Good exposition.

It turns out it is incorrect to say "z/OS allows '$' in dataset names." It
is actually a case of "z/OS allows x'5B' in dataset names."

Your glyphs may vary.

Charles

-----Original Message-----
From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf
Of Tony Harminc
Sent: Friday, May 11, 2012 12:44 PM
To: [email protected]
Subject: Re: Codepages and locales

On 11 May 2012 08:27, Phil Smith <[email protected]> wrote:
> Paul Gilmartin wrote:
>>"Different" from what.  Everything is different from something else.
>>Perhaps different from IBM-1047?
>
> Doh, yeah, sorry. "It seemed obvious at the time..." - but of course
you're right, it wasn't.

It's perhaps even a little less well defined than that. Not everything on
z/OS uses a system-defined codepage, "different" or not. Much of the
traditional MVS infrastructure knows nothing of code pages at all.
RACF is a good example; you can use the characters $, #, @ in passwords, but
there is no codepage support to this, so the mappings are always $=5B, #=7B,
@=7C, right off the green card. These are CP
037 and 1047 mappings, but not those for e.g. UK CP 285, which has the
Sterling sign (£) at 5B, and the dollar sign at 4A. So typical UK users of
3270s and printers and such have always thought of £ as a valid character in
passwords (and assembler language identifiers, dataset names, etc. etc.) No
matter until you start to exchange data with other systems. So in the case
of our products, where we transfer passwords between systems in various
contexts, if e.g. a UK Windows user enters a £ (ASCII, and in this case also
UNICODE A3), we must translate it to 5B in EBCDIC, even though there is no
notion of CP 285 being in effect in RACF, and no reliable place we can query
to find out that we should do this. If the UK company has users in Brazil or
the USA, we must translate their $ to 5B instead, and the £ to B1, where it
will probably correctly be treated as an error.

There's a lot of infrastructure that just ain't there yet.

Tony H.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions, send email
to [email protected] with the message: INFO IBM-MAIN

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to