Testing Questions

[David Cantrell]

On Mon, May 14, 2007 at 06:03:03AM -0700, Ovid wrote:
> --- Pete Sergeant <pete at clueball.com> wrote:
> > * To what degree do you try and separate programmer-generated unit
> > tests, and tests generated by a test engineer?

It would be nice to be able to afford a seperate person for that, but
it's not practical.  And of course you quite often find that you can't
really write all the tests seperately from the code, which makes a
test engineer an even more expensive waste of time, at least as far as
regression tests go.

> >                                                  Do you make any
> > distinction between regression tests and acceptance tests?

Yes.  Regression tests make sure that when you change one small corner
of your code (eg, the line that connects to PostgreSQL instead of
SQLite) nothing else breaks.  Acceptance tests make sure that the users
are given what they've asked for.

> Some object to [Test::Pod::Coverage] due to the assumptions
> it makes about documentation, but it's really nice to have the
> documentation format standardized and enforced.

It seems to cope OK with somewhat unusual formats anyway.  At least,
it's not broken with anything I've thrown at it yet.

-- 
David Cantrell | Cake Smuggler Extraordinaire

  Sobol's Law of Telecom Utilities:
    Telcos are malicious; cablecos are simply clueless.