
	display as themselves (that is, characters understood by the user's
	display terminal) may be given using the "printing-low" and
	"printing-high" settings (see above).  Useful values for these
	settings are 160 and 255, which correspond to the printable range
	of the ISO-Latin-1 character set.

	Input
	-----
	If the user's input device can generate all characters, and if the
	terminal settings are such that these characters pass through
	unmolested (Using "stty cs8 -parenb -istrip" works for me, on an
	xterm.  Real serial lines may take more convincing, at both ends.),
	then vile will happily incorporate them into the user's text, or
	act on them if they are bound to functions.  Users who have no need
	to enter 8-bit text may want access to the meta-bound functions
	while in insert mode as well as command mode.  The mode
	"meta-insert-bindings" controls whether functions bound to meta-
	keys (characters with the high bit set) are executed only in
	command mode, or in both command and insert modes.  In either case,
	if a character is _not_ bound to a function, then it will be
	self-inserting when in insert mode.  (To bind to a meta key in the
	.vilerc file, one may specify it as itself, or in hex or octal, or
	with the shorthand 'M-c' where c is the corresponding character
	without the high bit set.

------------------------------------
also from vile's Help file, these are the settable modes which affect
8-bit operation:

	meta-insert-bindings (mib)  Controls behavior of 8-bit characters
		during insert.  Normally, key-bindings are only operational
		when in command mode: when in insert mode, all characters
		are self-inserting.  If this mode is on, and a meta-character
		is typed which is bound to a function, then that function
		binding will be honored and executed from within insert
		mode.  Any unbound meta-characters will remain self-inserting.
		(B)

	printing-low The integer value representing the first of the
		printable set of "high bit" (i.e. 8-bit) characters. 
		Defaults to 0.  Most foreign (relative to me!) users would
		set this to 160, the first printable character in the upper
		range of the ISO 8859/1 character set.  (U)

	printing-high The integer value representing the last character of
		the printable set of "high bit" (i.e. 8-bit) characters. 
		Defaults to 0.  Set this to 255 for ISO 8859/1
		compatibility.  (U)

	unprintable-as-octal (uo) If an 8-bit character is non-printing, it
		will normally be displayed in hex.  This setting will force
		octal display.  Non-printing characters whose 8th bit is
		not set are always displayed in control character (e.g. '^C')
		notation.  (B)
************************* REQUIRES A RE-WRITE ********************************

14.11.2 vim
vim was developed on an Amiga in Europe, and supports a mechanism
similar to vile.  'vim' supports input digraphs for entering 8-bit
chars, the output convention is similar to vile -- raw or nothing.

Details are unkonwn.  (If you know more about vim,
please let me know.  A request to comp.editors should yield additional
information.)

14.11.3 nvi
A recent vi-rewrite which should also should support 8 bit characters.
(Keith Bostic (bostic@cs.berkeley.edu) is the author and should know
more about nvi.)



15. Terminals
15.1 X11 Terminal Emulators
See section 4 on X11 for bug fixes for X11 clients. 

15.1.1 xterm
If you are using X11 and xterm as your terminal emulator, you should 
place the following line in ~/.Xdefaults (this seems to be required in
some releases of X11, not in all):
-------------------------
XTerm*EightBitInput: True
-------------------------

15.1.2 rxvt
rxvt is another terminal emulator used for X11, mostly under
Linux.  Invoke rxvt with the 'rxvt -8' command line.


15.2 VT2xx, VT3xx
The character encoding used in VT2xx terminals is a preliminary
version of the ISO-8859-1 standard (DEC MCS), so some characters (the
more obscure ones) differ slightly.  However, these terminals can be
used with ISO 8859-1 characters without problems.

The newer VT3xx terminals use the official ISO 8859-1 standard. 

The international versions of the VT[23]xx terminals have a COMPOSE
key which can be used to enter accented characters, e.g. 
<COMPOSE><e><'> will give an e with accent aigu ().


15.3 Various UNIX terminals
Some terminals support down-loadable fonts.  If characters sent to
these terminals can be 8 bit wide, you can down-load your own ISO
characters set.  To see how this can be achieved, take a look at the
/pub/culture/russian/comp/cyril-term on nic.funet.fi.


