Re: Support for Unicode on IBM mainframes

From: Tony Harminc (
Date: Thu Oct 14 1999 - 17:02:03 EDT

On 14 Oct 99, at 13:08, Timothy Partridge wrote:

> How much support is there for Unicode in the world of EBCDIC? The
> UTF-EBCDIC transformation allows Unicode data to be stored on an IBM
> mainframe. Does anyone know of further support, e.g. printing? Is
> there a Double Byte Character Set which supports Unicode or is
> UTF-EBCDIC the only way of storing it?

You may have a misunderstanding here. UTF-EBCDIC is not in any sense
necessary for storing or manipulating Unicode data on any of the
three extant IBM mainframe operating systems I know (VM, MVS(OS/390),
and VSE). (I know almost nothing about TPF, the fourth, but I doubt
is is different in this regard.) There is nothing in any of these
systems to prevent Unicode data from being used in UTF-16 or UTF-8 or
whatever form you like. UTF-EBCDIC is there to provide some of the
nice properties that UTF-8 provides to the ASCII world, in
particular, compatibility with basic alpha and special characters,
reasonable treatment of typical EBCDIC-world assumptions, etc. This,
like UTF-8, is mainly to allow legacy applications and APIs to work
to a point with Unicode data.

That said, I am not aware of much "support" for Unicode data on the
mainframes. CDRA defines APIs, some of which are implemented on at
least OS/390, that will translate between Unicode and various DBCS
and SBCS codepages. And IBM's Unicode Java classes will certainly
work in the mainframe environments. Printing - I doubt it yet. The
DB2 relational database will store and manipulate Unicode data, I

Tony H.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT