On Wed, 21 Jan 2009 11:53:01 -0600, Tom Marchant <m42tom-ibmm...@yahoo.com> wrote:
>On Wed, 21 Jan 2009 11:06:41 -0600, John McKown wrote: > >>On Wed, 21 Jan 2009 10:38:57 -0600, Martin Kline wrote: >>> >>>You can't satisfy everyone. I suspect it was a performance choice made many >>>years ago. For whatever reason, it is what it is. Deal with it or get over it. >> >>Correct. I remember CVOLs. The original OS/360 catalog structure. Many of >>the current ills we have in this area are due to compatibility with the >>original CVOL structure. > >I spent a *lot* of time in the microfiche, reading the CVOL code. Whatever >the reason was for concatenating the generation data sets in reverse order, >I don't think it was for performance. > >-- >Tom Marchant Well, I'll defer to you. I did not spend a lot of time in the fiche. What I remember, and perhaps you can correct me, is that a CVOL was much like a PDS directory. The data portion of a record was 256 bytes in length. And it had an 8 byte hardware key. The key on the record was the highest value of the "node" within the record. The keys on a track were always in order so that a search key GE CCW would read the correct record from the track. I also remember that the key for the GnnnnVnn portion of the node was "reversed" by doing an XC with 8X'FF' so that the largest goovoo number would be at the beginning of the record, thus easier to find. I.e. to find the current goovoo, the search ke GE would simply be 8X'00'. That's what I remember. But I've been getting ECC errors on my memory for a while now. <grin>. -- John ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