15.4 MS-DOS PCs
MS-DOS PCs normally use a different encoding for accented characters,
so there are two options:

* you can use a terminal emulator which will translate between the
  different encodings.  If you use the PROCOMM PLUS, TELEMATE and
  TELIX modem programs, you can down-load the translation tables 
  from URL ftp://oak.oakland.edu/pub/msdos/modem/xlate.zip.  (You need
  to install CP850 for this to work.)

* you can reconfigure your MS-DOS PC to use an ISO-8859-1 code page.
  Either install IBM code page 819 (see section 19), or you can get
  the free ISO 8859-X support files from the anonymous ftp archive
  ftp://ftp.uni-erlangen.de/pub/doc/ISO/charsets, which contains data
  on how to do this (and other ISO-related stuff).  The README file
  contains an index of the files you need. 

Note that many terminal emulations for PCs strip the 8th bit when in
text transmission mode.  If you are using such a program to dial up
a computer, you may have to configure your terminal program to
transmit all 8 bits.


16. Programming applications which support the use of ISO 8859-1
For information on how to write applications with support for
localization (to the ISO 8859-1 and other character representations)
check out URL ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/.



17. Other relevant i18n FAQs
This is a list of other FAQs on the net which might be of interest. 
Topic			Newsgroup(s)		Comments
Nordic graphemes	soc.culture.nordic	interesting stuff about 
						handling nordic letters
accents sur Usenet	soc.culture.french,...	Accents on Usenet (French)
						+ more
Programming for I18N	comp.unix.questions,...	see section 16.
International fonts	ftp://ftp.vlsivie.tuwien.ac.at/pub/8bit/
						Discusses international fonts
						and where to find them
I18N on WWW		http://www.vlsivie.tuwien.ac.at/mike/i18n.html
German-HowTo for Linux	ftp://ftp.univie.ac.at/systems/linux/sunsite/docs/HOWTO/German-HOWTO

Using 8 bit characters  ftp://ftp.ulg.ac.be/pub/docs/iso8859/*     (1)

Much charactersets info ftp://kermit.columbia.edu/kermit/charsets/
                        http://www.columbia.edu/kermit/            (2)

(1) written to "convey" the problem to the ASCII programmer, hence
more theoretical background.
(2) Kermit is second to none (in time and quality) for character sets
support and deserves a pointer in this FAQ.



18. Operating Systems and ISO 8859-1
18.1 UNIX
Most Unix implementations use the ISO 8859--1 character set, or at
least have an option to use it.  Some systems may also support other
encodings, e.g.~Roman8 (HP/UX), DEC MCS (DEC Ultrix, see the section
on VMS), etc.


18.2 NeXTSTEP
NeXTSTEP uses a proprietary character set.


18.3 MS DOS
IBM code page 819 _is_ ISO 8859-1.  Code Page 850 has the same
characters as ISO 8859-1, BUT the characters are in different
locations (i.e., you can translate 1-to-1, but you do have to
translate the characters.)


18.4 MS-Windows
Microsoft Windows uses an ISO 8859-1 compatible character set (Code
Page 1252), as delivered in the US, Europe (except Eastern Europe) and
Latin America.  In Windows 3.1, Microsoft has added additional characters
in the 0x80-0x9F range.


18.5 DEC VMS
DEC VMS uses the DEC MCS character set, which is practically
equivalent to ISO 8859-1 (it is a fromer ISO 8859--1 draft standard).
The only characters which differ between DEC MCS and ISO 8859-1 are
the Icelandic characters (eth and thorn) at locations 0xD0, 0xF0, 0xDE
and 0xFE.




19. Table of ISO 8859-1 Characters
This section gives an overview of the ISO 8859-1 character set.  The
ISO 8859-1 character set consists of the following four blocks:

00	19	CONTROL CHARACTERS
20	7E	BASIC LATIN
80	9F	EXTENDED CONTROL CHARACTERS
A0	FF	LATIN-1 SUPPLEMENT

The control characters and basic latin blocks are similar do those
used in the US national variant of ISO 646 (US-ASCII), so they are not
listed here.  Nor is the second block of control characters listed,
for which not functions have yet been defined.  

