[TenTec] Software Testing

[Richard Detweiler]

I noticed an artical a while back,

One vendor sent a note to the QA department that read like this,

==============
Please evaluate this code.  There are 25 defects.  Please identify them.
==============

They were releasing a set of code that had x number of lines and came up 
with that number of bugs based on the history of bugs identified in previous 
products.

The QA departement was then challenged to find them not knowing the 
management didn't know what the bugs were....

Best 73's
Rich
K5SF




>From: "Mark Erbaugh" <mark at microenh.com>
>Reply-To: tentec at contesting.com
>To: <tentec at contesting.com>
>Subject: Re: [TenTec] Software Testing
>Date: Mon, 20 Sep 2004 08:30:30 -0400
>
> > Has the art of software testing made any progress
> > since Dijkstra's conjecture (almost 40 years ago)
> > that it can only prove the presence of errors,
> > but not their absence?
>
>Software testing still can't (except in very trivial cases) prove the
>absence of errors, but there have been vast improvements in recent years.
>Techniques such as test driven development all the developers to thoroughly
>test (again, the thoroughness of the test is up to the skill of the
>developer) the software under development and more importantly, verify that
>code written to add a new feature or fix a problem doesn't break a
>previously working feature.
>
>When a new version is ready for release to end users, automated test
>programs can simulate the testing of many users and can test the code more
>thoroughly.  For example, if it takes a complex series of steps to test a
>particular feature of a program, users are less likely to test all the
>possibilities at that stage than they are to test the features that are 
>more
>easily accessible. An automated test program doesn't care how complex a 
>step
>is.
>
>73,
>Mark
>
>_______________________________________________
>TenTec mailing list
>TenTec at contesting.com
>http://lists.contesting.com/mailman/listinfo/tentec