Re: Test first as specification
- From: "paul campbell" <paulc@.o.b.je.ct.vis.sion.co.uk>
- Date: Mon, 01 Aug 2005 20:15:23 +0100
On Sun, 31 Jul 2005 15:10:49 GMT, Roger L. Cauvin <roger@xxxxxxxxxxxx> wrote:
"Robert C. Martin" <unclebob@xxxxxxxxxxxxxxxx> wrote in message news:vf0dd1tp1evjjmm3b5a2t16g3efuck2d0e@xxxxxxxxxx
You should specify the requirements in terms of automated acceptance tests (e.g. take a look at fitnesse.org). The developers (in india) should write the unit tests, and make all the tests pass.
Just a bit of an off-topic quibble. While I cannot disagree with the notion
of acceptance tests as specifications, sometimes they are not, and cannot
be, exactly equivalent to requirements. For details, see:
http://cauvin.blogspot.com/2005/07/should-all-requirements-be-testable.html
Just to add another sub discussion to this ....
My own personal favourite counter-example for specfication by tests is a programming language.
However this is an extreme example and most business IT systems are more linear in nature and thus ameanable to specification by test.
Paul C. .
- Follow-Ups:
- Re: Test first as specification
- From: Steven Woody
- Re: Test first as specification
- From: Nick Malik [Microsoft]
- Re: Test first as specification
- From: Laurent Bossavit
- Re: Test first as specification
- References:
- Re: Test first as specification
- From: Roger L. Cauvin
- Re: Test first as specification
- Prev by Date: Re: Newbie - Design Document
- Next by Date: Re: Test first as specification
- Previous by thread: Re: Test first as specification
- Next by thread: Re: Test first as specification
- Index(es):