From: patrick-may@... (Patrick May)
Date: 2001-12-04T06:41:49+09:00
Subject: [ruby-talk:27376] Re: Quantitative Unit-testing? (Re: John Roth dolt (  Re: A challenge  to   proponents of Unit Testing. )

> Right.  It's the generally improper "touch and feel" that
> characterizes the XP/Alliance way of doing things.  Everything is
> merged and "squashed" by them into one, because to them everything is
> the same, and there is no objective truth.  So we can just do
> everything at the same time without sequencing  and see how things
> evolve.  Hackery pure and simple.

This seems to be a bit of a distortion.

When you are regularly unit testing, you use the tests to describe the
known and expected behaviour of the system and the components.  Each
project has different levels of expectations -- little might be
expected of a text munching script, while a mission-critical business
app expected to do alot more.

The depth of the tests determines the depth of the code that needs to
be written.  It isn't hackery to solve a simple problem appropriately.

Getting back to Panu's concern:

Unit tests in XP are for internal use, to encode expectations into
something that can be regularly tested.  Documenting those
expectations is another issue, though well written tests can help.