On 2017-02-06, at 07:22, R.S. wrote:
>
> It's worth to mention that performance of reverse processing dataset on real
> tape is worse than horrible.
> And IMHO it's not argument for using virtual tapes but for not using datasets
> on tape for applications. Tapes are for backup and ML2.
>
I had understood that old-fashioned SORT with SORTWK on tape employed
read backwards very effectively. Perhaps the SORT expert will jump in:
https://en.wikipedia.org/wiki/Oscillating_merge_sort
It was spectacular to watch. Are the data un-reversed by hardware,
firmware, or software?
On 2017-02-06, at 10:24, Binyamin Dissen wrote:
> Also, reading a PDS member by member requires BPAM (which is really BSAM).
>
Which was our motivation, when I was in a small group supporting a FOSS
compiler and runtime library, for implementing BPAM and BSAM but not
QSAM. Once we had the BPAM interface, most of the interface could likewise
be used for BSAM.
Ironically, not burdened by the extreme storage constraints of OS/360, our
compiler eschewed most of BPAM. At a COPY-type operation we simply opened
another DCB, processed the member to the end, then popped back to the parent
member; no NOTE or POINT involved. Worked great on MVS; not at all on
CMS "OS Emulation".
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
