On 2017-03-09, at 00:18, Vernooij, Kees (ITOPT1) - KLM wrote:
>
> How many DLM's do you use per year? How many department codes, projectnames,
> daily/weekly variations and sequence numbers do you have to code into 8
> chars? All problems are serious, but some problems are more serious than
> others. Come on, the 18 chars DLM limit will be at the bottom of my list.
>
There's a good argument here for using UNIX files where possible
instead of Legacy data sets: the former are blessedly free of
the 44(8) limitation. (Yes, there is a 1023(255) limitation;
lots more elbow room for department codes, projectnames, daily/weekly
variations and sequence numbers.)
For comparison, I tried the analog of DLM= on Linux and MacOS.
Both were happy with 380-character strings. ("No one should
ever need more ...") I didn't probe for the ultimate limit.
Both allowed broken/continued strings on the defining occurrence.
MacOS allowed broken/continued at the terminating occurrence;
Linux did not.
Why should z/OS strive for less than other OSes have supported
for decades?
-- gil
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN
