GNU recode, version {No value for ‘VERSION’}

The character set transliterator

Edition {No value for ‘EDITION’}, {No value for ‘UPDATED’}

{No value for ‘Francois’} Pinard

Permission is granted to make and distribute verbatim copies of this manual provided the copyright notice and this permission notice are preserved on all copies.

Permission is granted to copy and distribute modified versions of this manual under the conditions for verbatim copying, provided that the entire resulting derived work is distributed under the terms of a permission notice identical to this one.

Permission is granted to copy and distribute translations of this manual into another language, under the above conditions for modified versions, except that this permission notice may be stated in a translation approved by the Foundation.

1 What is the purpose of this program

This recode program has the purpose of converting files between various character sets and usages. When exact transliterations are not possible, as it is often the case, the program may get rid of the offending characters or fall back on approximations.

Let us coin the term charset to represent, without distinction, a character set “per se” or a particular usage of a character set. This program recognizes or produces around 150 such charsets. Since it can convert each charset to almost any other one, many thousands of different conversions are possible.

This tool pays special attention to superimposition of diacritics for French representation. This orientation is mostly historical, it does not impair the usefulness, generality or extensibility of the program.

1.1 Overview of charsets

Recoding is currently possible between most of the charsets described in RFC 1345. See section Charsets from RFC 1345.

Recode also handles some charsets in more specialized ways. These are:

usual 7-bit ASCII: without any diacritics, or else: using backspace for overstriking; Unisys’ ICON convention; TeX/LaTeX coding; easy French conventions for electronic mail;
8-bit extensions to ASCII: ISO Latin-1, Atari ST code, IBM’s code for the PC, Apple’s code for the Macintosh, NeXT code;
6-bit escaped ASCII based on CDC display code: 6/12 code from NOS; bang-bang code from Université de Montréal;
non-ASCII codes: three flavors of EBCDIC.

The recent introduction of RFC 1345 in GNU recode has brought with it a few charsets having the functionality of older ones, but yet being different in subtle ways. The effects have not been fully investigated yet, so for now, clashes are avoided, the old and new charsets are kept well separate. For example, wizards would be interested in comparing the output of these two commands:

recode -vh IBM-PC:Apple-Mac
recode -vh IBM437:macintosh

The first command uses only charsets prior to RFC 1345 introduction. Both methods give different recodings, the first also properly recodes end of lines. These differences are annoying, the fuzziness will have to be explained and settle down one day.

1.2 Contributions and bug reports

Even being the recode author and current maintainer, I am no specialist in charset standards. I only made recode along the years to solve my own needs, but felt it was applicable for the needs of others. Some GNU people liked the program structure and suggested to make it more widely available. I rely on GNU users judgment for what is best to be done next.

Properly protecting GNU recode about possible copyright fights is a pain for me and for contributors, but we cannot avoid addressing the issue in the long run. Besides, the Free Software Foundation, which mandates the GNU project, is very sensible to this matter. GNU standards require that I be cautious before looking at copyrighted code. The safest and simplest way for me is to gather ideas and reprogram them anew, even if this might slow me down considerably. For contributions going beyond a few lines of code here and there, the FSF definitely requires employer disclaimers and copyright assignments in writing.

Many users contributed to GNU recode already, I am grateful to them for their interest and involvement. Some suggestions can be integrated quickly while some others have to be delayed, I have to draw a line somewhere when time comes to make a new release, about what would go in it and what would go in the next. Also, when you contribute something to recode, please explain what it is about. Do not take for granted that I know those charsets which are familiar to you. Your explanations could well find their way into this documentation, too.

Mail suggestions, documentation errors and bug reports to bug-gnu-utils@prep.ai.mit.edu or, if you prefer, directly to Francois Pinard ‘pinard@iro.umontreal.ca’. Do not be afraid to report details, because this program is the mere aggregation of hundreds of details.

2 How to use this program

The general format of the program call is one of:

recode [option]… [charset]
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.

The available options are:

-C

--copyright

Given this option, all other parameters and options are ignored. The program prints briefly the Copyright and copying conditions. See the file ‘COPYING’ in the distribution for full statement of the Copyright and copying conditions.

-a

--auto-check

In this special mode, recode diagnostics itself by analyzing connectivity of the various charsets and reporting on standard output. No file will be recoded.

There might be one non-option argument, in which case it is interpreted as a charset name, possibly abbreviated to any non ambiguous prefix. recode will then study all recodings having the given charset as a starting or ending point. If there is no such non-option argument, recode will study all possible recodings.

For each possible pair of different charsets, it prints on standard output how many single steps are needed for achieving the recoding and how many can be saved by step merging. If a recoding cannot be done, the word ‘UNACHIEVABLE’ is printed instead. However, this special line is completely suppressed if option -x specified some charset to ignore.

The option -hname affects the resulting output, because there are more merging rules when this option is in effect. Other options affect the result: -d, -g and, notably, -s.

There was a time, in GNU recode development, when this option was reasonably interesting. With the greater number of handled charsets, it became inordinately slow, taking on the order of one hour of wall clock time, while generating a great deal of output. This option is not practical anymore when used without a charset parameter. However, it can be made slightly more usable, together with option -x., which effectively disables most RFC 1345 charsets from the report.

-c

--colons

With Texte Easy French conventions, use the column : instead of the double-quote " for marking diaeresis. See section ASCII with easy French conventions.

-d

--diacritics

While converting to or from one of HTML or LaTeX charset, limit conversion to diacritics only. This is particularly useful when people write what would be valid HTML, TeX or LaTeX files, if only they were using provided sequences for applying diacritics instead of using the diacriticized characters directly from the underlying character set.

