Changes between Version 1 and Version 2 of WhitePaper

Timestamp:

03/02/09 11:36:30 (9 months ago)

Author:

mbligh (IP: 216.239.45.4)

Comment:

--

WhitePaper

@@ -28 +28 @@
 We want bug identification, investigation, and fixing to be done earlier in the cycle. This allows multiple debugging efforts to be done in parallel, without affecting others, as well as many other advantages, such as the problem still being fresh in the developers mind, and not interacting with other later changes. This means running tests on multiple codebases (development trees), with high frequency - how can we scale to this? Fully automated testing. Machines are cheap, people are expensive - this is the reality of the modern age. For Linux, the problem is compounded by the staggering diversity of hardware and kernel configurations that we support.
 
-Moreover, a test system is not just about simple functional tests; we should also identify performance regressions and trends, adding statistical analysis. A broad spectrum of tests are necessary -- boot testing, regression, function, performance, and stress testing; from disk intensive to compute intensive to network intensive loads. A fully automated test harness also empowers other other techniques that are impractical when testing manually, in order to make debugging and problem identification easier, e.g. automated binary chop search amongst thousands of patches to weed out dysfunctional changes.
+Moreover, a test system is not just about simple functional tests; we should also identify performance regressions and trends, adding statistical analysis. A broad spectrum of tests are necessary -- boot testing, regression, function, performance, and stress testing; from disk intensive to compute intensive to network intensive loads. A fully automated test harness also empowers other techniques that are impractical when testing manually, in order to make debugging and problem identification easier, e.g. automated binary chop search amongst thousands of patches to weed out dysfunctional changes.
 
 It's critical that when operating in an open community, we can share and compare test results - that necessitates consistency of results formats. The easiest way to acheive this is to share one common test harness.