On Mon, 28 Nov 2016 19:40:11 +0800, David Crayford wrote: > >... I've ported many >open source projects to z/OS over the >years and EBCDIC is the killer. > >> I do not whinge about EBCDIC. EBCDIC is part of the austere and >> idiosyncratic nature of traditional IBM record-based business systems. >> Errr...
>EBCDIC is no beauty! She's an ugly hag! Medusa comes to mind with every >snakes head being a different EBCDIC code page... > Indeed. This strikes me as the antithesis of "austere". >... I recently received an >e-mail from a sysprog at a Danish bank >asking me if it was possible to enable my z/OS Lua port to handle a >Scandinavian code page. I hacked up a solution but running iconv to >convert source modules added significant overhead >to a fast language. > The situation may be comfortable for someone such as Ed G., who I suspect works in one location where all the users have terminals compatible with the same code page. It's brutal for the ISV who must court customers in varied locales. ISPF does surprisingly well with UNIX files tagged UTF-8. It displays characters in the terminal's particular code page and considers others "nondisplayable". (But I still need to submit an SR on how badly "FIND P'.'" performs on such characters.) I haven't tried UTF-8 with a MBCS 3270. Are there such? I don't believe I have access to one. UTF-8 is the modal text representation for desktop systems and WWW. There ought to be little need to use other character sets. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