While converting to HTML or LaTeX charset, this option assumes that non-diacriticized special characters are properly coded or protected, so recode will transmit them literally. While converting the other way, this option prevents all attempts at recognizing coded or protected versions of non-diacriticized special characters of the other charset. See section World Wide Web representations. See section ASCII with LaTeX codes.

-f

--force

It is planned that some future version of recode will protect you against recoding a file irreversibly over itself. However, please keep vividly in mind that this protection is not yet active in recode. When the protection will be enforced, option ‘-f’ will become mandatory for a file to be replaced by some recoding of its contents, if such conversion is loosing information. For now, recode acts as if option ‘-f’ was always selected.

In preparation for the time this option will become mandatory, you may start using ‘-f’ right away in scripts calling recode, when you know this is the reasonnable thing to do.

-g

--graphics

This option is only meaningful while getting out of the IBM-PC charset. In this charset, characters 176 to 223 are used for constructing rulers and boxes, using simple or double horizontal or vertical lines. This option forces the automatic selection of ASCII characters for approximating these rulers and boxes, at cost of making the transformation irreversible. Option -g implies -f.

-h[name]

--header[=name]

Instead of recoding files, recode writes a C source file on standard output and exits. This source is meant to be included in a regular C program: its purpose is to declare and initialize an array, named name, which represents the requested recoding. If name is not specified, then it defaults to before_to_after, where before is the starting charset and after is the goal charset.

Even if recode tries its best, this option does not always succeed in producing the requested C table. It will however, provided the recoding can be internally represented by only one step after the optimization phase, and if this merged step conveys a one-to-one or a one-to-many explicit table. But this is all fairly technical. Better try and see!

Beware that other options might affect the produced C tables, these are: -d, -g and, particularly, -s.

-i

--sequence=files

When the recoding requires a combination of two or more elementary recoding steps, this option forces many passes over the data, using intermediate files between passes. This is the default behavior when files are recoded over themselves. If this option is selected in filter mode, that is, when the program reads standard input and writes standard output, it might take longer for programs further down the pipe chain to start receiving some recoded data.

-l[format]

--list[=format]

This option asks for information about all charsets, or about one particular charset. No file will be recoded.

If there is no non-option arguments, recode ignores the format value of the option, it writes a sorted list of charset names on standard output, one per line. When a charset name have aliases or synonyms, they follow the true charset name on its line, presented in lexicographical order from left to right. This list is over one hundred lines. It is best used with grep, as in:

recode -l | grep greek

There might be one non-option argument, in which case it is interpreted as a charset name, possibly abbreviated to any non ambiguous prefix. This particular usage of the -l option is obeyed only for charsets having an RFC 1345 style internal description. Even if most charsets have this property, some do not, then option -l cannot be used to detail these particular charsets. For knowing if a particular charset can be listed this way, you should merely try and see if this works. The format value of the option is a keyword from the following list. Keywords may be abbreviated by dropping suffix letters, and even reduced to the first letter only:

decimal: This format asks for the production on standard output of a concise tabular display of the charset, in which character code values are expressed in decimal.
octal: This format uses octal instead of decimal in the concise tabular display of the charset.
hexadecimal: This format uses hexadecimal instead of decimal in the concise tabular display of the charset.
full: This format requests an extensive display of the charset on standard output, using one line per character showing its decimal, hexadecimal and octal code values, and also a descriptive comment which is indeed the 10646 character name.

When option -l is used together with a charset argument, the format defaults to decimal.

-o

--sequence=popen

When the recoding requires a combination of two or more elementary recoding steps, this option forces the creation of a chain of program instances initiated through the popen(3) library call, all operating in parallel. In filter mode, costing the overhead of multiple program initializations, recoded data will be available soon after the program starts, even if many elementary recoding steps are required.

If, at installation time, the popen(3) call is said to be unavailable, selecting option -o is equivalent to selecting option -i.

-p

--sequence=pipe

When the recoding requires a combination of two or more elementary recoding steps, this option forces the program to fork itself into a few copies interconnected with pipes, using the pipe(2) system call. All copies of the program operate in parallel. This method is similar to the method used through option -o, but is more efficient because the program initializes only once. This is the default behavior in filter mode. If this option is used when files are recoded over themselves, this should also save disk space because some temporary files might not be needed, at cost of more system overhead.

If, at installation time, the pipe(2) call is said to be unavailable, selecting option -p is equivalent to selecting option -o. If both pipe(2) and popen(3) are unavailable, selecting option -p is equivalent to selecting option -i.

-q

--quiet

--silent

This option has the sole purpose of inhibiting diagnostic messages about irreversible recodings.

This option is set automatically for the children processes, when recode splits itself in many collaborating copies. Doing so, the diagnostic is issued only once by the parent. See options -o and -p.

-s

--strict

By using this option, the user requests that recode be very strict while recoding a file, merely loosing in the transformation any character which is not explicitly mapped from a charset to another. This option renders the recoding less likely reversible, so it also implies option -f. Also See section Reversibility issues.

-t

--touch

The touch option is meaningful only when files are recoded over themselves. Without it, the time-stamps associated with files are preserved, to reflect the fact that changing the code of a file does not really alter its informational contents. When the user wants the recoded files to be time-stamped at the recoding time, this option inhibits the automatic protection of the time-stamps.

