"LS" == Laurent Siebenmann writes:

 LS> Bad news, the troika 866, 1251, KOI8 had an air of tolerable
 LS> complexity.

complexity in the number of encodings (3)? then, latin encodings have
a non-smaller complexity: latin1, latin2, latin3, latin4, latin5, etc
encodings...

after all, one can use on universal encoding utf-8 these days for
texts in any language.  i've just put an experimental utf-8 input
encoding support to CTAN:macros/latex/contrib/supported/t2/etc/utf-8/

it should work with "ordinary" TeX (Omega not required). currently,
latin and cyrillic scripts are supported, i.e. if a text file contains
accented latin + cyrillic at once, then it could be processed directly
by TeX. it is not hard to add support for other scripts (e.g. greek).

 LS> Malyshev has listed DEC as using ISO 8859-5.  VMS as well as UNIX
 LS> I presume.  Do you know another major unix flavor that uses ISO
 LS> 8859-5?

SUN (Solaris) and IBM (AIX) unixes support cyrillic in ISO 8859-5 (now
it is changing, -- other encoding options are being added).

 LS> If anyone listening could spot the codes for any of << >> < > ,,
 LS> `` '' and the number sign for DEC and followers, I would be
 LS> grateful.  The standard itself refers only to letters, I have
 LS> added it to yesterday's ftp posting.

ISO 8859-5 does not seem to contain any guillemets (either single or
double) and quotes listed above. numero sign has decimal code 240
in ISO 8859-5.

 >> cyrillic text exchanges abroad russia
 LS> Could you give an example?  I am not sure I understand.

i meant primarily software (mail readers, browsers, etc) which
supported only standard ISO encodings but ignored encodings such as
koi8-r or ms codepages. it was mainly developped not in russia.

Best regards, -- Vladimir.