2009-03-06 Hal Merritt <hmerr...@jackhenry.com> wrote:
> IMHO: exits as a subspecies are evil critters. They become an ongoing 
> maintenance challenge and tend to attract unwelcome attention from auditors. 
> Exits are hard to write, hard to stress test, and introduce a level of risk. 
> You need extraordinary measures in place to protect the code.

This is kind of amusing... Twenty years or so ago, the big controversy
was over IBM removing the source code to its products. IBM's argument
then was that too many shops were modifying their systems, and thus
delaying the move to newer OS versions, and that exits would be the
stable way for customers to make the systems behave the way they
wanted. The big shops at the time argued that IBM provided far too few
exits, and with too low functionality to duplicate what they did with
mods.

So now that there are lots of exit points, using them has become the
new evil? What's next on the list - config files? Parmlib options?
Local naming conventions that don't match the defaults?

[...]

> I once worked in an exit happy shop. Getting the exits updated and tested 
> tended to be the single biggest bottleneck in rolling out new operating 
> system levels.

Perhaps so. But SHARE showed fairly clearly back in the 1980s that the
most heavily modified ("modification happy"?) shops were also those at
the newest OS releases. In some cases updating the exits is worth the
resources.

> Of course, if you have a compelling business/technical need, then lock and 
> load.

Using exits, just like modifying source code, or choosing build vs buy
for an application, is a business decision, that has to be taken with
awareness of all the costs, risks, and benefits. If you have a
business requirement that is best and most cost effectively addressed
by using an exit, then do it. If you have programmers running around
writing and installing exits on production systems for fun, then you
have a management problem.

Tony H.

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

Reply via email to