+----+-----+---+------------------------------------------------------
|Hex | Dec |Car| Description ISO/IEC 10646-1:1993(E)
+----+-----+---+------------------------------------------------------
|    |     |   |
| A0 | 160 |   | NO-BREAK SPACE
| A1 | 161 |  | INVERTED EXCLAMATION MARK
| A2 | 162 |  | CENT SIGN
| A3 | 163 |  | POUND SIGN
| A4 | 164 |  | CURRENCY SIGN
| A5 | 165 |  | YEN SIGN
| A6 | 166 |  | BROKEN BAR
| A7 | 167 |  | SECTION SIGN
| A8 | 168 |  | DIAERESIS
| A9 | 169 |  | COPYRIGHT SIGN
| AA | 170 |  | FEMININE ORDINAL INDICATOR
| AB | 171 |  | LEFT-POINTING DOUBLE ANGLE QUOTATION MARK
| AC | 172 |  | NOT SIGN
| AD | 173 |  | SOFT HYPHEN
| AE | 174 |  | REGISTERED SIGN
| AF | 175 |  | MACRON
|    |     |   |
| B0 | 176 |  | DEGREE SIGN
| B1 | 177 |  | PLUS-MINUS SIGN
| B2 | 178 |  | SUPERSCRIPT TWO
| B3 | 179 |  | SUPERSCRIPT THREE
| B4 | 180 |  | ACUTE ACCENT
| B5 | 181 |  | MICRO SIGN
| B6 | 182 |  | PILCROW SIGN
| B7 | 183 |  | MIDDLE DOT
| B8 | 184 |  | CEDILLA
| B9 | 185 |  | SUPERSCRIPT ONE
| BA | 186 |  | MASCULINE ORDINAL INDICATOR
| BB | 187 |  | RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK
| BC | 188 |  | VULGAR FRACTION ONE QUARTER
| BD | 189 |  | VULGAR FRACTION ONE HALF
| BE | 190 |  | VULGAR FRACTION THREE QUARTERS
| BF | 191 |  | INVERTED QUESTION MARK
|    |     |   |
| C0 | 192 |  | LATIN CAPITAL LETTER A WITH GRAVE ACCENT
| C1 | 193 |  | LATIN CAPITAL LETTER A WITH ACUTE ACCENT
| C2 | 194 |  | LATIN CAPITAL LETTER A WITH CIRCUMFLEX ACCENT
| C3 | 195 |  | LATIN CAPITAL LETTER A WITH TILDE
| C4 | 196 |  | LATIN CAPITAL LETTER A WITH DIAERESIS
| C5 | 197 |  | LATIN CAPITAL LETTER A WITH RING ABOVE
| C6 | 198 |  | LATIN CAPITAL LIGATURE AE
| C7 | 199 |  | LATIN CAPITAL LETTER C WITH CEDILLA
| C8 | 200 |  | LATIN CAPITAL LETTER E WITH GRAVE ACCENT
| C9 | 201 |  | LATIN CAPITAL LETTER E WITH ACUTE ACCENT
| CA | 202 |  | LATIN CAPITAL LETTER E WITH CIRCUMFLEX ACCENT
| CB | 203 |  | LATIN CAPITAL LETTER E WITH DIAERESIS
| CC | 204 |  | LATIN CAPITAL LETTER I WITH GRAVE ACCENT
| CD | 205 |  | LATIN CAPITAL LETTER I WITH ACUTE ACCENT
| CE | 206 |  | LATIN CAPITAL LETTER I WITH CIRCUMFLEX ACCENT
| CF | 207 |  | LATIN CAPITAL LETTER I WITH DIAERESIS
|    |     |   |
| D0 | 208 |  | LATIN CAPITAL LETTER ETH
| D1 | 209 |  | LATIN CAPITAL LETTER N WITH TILDE
| D2 | 210 |  | LATIN CAPITAL LETTER O WITH GRAVE ACCENT
| D3 | 211 |  | LATIN CAPITAL LETTER O WITH ACUTE ACCENT