-v

--verbose

Before doing any recoding, the program will first print on ‘stderr’ the list of all intermediate charsets planned for recoding, starting with the before charset and ending with the after charset. It also prints an indication of the recoding quality, as one of the word ‘reversible’, ‘one to one’, ‘one to many’, ‘many to one’ or ‘many to many’.

This information will appear once or twice. It is shown a second time only when the optimization and step merging phase succeeds in creating a new single step.

This option also has a second effect. The program will print on ‘stderr’ one message per file recoded, so to let the user informed of the progress of its command.

An easy way to know beforehand the sequence or quality of a recoding is by using the command such as:

recode -v before:after < /dev/null

using the fact that, so far in recode, an empty input file produces an empty output file.

-x=charset

--ignore=charset

This option tells the program to ignore any recoding path through the specified charset, so disabling any single step using this charset as a start or end point. This may be used when the user wants to force recode in using an alternate recoding path.

charset may be abbreviated to any unambiguous prefix. For convenience, the value ‘.’ is an alias for ‘RFC 1345’, so the option -x. effectively disables all RFC 1345 tables at once.

--help

The program merely prints a page of help on standard output, and exits without doing any recoding.

--version

The program merely prints its version numbers on standard output, and exits without doing anything else.

The before:after argument specifies the start charset and the goal charset. The allowable values for before or after are described in the remainder of this document. Charsets may have predefined alternate names, or aliases, which are equally acceptable.

In the before:after argument only, a backslash may be used to quote the next character of a charset name. This might be useful for preventing a colon to be mistakenly interpreted as the separator between before and after. Rather, the colon could be omitted, because while recognizing a charset name or alias, GNU recode ignores all characters besides letters and digits. There is also no distinction between upper and lower case. Charset names or aliases may always be abbreviated to any unambiguous prefix.

One or both of the before or after keywords may be omitted, but the colon which separates them cannot. An omitted keyword implies the usual or default code in usage on the system where this program is installed. Usually, this default code is Latin-1 for UNIX systems or IBM-PC for MS-DOS machines.

3 Reversibility issues

Even if GNU recode tries hard at keeping the recodings reversible, you should not develop an unconditional confidence in its ability to do so. You ought to keep only reasonable expectations about reverse recodings. In particular, consider:

Most transformations are fully reversible for all inputs, but loose this property whenever -s is specified.
A few transformations are not meant to be reversible, by design.
Reversibility sometimes depends on actual file contents and cannot be ascertained beforehand, without reading the file.
Reversibility is never absolute across successive versions of this program. Even correcting a small bug in a mapping could induce slight discrepancies later.
Reversibility is easily lost by merging. This is best explained through an example. If you reversibly recode a file from charset ‘A’ to charset ‘B’, then you reversibly recode the result from charset ‘B’ to charset ‘C’, you cannot expect to recover the original file by merely recoding from charset ‘C’ directly to charset ‘A’. You will instead have to recode from charset ‘C’ back to charset ‘B’, and only then from charset ‘B’ to charset ‘A’.
Faulty files create a particular problem. Consider an example, recoding from IBM-PC to Latin-1. End of lines are represented as ‘\r\n’ in IBM-PC and as ‘\n’ in Latin-1. There is no way by which a faulty IBM-PC file containing a ‘\n’ not preceded by ‘\r’ be translated into a Latin-1 file, and then back.
There is another difficulty arising from code equivalences. For example, in a LaTeX charset file, the string ‘\^\i{}’ could be recoded back and forth through another charset and become ‘\^{\i}’. Even if the resulting file is equivalent to the original one, it is not identical.

Unless option -s is used, recode automatically tries to fill mappings with invented correspondences, often making them fully reversible. This filling is not made at random. The algorithm tries to stick to the identity mapping and, when this is not possible, it prefers generating many small permutation cycles, each involving only a few codes.

For example, here is how IBM-PC code 186 gets translated to control-U in Latin-1. Control-U is 21. Code 21 is the IBM-PC section sign, which is 167 in Latin-1. recode cannot reciprocate 167 to 21, because 167 is the masculine ordinal indicator on IBM PC’s, which is 186 in Latin-1. Code 186 in IBM PC’s has no Latin-1 equivalent; by assigning back to 21, recode closes this short permutation loop.

As a consequence of this map filling, recode may sometimes produce funny characters. They may look annoying, they are nevertheless helpful when one changes his/her mind and wants to revert to the prior recoding. If you cannot stand these, use option -s, which asks for a very strict recoding.

This map filling sometimes has another surprising consequence. In some cases, recode seems to copy a file without recoding it. But in fact, it does. As an illuminating example, consider you requested:

recode l1:us < File-Latin1 > File-ASCII
cmp File-Latin1 File-ASCII

then cmp will not report any difference. This is quite normal. Latin-1 gets correctly recoded to ASCII for charsets commonalities (which are the first 128 characters, in this case). The remaining last 128 Latin-1 characters have no ASCII correspondent. Instead of loosing them, recode elects to map them to unspecified characters of ASCII, so making the recoding reversible. The simplest way of achieving this is merely to keep those last 128 characters unchanged. The overall effect is copying the file verbatim.

If you feel this behavior is too generous and if you do not wish to care about reversibility, simply use option -s. By doing so, recode will strictly map only those Latin-1 characters which have an ASCII equivalent, and will merely drop those which do not. Then, there is more chance that you will observe a difference between the input and the output file.

4 Charsets from RFC 1345

In the GNU recode distribution, there is a copy of RFC 1345:

“Character Mnemonics & Character Sets”, K. Simonsen, Request for Comments no. 1345, Network Working Group, June 1992.

This document is also available by anonymous ftp at ‘nic.ddn.mil’ in directory ‘rfc’ as file ‘rfc1345.txt’. This report defines many character mnemonics and character sets.

GNU recode implements most of RFC 1345, however:

It does not recognize 16-bits charsets: GB_2312-80, JIS_C6226-1978, JIS_C6226-1983, JIS_X0212-1990 and KS_C_5601-1987.
It does not recognize those charsets which combine two characters for representing a third: ANSI_X3.110-1983, ISO_6937-2-add, T.101-G2, T.61-8bit, iso-ir-90 and videotex-suppl.
It interprets the charset isoir91 as NATS-DANO (alias iso-ir-9-1), not as JIS_C6229-1984-a (alias iso-ir-91). So better avoid using these two alias names.
It interprets the charset isoir92 as NATS-DANO-ADD (alias iso-ir-9-2), not as JIS_C6229-1984-b (alias iso-ir-92). So better avoid using these two alias names.
It ignores all about code overloading, but still processes correctly the remainder of dk-us and us-dk.

Keld Simonsen ‘keld@dkuug.dk’ did most of RFC 1345 himself, with some funding from Danish Standards and Nordic standards (INSTA) project. He also did the character set design work, with substantial input from Olle Jaernefors. Keld typed in almost all of the tables, some have been contributed. A number of people have checked the tables in various ways. The RFC lists a number of people who helped.

Internally, RFC 1345 associates which each character an unambiguous mnemonic of (usually) one or two characters, taken from ISO 646, a minimal set of 83 characters. The charset made up by these mnemonics is available in recode under the name RFC 1345, with . being accepted as a short alias.

Even if the mnemonics are unambiguous taken separately, strings made up by concatenating these mnemonics are ambiguous and cannot be safely interpreted. So recode only allows converting to RFC 1345, never from it. However, special machinery in the program allows for converting through RFC 1345, when RFC 1345 is neither the initial nor the final charset of the conversion sequence.

Recoding directly to . has the main goal of letting the user examine foreign charsets. We cannot do much, mechanically, with the result. For increased readability, as a matter of convenience, SP is left as a single space and LF becomes a newline.

5 Charsets based on ASCII

5.1 Usual ASCII

This charset is available in recode under the name ASCII. In fact, it’s true name is ANSI_X3.4-1968 as per RFC 1345, accepted aliases being ANSI_X3.4-1986, ASCII, IBM367, ISO646-US, ISO_646.irv:1991, US-ASCII, cp367, iso-ir-6 and us. The shortest way of specifying it in recode is us.

This documentation used to include ASCII tables. They have been removed since recode can now recreate these (and a lot of others) easily:

recode -lf us                   for commented ASCII
recode -ld us                   for concise decimal table
recode -lo us                   for concise octal table
recode -lh us                   for concise hexadecimal table

5.2 ASCII extended by Latin Alphabets

This charset is available in recode under the name Latin-1. In fact, it’s true name is ISO_8859-1:1987 as per RFC 1345, accepted aliases being CP819, IBM819, ISO-8859-1, ISO_8859-1, iso-ir-100, l1 and Latin-1. The shortest way of specifying it in recode is l1.

This charset corresponds to the ISO Latin Alphabet 1. It is an eight-bit code which coincides with ASCII for the lower half.

This documentation used to include Latin-1 tables. They have been removed since recode can now recreate these (and a lot of others) easily:

recode -lf l1                   for commented ISO Latin-1
recode -ld l1                   for concise decimal table
recode -lo l1                   for concise octal table
recode -lh l1                   for concise hexadecimal table

The following from ‘lasko@video.dec.com’ (Tim Lasko), with no date.

ISO Latin-1, or more completely ISO Latin Alphabet No 1, is now an international standard as of February 1987 (IS 8859, Part 1). For those American USEnet’rs that care, the 8-bit ASCII standard, which is essentially the same code, is going through the final administrative processes prior to publication.

ISO Latin-1 (IS 8859/1) is actually one of an entire family of eight-bit one-byte character sets, all having ASCII on the left hand side, and with varying repertoires on the right hand side:

Latin Alphabet No 1 (caters to Western Europe - now approved).
Latin Alphabet No 2 (caters to Eastern Europe - now approved).
Latin Alphabet No 3 (caters to SE Europe + others - in draft ballot).
Latin Alphabet No 4 (caters to Northern Europe - in draft ballot).
Latin-Cyrillic alphabet (right half all Cyrillic - processing currently suspended pending USSR input).
Latin-Arabic alphabet (right half all Arabic - now approved).
Latin-Greek alphabet (right half Greek + symbols - in draft ballot).
Latin-Hebrew alphabet (right half Hebrew + symbols - proposed).

5.3 ASCII 7-bits, <BS> to overstrike

This charset is available in recode under the name ASCII-BS, with BS as an acceptable alias.

The file is straight ASCII, seven bits only. According to the definition of ASCII: diacritics are applied by a sequence of three characters: the letter, one <BS>, the diacritic mark. We deviate slightly from this by exchanging the diacritic mark and the letter so, on a screen device, the diacritic will disappear and let the letter alone. At recognition time, both methods are acceptable.

The French quotes are coded by the sequences: < <BS> " or " <BS> < for the opening quote and > <BS> " or " <BS> > for the closing quote. This artificial convention was inherited in straight ASCII-BS from habits around Bang-Bang entry, and is not well known. But we decided to stick to it so that ASCII-BS charset will not loose French quotes.

