Usenet 1994 January/usenetsourcesnewsgroupsinfomagicjanuary1994.iso/sources/std_unix/volume.30/text0002.txt

Text File | 1993-03-11 | 1.6 KB | 41 lines

Submitted-by: [e-mail] (Chuck Karish)

In article <[e-mail]> [e-mail]
(John R. Levine) writes:
>The problems that people have pointed out with the POSIX tests demonstrate
>that it is not appropriate to standardize test suites, particularly not
>test suites that haven't themselves been tested by a significant amount of
>use.

We haven't been discussing standardizing test suites but rather
test methods (lists of assertions).

>I'm all in favor of test suites.  I think there should be lots of them,
>since they make the implementor's design so much easier, and raise the
>user's confidence that a product that passed the tests probably works.
>But test suites make poor standards.

That's why the current practice is to standardize just the
test methods, largely a specification of the assertions
that a test suite has to test.

Once the standards process has determined that the test
methods properly express the base specification, the test
suite implementors have a specification for the tests and
the users of the tests have a way to examine the tests for
correct behavior.

This raises the hope that implementors will be able to
produce products that faithfully reflect the base standards,
without ever having to work around test suite bugs.  This
requires, of course, that the certifying authorities, the
test suite implementors, and the standards committees all
be responsive to problem reports.

--
Chuck Karish          [e-mail]
[phone] x117   [e-mail]

Volume-Number: Volume 30, Number 3