cyr-conv — Automatic conversion between Cyrillic encodings

• Get the source here.

• cyr-conv is a Perl script. It can work in two modes: manual, where it converts from one specified encoding into another, and automatic, where it guesses the initial encoding by frequency analysis and converts into the encoding you specify.

• You can use this script on your own system. For automatic mode, say

  cyr-conv x

  where x stands for the encoding into which you wish to convert: k for KOI8, a for DOS866 ("alternative"), w for WIN1251, m for Macintosh, i for ISO8859-5 ("main"), l for Latin (transliterated). For manual mode, say

  cyr-conv xy

  to convert from x into y.

• In the manual mode, x cannot be l. For reverse transliteration, use another script lat-koi.

• In both modes, source text has to be given on standard input and converted text appears on standard output. (This is very convenient e.g. for usage as a mail filter.) So if you want, say, to convert file-src in any encoding into file-dest in KOI8, the command is

  cyr-conv k < file-src > file-dest 

• The success rate in determining the original encoding, provided the text is more than one-two words long, should be very high if the text is normal or IN ALL CAPITALS, but may drop if unUsuAl cAPitaLizAtIOn is used. (No, I can't convert the text to lower case before analyzing, because to do that, I need to know the encoding. :-)

• It doesn't work... You input some abracadabra and get back but some more abracadabra. What's wrong?
  Well, apart from trivial things like selecting the correct encoding in your browser, make sure that your input is what you think it is. Unexpected things can happen to 8-bit characters passing through imprudent mailers, terminal drivers, ... — even during cut-and-paste. Obviously, I can't take the blame for all that. :-)
  If it doesn't work in automatic mode, try one by one all the "from" encodings in manual mode (unless you are absolutely sure you know the source encoding). If none works, your input is not in one of these standard encodings and there is nothing we can do for you. If one does work, this means the guessing algorithm fails. In this case, please send me the original text and tell me what happened.
  Occasionally, you may see no output at all (and possibly something like "Document contains no data" from your browser). This, unless a network failure, probably indicates a bug in the converter. Again, please send me the exact input. Your cooperation will be greatly appreciated.

• Comments, suggestions, problems? Please email me!