The ASCII-BS charset is independent of ASCII, and different. The following examples demonstrate this, knowing at advance that ‘!2’ is the Bang-Bang way of representing an e with an acute accent. Compare:

% echo \!2 | recode -v bang:us | od -bc
Bang-Bang -> ISO_8859-1:1987 -> RFC 1345 -> ANSI_X3.4-1968 (many to one)
Simplified to: Bang-Bang -> ISO_8859-1:1987 -> ANSI_X3.4-1968 (many to one)
0000000 351 012
        351  \n
0000002

with:

% echo \!2 | recode -v bang:bs | od -bc
Bang-Bang -> ISO_8859-1:1987 -> ASCII-BS (many to many)
0000000 047 010 145 012
          '  \b   e  \n
0000004

In the first case, the e with an acute accent is merely transmitted by the Latin-1:ASCII mapping, not having a special recoding rule for it. In the Latin-1:ASCII-BS case, the acute accent is applied over the e with a backspace: diacriticized characters have special rules. For the ASCII-BS charset, reversibility is still possible, but there might be difficult cases.

5.4 ASCII without diacritics nor underline

This charset is available in recode under the name flat.

This code is ASCII expunged of all diacritics and underlines, as long as they are applied using three character sequences, with <BS> in the middle. Also, despite slightly unrelated, each control character is represented by a sequence of two or three graphic characters. The newline character, however, keeps its functionality and is not represented.

Note that charset flat is a terminal charset. We can convert to flat, but not from it.

6 Charsets based on IBM

6.1 EBCDIC code

This charset is the IBM’s external binary coded decimal for interchange coding. This is an eight bits code. The following three variants were implemented in GNU recode independently of RFC 1345:

EBCDIC

GNU recode us:ebcdic conversion is identical to GNU dd ebcdic conversion, and recode ebcdic:us conversion is identical to GNU dd ascii conversion. This charset also represents the way Control Data Corporation relates EBCDIC to 8-bits ASCII.

EBCDIC-CCC

GNU recode us:ebcdic-ccc or ebcdic-ccc:us conversions represent the way Concurrent Computer Corporation (formerly Perkin Elmer) relates EBCDIC to 8-bits ASCII.

EBCDIC-IBM

GNU recode us:ebcdic-ibm conversion is almost identical to GNU dd ibm conversion. Given the exact dd ibm conversion table, recode once said:

Codes  91 and 213 both recode to 173
Codes  93 and 229 both recode to 189
No character recodes to  74
No character recodes to 106

So I arbitrarily chose to recode 213 by 74 and 229 by 106. This makes the EBCDIC-IBM recoding reversible, but this is not necessarily the best correction. In any case, I believe GNU dd should be corrected, and preferably, GNU dd and GNU recode should agree on the same correction. So, this table may change once again.

RFC 1345 brings in recode 15 other EBCDIC charsets, and 21 other charsets having EBCDIC in at least one of their alias names. You can get a list of all these by executing:

recode -l | grep ebcdic

6.2 IBM’s PC code

This charset is available in recode under the name IBM-PC. There are a few discrepancies between this charset and the very similar RFC 1345 charset ibm437, which have not been analyzed yet, so the charsets are being kept separate for now. This might change in the future.

The file was obtained or is aimed towards a PC microcomputer from IBM or any compatible. This is an eight-bit code.

6.3 Unisys’ ICON code

This charset is available in recode under the name Icon-QNX, with QNX as an acceptable alias.

The file is using Unisys’ Icon way to represent diacritics with code 25 escape sequences, under the system QNX. This is a seven-bit code, even if eight-bit codes can flow through as part of IBM-PC charset.

7 Charsets based on CDC

7.1 Control Data’s Display Code

This code is not available in recode, but repeated here for reference. This is a 6-bit code used on CDC mainframes.

Octal display code to graphic       Octal display code to octal ASCII

00  :    20  P    40  5   60  #     00 072  20 120  40 065  60 043
01  A    21  Q    41  6   61  [     01 101  21 121  41 066  61 133
02  B    22  R    42  7   62  ]     02 102  22 122  42 067  62 135
03  C    23  S    43  8   63  %     03 103  23 123  43 070  63 045
04  D    24  T    44  9   64  "     04 104  24 124  44 071  64 042
05  E    25  U    45  +   65  _     05 105  25 125  45 053  65 137
06  F    26  V    46  -   66  !     06 106  26 126  46 055  66 041
07  G    27  W    47  *   67  &     07 107  27 127  47 052  67 046
10  H    30  X    50  /   70  '     10 110  30 130  50 057  70 047
11  I    31  Y    51  (   71  ?     11 111  31 131  51 050  71 077
12  J    32  Z    52  )   72  <     12 112  32 132  52 051  72 074
13  K    33  0    53  $   73  >     13 113  33 060  53 044  73 076
14  L    34  1    54  =   74  @     14 114  34 061  54 075  74 100
15  M    35  2    55      75  \     15 115  35 062  55 040  75 134
16  N    36  3    56  ,   76  ^     16 116  36 063  56 054  76 136
17  O    37  4    57  .   77  ;     17 117  37 064  57 056  77 073

7.2 ASCII 6/12 from NOS

This charset is available in recode under the name CDC-NOS, with NOS as an acceptable alias.

This is one of the charset in use on CDC Cyber NOS systems to represent ASCII, sometimes named NOS 6/12 code for coding ASCII. This code is also known as caret ASCII. It is based on a six bits character set in which small letters and control characters are coded using a ^ escape and, sometimes, a @ escape.