| D4 | 212 |  | LATIN CAPITAL LETTER O WITH CIRCUMFLEX ACCENT
| D5 | 213 |  | LATIN CAPITAL LETTER O WITH TILDE
| D6 | 214 |  | LATIN CAPITAL LETTER O WITH DIAERESIS
| D7 | 215 |  | MULTIPLICATION SIGN
| D8 | 216 |  | LATIN CAPITAL LETTER O WITH STROKE
| D9 | 217 |  | LATIN CAPITAL LETTER U WITH GRAVE ACCENT
| DA | 218 |  | LATIN CAPITAL LETTER U WITH ACUTE ACCENT
| DB | 219 |  | LATIN CAPITAL LETTER U WITH CIRCUMFLEX ACCENT
| DC | 220 |  | LATIN CAPITAL LETTER U WITH DIAERESIS
| DD | 221 |  | LATIN CAPITAL LETTER Y WITH ACUTE ACCENT
| DE | 222 |  | LATIN CAPITAL LETTER THORN
| DF | 223 |  | LATIN SMALL LETTER SHARP S
|    |     |   |
| E0 | 224 |  | LATIN SMALL LETTER A WITH GRAVE ACCENT
| E1 | 225 |  | LATIN SMALL LETTER A WITH ACUTE ACCENT
| E2 | 226 |  | LATIN SMALL LETTER A WITH CIRCUMFLEX ACCENT
| E3 | 227 |  | LATIN SMALL LETTER A WITH TILDE
| E4 | 228 |  | LATIN SMALL LETTER A WITH DIAERESIS
| E5 | 229 |  | LATIN SMALL LETTER A WITH RING ABOVE
| E6 | 230 |  | LATIN SMALL LIGATURE AE
| E7 | 231 |  | LATIN SMALL LETTER C WITH CEDILLA
| E8 | 232 |  | LATIN SMALL LETTER E WITH GRAVE ACCENT
| E9 | 233 |  | LATIN SMALL LETTER E WITH ACUTE ACCENT
| EA | 234 |  | LATIN SMALL LETTER E WITH CIRCUMFLEX ACCENT
| EB | 235 |  | LATIN SMALL LETTER E WITH DIAERESIS
| EC | 236 |  | LATIN SMALL LETTER I WITH GRAVE ACCENT
| ED | 237 |  | LATIN SMALL LETTER I WITH ACUTE ACCENT
| EE | 238 |  | LATIN SMALL LETTER I WITH CIRCUMFLEX ACCENT
| EF | 239 |  | LATIN SMALL LETTER I WITH DIAERESIS
|    |     |   |
| F0 | 240 |  | LATIN SMALL LETTER ETH
| F1 | 241 |  | LATIN SMALL LETTER N WITH TILDE
| F2 | 242 |  | LATIN SMALL LETTER O WITH GRAVE ACCENT
| F3 | 243 |  | LATIN SMALL LETTER O WITH ACUTE ACCENT
| F4 | 244 |  | LATIN SMALL LETTER O WITH CIRCUMFLEX ACCENT
| F5 | 245 |  | LATIN SMALL LETTER O WITH TILDE
| F6 | 246 |  | LATIN SMALL LETTER O WITH DIAERESIS
| F7 | 247 |  | DIVISION SIGN
| F8 | 248 |  | LATIN SMALL LETTER O WITH OBLIQUE BAR
| F9 | 249 |  | LATIN SMALL LETTER U WITH GRAVE ACCENT
| FA | 250 |  | LATIN SMALL LETTER U WITH ACUTE ACCENT
| FB | 251 |  | LATIN SMALL LETTER U WITH CIRCUMFLEX ACCENT
| FC | 252 |  | LATIN SMALL LETTER U WITH DIAERESIS
| FD | 253 |  | LATIN SMALL LETTER Y WITH ACUTE ACCENT
| FE | 254 |  | LATIN SMALL LETTER THORN
| FF | 255 |  | LATIN SMALL LETTER Y WITH DIAERESIS
+----+-----+---+------------------------------------------------------

