Re: Testing relational databases

From: Phlip <>
Date: Mon, 10 Jul 2006 16:03:06 GMT
Message-ID: <_Susg.129129$>

Tony D wrote:

>S Perryman wrote:

>> For example, Praxis CS (safety-critical s/w developers) showed on one
>> project
>> quite significant ROI on defect prevention and detection by using
>> proof-based
>> techniques, compared to the usual test approaches (unit testing etc) .

That compares proof-first to test-last.

> Another question is, who designs the tests ?

So who designs the proofs?

> Programmers are notorious
> for testing to "prove" that their software works.

Absolutely. Without any tests or proofs, they tend to debug the code, then manually run a few "happy cases" where they know the code will work. This is nothing to fault them for; they are not professional manual testers, so they should indeed spend their time developing.

However, if they only write lines of code in response to failing tests, then the tests will automatically defend not just the happy cases, but all the cases the team found were important. This, in turn, frees the developer to focus only on the one feature they are currently adding.

> Testers (when
> available) are notorious for testing to "prove" that the same software
> fails. Who would you rather have design your tests ?

Teams using TDD still have a role for pro testers. They typically work with the "business analyst" or "customer liaison" roles, to make sure that all requirements are specified as tests.

(Nice to see everyone's actually discussing the cited website here!)

  Phlip <-- NOT a blog!!! 
Received on Mon Jul 10 2006 - 18:03:06 CEST

Original text of this message