The routines given here presume that the six bits code is already expressed in ASCII by the communication channel, with embedded ASCII ^ and @ escapes.

Here is a table showing which characters are being used to encode each ASCII character.

000  ^5  020  ^#  040     060  0  100 @A  120  P  140  @G  160  ^P
001  ^6  021  ^[  041  !  061  1  101  A  121  Q  141  ^A  161  ^Q
002  ^7  022  ^]  042  "  062  2  102  B  122  R  142  ^B  162  ^R
003  ^8  023  ^%  043  #  063  3  103  C  123  S  143  ^C  163  ^S
004  ^9  024  ^"  044  $  064  4  104  D  124  T  144  ^D  164  ^T
005  ^+  025  ^_  045  %  065  5  105  E  125  U  145  ^E  165  ^U
006  ^-  026  ^!  046  &  066  6  106  F  126  V  146  ^F  166  ^V
007  ^*  027  ^&  047  '  067  7  107  G  127  W  147  ^G  167  ^W
010  ^/  030  ^'  050  (  070  8  110  H  130  X  150  ^H  170  ^X
011  ^(  031  ^?  051  )  071  9  111  I  131  Y  151  ^I  171  ^Y
012  ^)  032  ^<  052  *  072 @D  112  J  132  Z  152  ^J  172  ^Z
013  ^$  033  ^>  053  +  073  ;  113  K  133  [  153  ^K  173  ^0
014  ^=  034  ^@  054  ,  074  <  114  L  134  \  154  ^L  174  ^1
015  ^   035  ^\  055  -  075  =  115  M  135  ]  155  ^M  175  ^2
016  ^,  036  ^^  056  .  076  >  116  N  136 @B  156  ^N  176  ^3
017  ^.  037  ^;  057  /  077  ?  117  O  137  _  157  ^O  177  ^4

7.3 ASCII “bang bang”

This charset is available in recode under the name Bang-Bang.

This is the local code in use on Cybers at Universite de Montreal, which grave and serious people there prefer to name ASCII code display. This code is also known as Bang-bang. It is based on a six bits character set in which capitals, French diacritics and a few others are coded using an ! escape followed by a single character, and control characters using a double ! escape followed by a single character.

The routines given here presume that the six bits code is already expressed in ASCII by the communication channel, with embedded ASCII ! escapes.

Here is a table showing which characters are being used to encode each ASCII character.

000 !!@  020 !!P  040    060 0  100 @   120 !P  140 !@ 160 P
001 !!A  021 !!Q  041 !" 061 1  101 !A  121 !Q  141 A  161 Q
002 !!B  022 !!R  042 "  062 2  102 !B  122 !R  142 B  162 R
003 !!C  023 !!S  043 #  063 3  103 !C  123 !S  143 C  163 S
004 !!D  024 !!T  044 $  064 4  104 !D  124 !T  144 D  164 T
005 !!E  025 !!U  045 %  065 5  105 !E  125 !U  145 E  165 U
006 !!F  026 !!V  046 &  066 6  106 !F  126 !V  146 F  166 V
007 !!G  027 !!W  047 '  067 7  107 !G  127 !W  147 G  167 W
010 !!H  030 !!X  050 (  070 8  110 !H  130 !X  150 H  170 X
011 !!I  031 !!Y  051 )  071 9  111 !I  131 !Y  151 I  171 Y
012 !!J  032 !!Z  052 *  072 :  112 !J  132 !Z  152 J  172 Z
013 !!K  033 !![  053 +  073 ;  113 !K  133 [   153 K  173 ![
014 !!L  034 !!\  054 ,  074 <  114 !L  134 \   154 L  174 !\
015 !!M  035 !!]  055 -  075 =  115 !M  135 ]   155 M  175 !]
016 !!N  036 !!^  056 .  076 >  116 !N  136 ^   156 N  176 !^
017 !!O  037 !!_  057 /  077 ?  117 !O  137 _   157 O  177 !_

8 Non-IBM micro-computer charsets

8.1 Apple’s Macintosh code

This charset is available in recode under the name Apple-Mac. There are a few discrepancies between this charset and the very similar RFC 1345 charset macintosh, which have not been analyzed yet, so the charsets are being kept separate for now. This might change in the future.

The file has been obtained or is aimed to a Macintosh micro-computer from Apple. This is an eight bit code. The file is the data fork only.

8.2 Atari ST code

This charset is available in recode under the name AtariST.

This is the character set used on the Atari ST/TT/Falcon. This is similar to IBM-PC, but differs in some details (includes some more accented characters, the graphic characters are mostly replaced by hebrew characters, and there is a true German <sharp s> different from Greek <beta>).

About the end-of-line conversions: the canonical end-of-line on the Atari is ‘\r\n’, but unlike IBM-PC, the OS makes no difference between text and binary input/output; it is up to the application how to interpret the data. In fact, most of the libraries that come with compilers can grok both ‘\r\n’ and ‘\n’ as end of lines. Many of the users who also have access to Unix systems prefer ‘\n’ to ease porting Unix utilities. So, for easing reversibility, recode tries to let ‘\r’ undisturbed through recodings.

8.3 NeXT international code

This charset is available in recode under the name NeXT.

The NeXT encoding is an extension to the ISO Latin-1 ASCII encoding used by NeXT under the system NeXTSTEP. It is identical to Latin-1 for the positions 0-127. In the position 128-255, NeXT added some chars and shuffled them around a little bit (for some unknown reason).

9 Some other charsets

Even if these charsets were originally added to recode for handling texts written in French, they find other uses. We did use them lot for writing French diacriticized texts in the past, so recode knows how to handle these particularly well for French texts.

