A New Title Thanks to Jake, I have a new title for my weblog.  6:16:47 PM    

DG Forum Got this from an email list:

There's a free service designed for just this problem: Take It Offline, now called QuickTopic. Go to http://www.quicktopic.com, start a topic about this, and then post the topic URL back (once!) to baylisa-jobs so that folks who are interested can go there and discuss it. QT is free, the folks who run it are great, they don't do anything at all with your info (you need to register to use it), and you can get topic updates via email instead of having to check it on the web. cheers,
Strata

Might be useful for blogs where the comment interface is not sufficient . . .  10:44:54 AM    

The Testing End-Game

Prioritize and order the testing based on what has changed and what you have not covered yet.

We're about to get the candidate for the 2nd beta release, and soon after that we'll get the FCS candidate.  How should we plan our testing?

(assumption #1 - we've already run smoke and sanity tests for the build - shallow and somewhat broad tests which cover basic functionality.  These should be automated)

First, I look at what is new (yes, new features during beta - this is the reality of the market), and what bugs have been fixed.  New tests that sanity-check these changes should be run first, followed by a small set of old tests to make sure stuff did not break.

I also look at what the field has been led to expect from this particular release.  Tests which cover these functional areas are bumped up in the list.

Next, we set up the long-running tests on dedicated testbeds.  The more automation here, the better, even if you have to "check in" on the health of the system every so often.

I then look at mixing up the in-depth tests of new features, and the tests not yet run (skipped in prior test cycles, and lower-priority bug fixes which have not yet been verified).  Each tester can choose the exact test cases based on their own discretion.

This is enough to start the test cycle.  I keep an eye on the bugs found (both in testing and in the field), and keep asking questions to "shape" the testing effort.

  10:32:49 AM