Footnote: ISO 10646 calls  a `ligature', but this is a
          letter in (at least some) Scandinavian languages.  Thus, it
          is not in the same, merely typographic `ligature' class as
          `oe' ({\oe} in {\LaTeX} convention) which was not included
          in the ISO8859-1 standard. 

***Tentative info***
Supposedly the Danish press, some months ago, reported that ISO has
changed the standard so from now on  and  are classified as
letters. 

If you can confirm or deny this, please let me know...
***Tentative info***



20. History
In April 1965, the ECMA (European Computer Manufacturer's Association)
stndardized ECMA-6.  This the character set is also (and more
commonly) also know under the names of ISO 646, US-ASCII or DIN 66003.

However, this standard only contained the basic Latin alphabet, with
no provisions for national characters in use all across Europe.  These
characters were later added by replacing several special characters
from the US-ASCII alphabet (such as {[|]}\ etc.).  These variants were
local to each country and were calle `national ISO 646 variants'.
Portability from one country to another was low, as each country had
their own national variant, and some of the special characters were
still needed (such as for programming C), which made this an
altogether unsatisfying solution.

In 1981, IBM released the IBM PC with an 8 bit character set, code
page 437.  The order of the characters added was somewhat confusing,
to say the least.  However, in 1982 the first hardware (DEC VT220 and
VT240 terminal) using a more satisfying character set, the DEC MCS
(Multilanguage Character Set) was released.

This character set was very similar to ISO 6937/2, which is
essentially equivalent to today's ISO 8859-1.  In March 1985, ECMA
standardized ECMA-94, which later came to be known as ISO 8859-1
through 8859-4.  However, ISO 8859-1 was officially stndardized by ISO
only in 1987.

1987 also saw the release of MS-DOS 3.3 which used Code Page 850.
Code Page 850 contains all characters from ISO 8859-1, making a
loss-free conversion possible.  Code Page 819 which was released later
goes one step further, as it is fully ISO 8859-1 compliant.

The ISO 8859-X standard was designed to allow as much interoperability
between character sets as possible.  Thus, all ISO 8859-X character
sets are a superset of US-ASCII and all character sets will render
English text properly.  Also, there is considerable overlap between
several character sets: a text written in German using the ISO 8859-1
character set can be correctly rendered in ISO 8859-2, the Eastern
European character set, where German is the primary foreign language
(-3, -4, -9, -10 supposedly also can display German text without
changes).

While ISO 8859-X was designed for considerable portability, texts are
still restricted mostly to their character set and portability to
other cultural areas is a problem.  One solution is to use a
meta-protocol (such as -> MIME) which specifies the character set
which was used to write a text and which causes the correct character
set to be used in displaying text.

A different approach to overloading the character set as done in the
ISO 8859-X standard (where the locations 0xa0 to 0xff are used to
encode national characters) is to use wider characters.  This is the
approach employed in Unicode (which is an enocing of Basic
MUlitlanguage Plane (BMP) of ISO/IEC 10646).  The downside to this
approach is that most of the software available today only accepts 8
bit wide characters (7 bit if you have bad luck :-( ), so the Unicode
approach is problematic.  This 8 bit restriction permeates nearly all
code in use today, including such system software (file systems,
process identifiers, etc.!).  To ease this problem somewhat, several
representations which map Unicode characters to a variable length 8
bit based encoding have been introduced (this encoding is called
UTF-8).  More information about Unicode can be obtained from URL
http://unicode.org.



