-------------------------<snip>--------------------------

 I originally wanted to start right out with a default of NO
in 1.8, but Bob Rogers talked me out of it, so I instead settled
for stern language in the description of the parameter in
Initialization and Tuning Reference. Even there, the writer of the manual (in China) wanted to water down "IBM recommends" to "IBM suggests" (because "IBM Style", whatever that is, discourages
the use of "IBM recommends").

 I still wanted to move to a default of NO quickly, as the
slow progress in eliminating user key CSA convinced me that
the sledgehammer approach was required. Hopefully, having to explain to security auditors why use of product X requires
overriding IBM's recommendation, default, and HealthChecker
will be a sufficiently large sledgehammer.
--------------------------<unsnip>-------------------------
Jim, maybe you could start with a tack-hammer; why can't SERVERPACK be shipped with an initial value of "NO"? Eventually, you should take the option away entirely and completely disallow the feature entirely. Anyone that's bright enough to use CSA storage should also be bright enough to assign a non-user key as well. Else they shouldn't be "mucking around" there to begin with. Cross-memory services are well enough developed that they should be embraced by developers.

----------------------------------------------------------------------
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

Reply via email to