Re: What can I do with an unloaded partitioned data set on Linux?
On Thu, 01 May 2008 19:44:50 -0400, Douglas A. Tutty wrote:
> On Thu, May 01, 2008 at 03:50:09PM +0000, Hendrik Boom wrote:
>> I am the proud owner of several unloaded partitioned data sets, dating
>> back to my days on OS/360 and OS/370.
>> Does anyone have any code I could use to decode these? I don't mean the
>> EBCDIC-to-ASCII (or to UTF-8) conversion -- I mean the weird file format.
>> Or documentation about this file format?
>> Or will I have to reverse-engineer it myself? Not that that's likely to
>> be an insurmountable task. I just suspect that somewhere, someone has
>> already done it.
> I remember that there are some old-timers on here who have mentioned in
> passing their history with mainframes, so hopefully you get an answer.
> I didn't start __administrating__ computers until OS/2 then Linux,
> although I am sure I _used_ a mainframe (probably doing that by sending
> this mail...).
> I must admit that until I read "days on OS/360...", I didn't know what
> you were talking about.
> On what medium are the data sets? Do you have a box that can physically
> read them? IIRC, there is a OS/360 emulator in the repository (I forget
> the name); I wonder if it can read it.
Ther were text files, source code, and object code. Record formats are
FB. VB, and VBS. Once I have liberated the text files and source code
from their ancient file format and EBCDIC I plan to mechanically translate
them into something that will run in the here and now. I don't
think I'll have any use for the object code.
As to what medium they're in, well, that's a long story. They were
copied onto a magnetic taps long long ago. That's when they changed from
being partitioned data sets into unloaded partitioned data sets.I still
have that tape somewhere, and it might even be readable. However, when it
looked as if mag tapes might be going obsolete, I copied the tapes, block
by block into an ordinary Unix file, together with block lengths, error
indications, and redundant synchronizing marks in case something went
wrong. These files were then chopped into pieces and stored on Amiga
floppy disks. Those were many years later, successfully transferred to
Linux (now that a reasonable OS is available to ordinary mortals, and the
Unix file that encodes the mag tape has been reconstructed. I've thrown
together a hex & EBCDIC display program, and have had tantalizing glimpses
into what I had been doing over 30 years ago.
> For documentation, have you tried either wikipedia or the IBM website?
> Its advanced search page is quite powerful.
There are too many new uses of the word "partitioned". The pages that
are on topic describe the IBM mainframe utilities for handling them
(which I don't have available, nor the mainframs to run them), and I've
found nothing (yet) that described the file format. Thanks for the
recommendation, though. There ought to be information on the IBM site
somewhere, but I haven't been able to find it.
It brings me back though. It's been a long time since I've seen some of
> I wish I had access to a mainframe (and the team to run it). I think
> I'd learn a lot of good, conservative, skills. Perhaps having people
> like you on DU is part of what makes Debian good and reasonably
> Good luck.