21. Glossary: Acronyms, Names, etc.
i18n	    I<-- 18 letters -->n = Internationalization
e13n	    Europeanization
l10n	    Localization
ANSI        American National Standards Institute, the US member of ISO 
ASCII       American Standard Code of Information Interchange 
CP          Code Page 
CP850       Code Page 850, the most widely used MS DOS code page 
CR          Carriage Return 
CTAN server Comprehensive TeX Archive Network, the world's largest
            repository for TeX related material. It consists of three
            sites mirrowing each other: ftp.shsu.edu, ftp.tex.ac.uk,
            ftp.dante.de. The current configuration, including known
            mirrows, can be obtained by fingering ctan_us@ftp.shsu.edu
DEC         Digital Equipment Corp. 
DIN         Deutsche Industrie Norm (German Industry Norm)
DOS         Disk Operating System 
EBCDIC      Extended Binary Coded Decimal Interchange Code
             ---a proprietary IBM character set used on mainframes  
ECMA        European Computer Manufacturer's Association
emacs       Editing Macros, a family of popular text editors 
ESMTP       Enhanced SMTP 
Esperanto   A synthetic, ``universal'' language developed by
              Dr.~Zamenhof in~1887.
FSF         Free Software Foundation 
FTP         File Transmission Protocol 
GNU         GNU's not Unix, an FSF project 
HP          Hewlett Packard 
HP/UX       HP Unix 
IBM         International Business Machines Corp.
IEEE        Institute of Electrical and Electronics Engineers 
INRIA       Institut National de Recherche en Informatique et Automation 
IP          Internet Protocol 
ISO         International Standards Organization 
KOI8        ???---a popular encoding for Cyrillic on UNIX workstations 
\LaTeX{}    A macro package for \TeX{} 
LF          Linefeed 
MCS         DEC's Multilingual Character Set---the ISO 8859--1 draft standard 
MIME        Multi-Purpose Internet Mail Extension 
MS-DOS      Microsoft's program loader 
MTA         mail transfer agent 
MUA         mail user agent 
OS          Operating System 
OSF         the Open Software Foundation 
OSF/1       the Open Software Foundation's Unix, Revision 1 
PGP         Pretty Good Privacy, an encryption package 
POSIX       Portable Operating System Interface (an IEEE UNIX standard) 
PS          PostScript, Adobe's printer language 
RFC         Request for Comment, an Internet standard 
sed         stream editor, a UNIX file manipulation utility 
SMTP        Simple Mail Transfer Protocol 
TCP         Transmission Control Protocol 
\TeX{}      Donald Knuth's typesetting program 
UDP         User Datagram Protocol 
URL         a WWW Uniform Resource Locator 
US-ASCII    the US national variant of ISO 646, see ASCII 
VMS         Virtual Memory System---DEC's proprietary OS 
W3          WWW 
WWW         World Wide Web 
X11         X Window System 



22. Comments
This FAQ is somewhat Sun-centered, though I have tried to include
other machine types.  If you have figured out how to configure your
machine type, please let me (mike@vlsivie.tuwien.ac.at) know so that I
can include it in future revisions of this FAQ.



23. Home location of this document
23.1 www
You can find this and other i18n documents under URL
http://www.vlsivie.tuwien.ac.at/mike/i18n.html.

23.2 ftp
The most recent version of this document is available via anonymous
ftp from ftp.vlsivie.tuwien.ac.at under the file name
/pub/8bit/FAQ-ISO-8859-1

-----------------

Copyright  1994,1995,1996 Michael Gschwind (mike@vlsivie.tuwien.ac.at)

This document may be copied for non-commercial purposes, provided this
copyright notice appears.  Publication in any other form requires the
author's consent. (Distribution or publication bundled with a product
requires the author's consent, as does publication in any book,
journal or other work.)

Dieses Dokument darf unter Angabe dieser urheberrechtlichen
Bestimmungen zum Zwecke der nicht-kommerziellen Nutzung beliebig
vervielfltigt werden.  Die Publikation in jeglicher anderer Form
erfordert die Zustimmung des Autors.  (Verteilung oder Publikation mit
einem Produkt erfordert die Zustimmung des Autors, wie auch die
Verffentlichung in Bchern, Zeitschriften, oder anderen Werken.)

Local IspellDict: english
Michael Gschwind, Institut f. Technische Informatik, TU Wien
snail: Treitlstrae 3-182-2 || A-1040 Wien || Austria
email: mike@vlsivie.tuwien.ac.at   PGP key available via www (or email)
www  : URL:http://www.vlsivie.tuwien.ac.at/mike/mike.html
phone: +(43)(1)58801 8156	   fax: +(43)(1)586 9697

User Contributions:

Comment about this article, ask questions, or add new information about this topic:


[ Usenet FAQs | Web FAQs | Documents | RFC Index ]

Send corrections/additions to the FAQ Maintainer:
mike@vlsivie.tuwien.ac.at





Last Update March 27 2014 @ 02:11 PM