On 8/2/2010 12:57 PM, Rick Fochtman wrote:
-----------------------------------<snip>-----------------------------------
On Mon, 2 Aug 2010 09:24:53 -0500, Rick Fochtman wrote:
---------------------------------------<snip>--------------------------------------
Because you didn't use system services to insulate yourself from
changes.
-----------------------------------<unsnip>------------------------------------
Most of those geometry-related "System Services" didn't exist! :-)
Not even by the advent of the 3390, the "last ever" conversion?
--------------------------------------<unsnip>---------------------------------------
That's right........
Rick
Actually, the biggest problem I had was not code that couldn't
handle the different geometries (though APT did have this problem).
Most code would use the tables provided by the OS to get track size,
cylinder size, etc.
The biggest problem was moving datasets to the new devices.
At the time some datasets still had CCHHR or TTR pointers in them
which would obviously be wrong of they were just copied.
IEBCOPY would correct TTR pointers in the directory, but if they
existed in a member, or a non PDS dataset you had problems.
SAS was also a big problem. At the time SAS created RECFM=U
datasets with a BLKSIZE that depended on the type of disk. If
you just copied the dataset with IEBGENER or some other non SAS
utility SAS wouldn't recognize it as a SAS dataset. So all
SAS dataset had to be identified and copied using a SAS utility.
--
Richard
----------------------------------------------------------------------
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
