On Fri, 25 Jun 2021 12:19:10 -0400, Matt Hogstrom wrote: 
>
>    ...  I’d love to share my more esoteric issues like unable to allocate a 
> dataset because of a hashing problem with a VTOC IX caused by a batch job 
> creating crappy data set names.  
>
It's easy to create "crappy data set names" by calling dynalloc from an 
assembler program.
Either:
o Allocation should be modified to prohibit creating such names, even though 
that
  would probably break existing customer code.  Or:
o VTOC indexing should be modified to deal with such names, even though that
  would break existing VTOC indices.

The OS should not rely on:
o The difficulty of creating such names with an 029 keypunch.
  - Consider multi-punch.
  - Consider hexadecimal strings.
o Paper documentation's calling such names invalid, despite lack of enforcement.
o IBM experts' calling such names invalid in fora such as this, despite
  lack of enforcement.

The Black Team had the right idea.  In the day they were respected for that.

-- gil

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to