Like old soldiers this topic never dies. It does not even fade away
monotonically.
Consult the archives. The last time we went round (and round) about it
o ROUND, while innocuous, no longer improved the performance of sequential
datasets; but
o ROUND did still measurably improve the performance of PDSs, though not of
PDSEs.
At that time there was a sort of commitment on the part of one of the senior
IBM posters here that he would, when he found the time, remedy this situation,
removing the exception for PDSs; and if he has done so I am sure he will chime
in to say that he has.
The hardware issue, when DASD did indeed spin, was that allocations in TRK
units set the device mask foir an interrupt at each end-of-track, for a run
through a long block of code IOCS to ensure that the next track was a part of
in the data set being processed, while allocations in CYL units set the device
mask for this interrupt only at each end-of-cylinder, i.e., for much less
frequent runs throug this code.
Old men and women in this business thus found themselves repeating the litany
"comma, comma, comma, round" over and over again to the young, who had
unwonted difficulty with positional parameters and vacuous-comma coding.
It would be nice to know that this issue has now been resolved definitively,
but in this OCO era only IBM can tell us that it has been.
John Gilmore Ashland, MA 01721-1817 USA
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: GET IBM-MAIN INFO
Search the archives at http://bama.ua.edu/archives/ibm-main.html