9.1 ASCII with LaTeX codes

This charset is available in recode under the name LaTeX and has ltex as an alias. It is used for ASCII files coded to be read by LaTeX or, in certain cases, by TeX.

Whenever you recode from another charset to LaTeX, beware that all occurrences of backslashes \ are usually translated into the string ‘\backslash{}’. However, in practice, people often use backslashes in the other charset for introducing TeX commands, compromising it: it is not pure TeX, nor it is pure other charset. This translation of backslashes into ‘\backslash{}’ can be rather inconvenient, it may be inhibited through the command option -d.

9.2 ASCII with easy French conventions

This charset is available in recode under the name Texte and has txte for an alias.

This charset is a seven bits code, identical to ASCII-BS, save for French diacritics which are noted using a slightly different convention.

At text entry time, these conventions provide a little speed up. At read time, they slightly improve the readability over a few alternate ways of coding diacritics. Of course, it would better to have a specialized keyboard to make direct eight bits entries and fonts for immediately displaying eight bit ISO Latin-1 characters. But not everybody is so fortunate. In several mailing environments, the eight bit is often willingfully destroyed.

Easy French has been in use in France for a while. I only slightly adapted it (the diaeresis option) to make it more comfortable to several usages in Québec originating from Université de Montréal. In fact, the main problem for me was not to necessarily to invent Easy French, but to recognize the “best” convention to use, (best is not being defined, here) and to try to solve the main pitfalls associated with the selected convention.

9.2.1 Diacritics

French quotes (sometimes called “angle quotes”) are noted the same way English quotes are noted in TeX, id est by `` and ''.

No effort has been put to preserve Latin ligatures (ae, oe) which are representable in several other charsets. So, these ligatures may be lost through Easy French conventions.

This is almost the French convention for simplified diacritics entry:

e': Acute accent
e`: Grave accent
e^: Circumflex accent
e": Diaeresis
c,: Cedilla

In some countries, : is used instead of " to mark diaeresis. recode support one convention on a single call, depending on the -c option of the recode command.

The convention is prone to loosing information, because the diacritic meaning overloads some characters that already have other uses. To alleviate this, some knowledge of the French language is boosted into the recognition routines. So, the following subtleties are systematically obeyed by the various recognizers.

A single quote which follows a e does not necessarily means an acute accent if it is followed by a single other one. For example:

e'

will give an e with an acute accent.

e''

will give a simple e, with a closing quotation mark.

e'''

will give an e with an acute accent, followed by a closing quotation mark.

There is a problem induced by this convention if there are English quotations with a French text. In sentences like:
```
There's a meeting at Archie's restaurant.
```
the single quotes will be mistaken twice for acute accents. So English contractions and suffix possessives could be mangled.
A double quote or colon, depending on -c option, which follows a vowel is interpreted as diaeresis only if it is followed by another letter. But there are in French several words that end with a diaeresis, the program also recognizes them. See section List of words ending with diaeresis, for a study of all the problematic cases.
A comma which follows a c is interpreted as a cedilla only if it is followed by one of the vowels a, o and u.

9.2.2 List of words ending with diaeresis

Here is a classification of all cases of a diaeresis at the end of a French word:

Words ending in “igue”
- - Feminine words without a relative masculine: ‘besaigue"’ and ‘cigue"’.
- - Feminine words with a relative masculine (1): ‘aigue"’, ‘ambigue"’, ‘contigue"’, ‘exigue"’, ‘subaigue"’ and ‘suraigue"’.
Words not ending in “igue”
- - Ended by “i” (2): ‘ai"’, ‘congai"’, ‘goi"’, ‘hai"kai"’, ‘inoui"’, ‘sai"’, ‘samurai"’, ‘thai"’ and ‘tokai"’.
- - Ended by “e”: ‘canoe"’.
- - Ended by “u” (3): ‘Esau"’.

Notes:

There are supposed to be seven words in this case. So, one is missing.
Look at one of the following sentences (the second has to be interpreted with the -c option):
```
"Ai"e!  Voici le proble`me que j'ai"
Ai:e!  Voici le proble`me que j'ai:
```
There is an ambiguity between an ‘ai"’, the small animal, and the indicative future of avoir (first person singular), when followed by what could be a diaeresis mark. Hopefully, the case is solved by the fact that an apostrophe always precedes the verb and almost never the animal.
I did not pay attention to proper nouns, but this one showed up as being fairly evident.

Just to complete this topic, note that it would be wrong to make a rule for all words ending in “igue” as needing a diaerisis. Here are counter-examples: ‘becfigue’, ‘be`sigue’, ‘bigue’, ‘bordigue’, ‘bourdigue’, ‘brigue’, ‘contre-digue’, ‘digue’, ‘d'intrigue’, ‘fatigue’, ‘figue’, ‘garrigue’, ‘gigue’, ‘igue’, ‘intrigue’, ‘ligue’, ‘prodigue’, ‘sarigue’ and ‘zigue’.

9.3 World Wide Web representations

This charset is available in recode under the name HTML and has w3 and WWW for aliases.

HTML texts used by World Wide Web limit themselves to 7-bit characters internally, special sequences beginning with an ampersand & and ending with a semicolon ; are used for representing characters from Latin-1 having the 8th bit set.

