EconomÃa de Western Sahara

The Extended Binary Coded Decimal Interchange Code (EBCDIC) is a mode of coding of the characters on 8 bits created by IBM at the time of the perforated cards. There exist at least 6 well documented different versions (and of many alternatives sometimes created by competitors of IBM), incompatible between them. This mode of coding was criticized for this reason, but also because certain characters of punctuation are not available in certain versions. These disparities were sometimes interpreted like a means for IBM of preserving its captive customers.

EBCDIC is still used in systems AS/450 from IBM like on the Mainframe S under MVS, VM or DOS/VSE.

History

The provision apparently “ étrange ” of coding EBCDIC is historical and comes from the former perforated card and tape readers, which were to be able to be read and bored manually by human by using the mnemotechnical decimal system (and it was itself derived from the old telegraphic character set on 5 bits).

The bits of strong weight appeared after and made it possible to code in an additional column of perforation the distinctions between figures and letters, or tiny and capital. The play was conceived so that to each character also at least a hole corresponds, in order to allow the synchronization of the tape readers perforated, this is why column 0 was at the unutilised origin for the letters, the figures having an additional dedicated hole corresponding to a bit of strong weight.

Play EBCDIC in the beginning was well a play on 7 bits, the eighth being then only used on the perforated cards to indicate that the column of perforation was well occupied by a coded character (this is why principal codes EBCDIC occupy second half of the table).

In the same way, the punctuations were also coded in a visual and mnemotechnical way, and could easily be distinguished from the figures and letters by the fact that they did not have any hole in the binary positions corresponding to the marks of figures or letters.

The absence of any perforation in a column of perforations was interpreted like stuffing not meaning, dependant on technology employed (this code at the beginning of table corresponds today to the control character “NULL” C0). In the same way, one could correct a manual error of perforation by perforating all the positions of a column, and thus the last position is also a stuffing character not meaning (this code at the end of the table corresponds today to the control character C1 “APC”, function rather reserved for the control character C0 “LED” of the ISO 646 and the ASCII but of which the use is more ambiguous).

The EBCDIC is thus the fruit of a historical evolution good older (developed long time by IBM in continuity with the old telegraphic systems) that ASCII (more practical to handle in the programs) which replaced it then practically everywhere and then gave place to a standardization in the ISO 646.

The bands and perforated cards however continued to be used with the EBCDIC until worms the years 1990, for example in Turkey for the military exchanges of information whereas the compatibility issues of the systems of tape recording or the networks were still far from being all solved.

The success of the standards of communication between heterogeneous systems (in particular TCP/IP and Internet, as well as the systems of Cryptography for the protected and less expensive transmission via public networks) put an end to the interest EBCDIC for all new development, and the many national versions of the EBCDIC also less and less have interest since the appearance of UTF-EBCDIC.

Example of coding

This table represents the coding of an alternative of EBCDIC compatible with the ISO 8859-1. The coded characters of 0x00 with 0x3F as 0xFF are control characters, 0x40 is space, 0x41 is the nonbreaking space. The character coded in 0x54 is the visible indent of Césure only at the end of the line.

This alternative is compatible, byte by byte, with the UTF-EBCDIC which uses the positions alternatives coloured in green to code the Unicode characters (out of the ASCII and the check codes) in the form of sequences of bytes.

Each national or international alternative of the EBCDIC is coded according to this table (however permutations are possible between two CCSID for the same characters). It should be noted that all the alternatives of the EBCDIC do not code the basic Latin small letters, contrary to the character sets based on the ISO 646 or ISO 8859.

In addition certain national alternatives of the ISO 646 (or other character sets compatible with this standard) contain positions additional, invariant alternatives in the character sets based on the EBCDIC.

Lastly, the positions 0x21 and 0x22 of the EBCDIC are variantes - the majority code the point of exclamation and the quotation mark English (double quote) - contrary to the corresponding characters of the national alternatives of the ISO 646.

It should be noted that on systems EBCDIC, the jump of line is normally coded with the control character C1 “NEL” (U+0085 in Unicode, or 0x25 in all standard alternatives EBCDIC) and not with the control characters C0 “CR” and/or “LF” of the ISO 646 and ASCII (U+000D and/or U+000A, i.e. 0x0D and/or 0x15 in EBCDIC, where these orders have a function well defined and single of management of position of the cursor on a terminal, or make it possible to distinguish the jumps of lines forced in the same paragraph, or made it possible to post a line in overprinting to produce bold characters, underlined, or accentuated additional).

Transcribing of the ISO 8859-1 towards the EBCDIC

The following table makes it possible to transcribe the ASCII (characters Unicode U+0000 with U+007F) and the command set C1 (characters Unicode U+0080 with U+009F) in EBCDIC. The characters of the extensions ISO 8859 (in greens) are indicated here in the order compatible with the UTF-EBCDIC, but this assignment does not correspond to a particular national alternative of the EBCDIC.

This table is the reverse of the preceding table and is compatible byte by byte with the second phase (of permutation of the values of bytes) of UTF-EBCDIC.

Units, these two tables make it possible to easily adapt any compatible character set ISO 646 or ISO 8859 for treatments on systems EBCDIC standards. On the other hand, to adapt to a precise version of a play EBCDIC (and to make recognize the nonASCII characters exactly), it will be necessary to substitute the positions marked in green in the two tables, according to their real assignment in the national alternatives of plays EBCDIC correspondents, and even sometimes to substitute the Latin small letters replaced in certain obsolete versions of certain plays EBCDIC (play EBCDIC Japanese or Cyrillic, except Russian plays EBCDIC which unify with same code EBCDIC certain Cyrillic and Latin letters like has, of identical appearances in the two writings).

The detail of the assignments of these positions (marked here in green in the two tables) in the national alternatives of the EBCDIC is referred on the IBM site mentioned in the external bonds. In particular, the alternative still most used is the EBCDIC CCSID 500 (version international of the Latin play n°1 to which the play is very close to that of the ISO 8859-1, but with some differences, and differently sorted tables presented here).

See too

Internal bond

ISO 646, ASCII, ISO 8859
EBCDIC-codepages with Latin-1-charset
EBCDIC-codepage 297 (France)
UTF-EBCDIC

External bond

Tables EBCDIC on the site IBM
http://home.arcor.de/wzwz.de/wiki/ebcdic/cc_fr.htm // EBCDIC-codepages with Latin-1-charset (Javascript)
http://home.arcor.de/wzwz.de/wiki/ebcdic/aa70_all_pages.zip // ZIPped version
compared http://homepages.cwi.nl/~dik/english/codes/stand.html character sets 7 or 8 bits.

Random links: