Re: [extremeprogramming] About placing configuration defaults

"Steve Gordon" <[email protected]> Tue, 19 Nov 2019 09:30:02 -0700
Newsgroups gmane.comp.programming.extreme-programming
Message-ID <CAF7qo6WbSQ59tO6VNPhBqbdO1__YUfyqb=mHcB800+_Cna5dCA@mail.gmail.com>
Another factor is testing.

If you are writing software whose behavior is affected by the
configuration, individual automated tests (especially functional tests)
need to be able to change configuration for only that test.  If not, then
you are only testing the software for the default or some globally set
configuration and you can expect a lot of unexpected bugs to occur when
customers change their configuration.

So, not matter how you organize them and where you store them, there should
be a way for tests to override them.

I would also add that in complex systems some combinations of
configurations make no sense and could lead to nonsensical results.  Faced
with such complexities, I believe the most robust solution would be to
program a rules engine front-end to the configuration system, although the
most common solution is to document the configurations and blame the
customer if they create a bad set of configurations.

>

-=-=-=-=-=-=-=-=-=-=-=-
Groups.io Links: You receive all messages sent to this group.

View/Reply Online (#160232): https://groups.io/g/extremeprogramming/message/160232
Mute This Topic: https://groups.io/mt/60557252/2417047
Group Owner: [email protected]
Unsubscribe: https://groups.io/g/extremeprogramming/leave/4902963/619838065/xyzzy  [[email protected]]
-=-=-=-=-=-=-=-=-=-=-=-