On Thu, 23 Aug 2007 14:19:05 -0500, John McKown wrote: >I'm reading the book again. In "z/OS: DFSMS Using Data Sets" >(SC26-7410-06), in section 3.1.3.2.4 Null segments, it states: > ><quote> >A 1 in bit position 0 of the SDW indicates a null segment. A null >segment means that there are no more segments in the block. Bits 1-7 of >the SDW and the remainder of the block must be binary zeros. A null >segment is not an end-of-logical-record delimiter. (You do not have to >be concerned about null segments unless you have created a data set >using null segments.) ></quote> > <snip> > >Anybody know why IBM came up with this? Does anything actually use it?
I wonder if this was intended to pad very short blocks on old tape devices, where a block shorter than 18 bytes might be taken as noise. I used to wonder if QSAM wrote one of these null segments automatically when a block would have been shorter than 18 bytes, and ignored the null segment when reading. Maybe the null segment was invented with that purpose in mind, and was never implemented by QSAM. All pure speculation on my part. Bill ---------------------------------------------------------------------- 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
