On Fri, 20 May 2005 13:51:55 -0500, Todd Burch <[EMAIL PROTECTED]> wrote:
>... >IBM, by implementing this, will not have "broken" anything. Your s/360 >program that is runing today from JCL is expecting no more than 100 bytes. >Fine. Why would you pass 64K bytes to a program that is expecting a max of >100? If you do, you've broken it, not IBM. >... Technically true but effectively false. It doesn't much matter who was right when production applications die. I assume a JCL error is produced if you create a parm larger than 100 bytes. (Either that or the data is just struncated.) Change the parm behavior and instead of a JCL error you get abends or incorrect execution of some programs. A change in program behavior with no change to the program (whose source may have been lost 13 years ago). Late night phone calls to support people who've never heard of the dying programs. Late night phone calls to the system programmers supporting the recently upgraded MVS. Lots of finger pointing. Everything is someone else's problem. All parts are Broken As Designed because nothing prevented the problem. All because of a finger check no longer causes a JCL error. IBM should be praised for approaching this cautiously, and is looking for a solution with the least negative effect. Pat O'Keefe ---------------------------------------------------------------------- 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
