On 29/05/2015 10:31 PM, Paul Gilmartin wrote:
On Fri, 29 May 2015 19:54:20 +0800, David Crayford wrote:
And how much has that dumb decision cost mainframe customers over the
years? Fiasco is the right word.
And IBM could have recovered, rather than compounding the fiasco at
the inception of OMVS by making OMVS ASCII based and providing
ASCII<-->EBCDIC conversion in the C RTL for Legacy data sets except
when fopen() was called with mode="*b". The kernel would have been
simpler for omitting autoconversion. (I believe Legacy I/O is not
handled by kernel.)
Legacy I/O is a handled quite well by Java. I had a good experience this
week with Java when I got it to push over 1GB of CSV data to a Linux
server in < 30 seconds over a Redis backplane.
It implicitly converted strings to ASCII (which saved me heaps of time)
and it was very fast.
And there would have been no EBCDIC obstacle to porting GNU and
other FOSS.
Fiasco ** 2.
Even yet, I wish IBM would complete the Enhanced ASCII support in the
C RTL. Significant omissions are Curses and X11; sockets is already
supported.
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