When you recode from another charset to HTML, beware that all occurrences of ampersands are usually translated into the string ‘&’, similarly, left angle brackets < are translated into ‘<’ and right angle brackets > are translated into ‘>’. However, in practice, people often use ampersands and angle brackets in the other charset for introducing HTML commands, compromising it: it is not pure HTML, not it is pure other charset. These three translations can be rather inconvenient, they may be specifically inhibited through the command option -d.

10 Internal aspects

Suppose that four elementary steps are selected at path optimization time. Then recode will split itself into four different tasks interconnected with pipes, logically equivalent to:

step1 <input | step2 | step3 | step4 >output

10.1 Overall organization

The main driver constructs, while initializing all conversion modules, a table giving all the conversion routines available (single steps) and for each, the starting charset and the ending charset. If we consider these charsets as being the nodes of a directed graph, each single step may be considered as oriented arc from one node to the other. A cost is attributed to each arc: for example, a high penalty is given to single steps which are prone to loosing characters, a low penalty is given to those which need studying more than one input character for producing an output character, etc.

Given a starting code and a goal code, recode computes the most economical route through the elementary recodings, that is, the best sequence of conversions that will transform the input charset into the final charset. To speed up execution, recode looks for subsequences of conversions which are simple enough to be merged, it then dynamically creates new single steps for these mergings.

A double step is a sequence of two single steps, the output of the first being the special charset rfc1345, the input of the second single step being also rfc1345. A special machinery dynamically produces efficient, reversible, merge-able single steps out of these double steps.

The main part of recode is written in C, as are most single steps. A few single steps need to recognize sequences of multiple characters, they are often better written in flex.

10.2 Adding new charsets

It is easy for a programmer to add a new charset to recode. All it requires is making a few functions kept in a single ‘.c’ file, adjusting ‘Makefile.in’, and remaking recode.

One of the function should convert from any previous charset to the new one. Any previous charset will do, but try to select it so you will not loose too much information while converting. The other function should convert from the new charset to any older one. You do not have to select the same old charset than what you selected for the previous routine. Once again, select any charset for which you will not loose too much information while converting.

If, for any of these two functions, you have to read multiple bytes of the old charset before recognizing the character to produce, you might prefer programming it in flex in a separate ‘.l’ file. Prototype your C or flex files after one of those which exist already, so to keep the sources uniform. Besides, at make time, all ‘.l’ files are automatically merged into a single big one by the script ‘mergelex.awk’, which requires sources to follow some rules. Mimetism is a simple approach which relieves me of explaining all these rules!

Each of your source files should have its own initialization function, named module_charset, which is meant to be executed quickly, once, prior to any recoding. It should declare the name of your charsets and the single steps (or elementary recodings) you provide, by calling declare_step one or more times. Besides the charset names, declare_step expects a description of the recoding quality (see ‘recode.h’) and two functions you also provide.

The first such function has the purpose of allocating structures, preconditioning conversion tables, etc. It is also the usual way of further modifying the STEP structure. This function is executed only if and when the single step is retained in an actual recoding sequence. If you do not need such delayed initialization, merely use NULL for the function argument.

The second function executes the elementary recoding on a whole file. There are a few cases when you can spare writing this function:

Some single steps do nothing else than a pure copy of the input onto the output, in this case, you can use the predefined function file_one_to_one, while having a delayed initialization for presetting the STEP field one_to_one to the predefined value one_to_same.
Some single steps are driven by a table which recodes one character into another; if the recoding does nothing else, you can use the predefined function file_one_to_one, while having a delayed initialization for presetting the STEP field one_to_one with your table.
Some single steps are driven by a table which recodes one character into a string; if the recoding does nothing else, you can use the predefined function file_one_to_many, while having a delayed initialization for presetting the STEP field one_to_many with your table.

If you have a recoding table handy in a suitable format but do not use one of the predefined recoding functions, it is still a good idea to use a delayed initialization to save it anyway, because recode option -h will take advantage of this information when available.

Finally, edit ‘Makefile.in’ to add the source file name of your routines to the C_STEPS or L_STEPS macro definition, depending on the fact your routines is written in C or in flex. For C files only, also modify the STEPOBJS macro definition.

1 What is the purpose of this program
- 1.1 Overview of charsets
- 1.2 Contributions and bug reports
2 How to use this program
3 Reversibility issues
4 Charsets from RFC 1345
5 Charsets based on ASCII
6 Charsets based on IBM
7 Charsets based on CDC
8 Non-IBM micro-computer charsets
9 Some other charsets
10 Internal aspects
- 10.1 Overall organization
- 10.2 Adding new charsets

About This Document

This document was generated on February 15, 2023 using texi2html 5.0.

The buttons in the navigation panels have the following meaning:

Button	Name	Go to	From 1.2.3 go to
[ << ]	FastBack	Beginning of this chapter or previous chapter	1
[ < ]	Back	Previous section in reading order	1.2.2
[ Up ]	Up	Up section	1.2
[ > ]	Forward	Next section in reading order	1.2.4
[ >> ]	FastForward	Next chapter	2
[Top]	Top	Cover (top) of document
[Contents]	Contents	Table of contents
[Index]	Index	Index
[ ? ]	About	About (help)

where the Example assumes that the current position is at Subsubsection One-Two-Three of a document of the following structure:

1. Section One
- 1.1 Subsection One-One
  - ...
- 1.2 Subsection One-Two
  - 1.2.1 Subsubsection One-Two-One
  - 1.2.2 Subsubsection One-Two-Two
  - 1.2.3 Subsubsection One-Two-Three <== Current Position
  - 1.2.4 Subsubsection One-Two-Four
- 1.3 Subsection One-Three
  - ...
- 1.4 Subsection One-Four