| Clark F. Morris, Jr. 2004-12-19, 3:55 pm |
| Howard Brazee wrote:
> On 7-Dec-2004, Robert Wagner <spamblocker-robert@wagner.net> wrote:
>
>
>
>
> I usually had code like this:
> MAIN-SECTION
> sort...
> SORT-OUTPUT SECTION.
> perform sort-output-paragraph.
> GET-RID-OF-THE-DAMN-SECTIONS SECTION.
> SORT-OUTPUT-PARAGRAPH.
> ....
>
>
The real reason dates back to sorting in real memory partitions. I was
used to 22 - 28K partitions on DOS and 100K - 160K regions on MVT. This
meant that you really didn't want to sort a large file using the COBOL
SORT verb because of the limited amount of memory available to the SORT.
This lessened with the 4 - 8 megabyte regions in MVS/SP and still
further in MVS/XA. Times change and now the major reasons to have
separate sorts may well be history and the greater ease of getting a
change to even complex SORT changes (INCLUDE/OMIT, multifile etc.) into
production versus changing a program.[color=darkred]
>
[color=darkred]
|