__/ [ Tim Smith ] on Monday 17 July 2006 08:27 \__
> In article <1631083.AYT2nUj6by@xxxxxxxxxxxxxxx>, Roy Schestowitz quoted
> someone:
>>| In their book "Software Engineering: Theory and Practice" Pfleeger and |
>>Atlee quote the results of the research on software reliability stating |
>>that 'normal' testing methods uncovered only one out of five bugs in | the
>>software code under investigation while the peer review process | found the
>>remaining four. | | Taking this research at face value implies that the
>>peer review | process results in enormously higher quality of software code
>>as | opposed to proprietary software developed behind the closed doors.
>
> No, it doesn't imply that at all. (Hint: proprietary code developed behind
> closed doors can be peer reviewed.
Not at quite the same level. You see, testers are paid to 'toy' with the
program (boring, exhaustive regression tests) whereas users (of which there
is *more*) /use/ it for their work (i.e. no employment fees) and can submit
patches. As an example, you are able to report a bug to your newsreader for
re-wrapping the text badly, unlike other news reader under the same
circumstance (probably opting not to re-wrap).
|
|
