Skip to main content
16 events
when toggle format what by license comment
Apr 16, 2024 at 14:15 comment added Edwin Buck @dodgethesteamroller When considering testing, I don't see context-driven testing acknowledging even the basic management of testing or the basic debugging of test failures. The ice-cream cone pattern typically yields a "system failed" result, which then needs post-run debugging to discover what failed, or basically "how we started testing in the 1970's" Yes, it is useful in some scenarios (especially when an independent component acts differently than expected), but a sufficient bank of unit tests was the approach to fix "it crashed again, but I can't find out why".
Aug 28, 2020 at 18:16 history edited Jeff Bowman CC BY-SA 4.0
Better summary up top, plus case study down below, plus link to test doubles.
May 26, 2018 at 9:35 comment added Nicholas Shanks Also, pyramids are a lot bigger than ice-cream cones. Bigger == more tests == better ;-)
Mar 16, 2017 at 19:52 vote accept mindplay.dk
Jun 16, 2016 at 0:28 history edited Jeff Bowman CC BY-SA 3.0
Add the most damning aspect of integration tests.
Nov 12, 2015 at 18:48 comment added Jeff Bowman @dodgethesteamroller Whether something is contrarian IS a matter of opinion or perspective; in a few years it may be the dominant line of thinking, for all I know. Don't read into my punctuation; my point is specifically not to judge it and merely to say that that if you think it's contrarian, and expound on it in 2 (now 3) comments, it's time for it to be an answer.
Nov 12, 2015 at 18:43 comment added dodgethesteamroller @JeffBowman Point taken on the comment vs. answer, although I was commenting specifically on the Testing Pyramid, which some people do take as gospel (even if you don't). But why do you put "contrarian school" in quotes as if it's a matter of opinion or a joke? Cem Kaner and like-minded engineers and academics are highly regarded members of the software testing community.
Nov 6, 2015 at 19:38 comment added rszalski Highly recommended presentation on the hierarchy and ratio of unittests to integration tests: vimeo.com/80533536 Very well explained.
Nov 4, 2015 at 17:01 history edited Jeff Bowman CC BY-SA 3.0
added 22 characters in body
Nov 3, 2015 at 23:01 comment added Jeff Bowman @dodgethesteamroller In-depth discussions of a "contrarian school" like that are probably best suited to their own answer. Personally, I find that the Google Testing Blog does a pretty good job describing the virtues of fast, tightly-scoped automated tests alongside system-in-context tests. In case you found it unclear, I list the test pyramid here as a useful model or starting point, not as an excuse to stop thinking as an engineer.
Nov 3, 2015 at 22:38 comment added dodgethesteamroller ... consequently, Fowler et al., in arguing that less effort should be spent on integration tests and user acceptance tests because they are too difficult to write in a robust and maintainable way, are really just providing an ex post facto excuse for why they haven't figured out how to test well at higher levels.
Nov 3, 2015 at 22:32 comment added dodgethesteamroller A contrarian school of software QA, the context-driven testing movement, is in part devoted to disputing that there exist any such useful general rules of thumb as the Testing Pyramid. In particular, the seminal texts of the movement give many examples where integration tests are far more valuable than other kinds of tests (because they test the system in context, as a system)....
Nov 3, 2015 at 22:26 comment added dodgethesteamroller +1 for "A test that is a perfect mirror of its implementation isn't really testing anything at all." All too common. I call this the Doppelganger Antipattern.
Nov 3, 2015 at 16:06 history edited Jeff Bowman CC BY-SA 3.0
Choose better words.
Nov 2, 2015 at 19:43 review First posts
Nov 2, 2015 at 21:41
Nov 2, 2015 at 19:39 history answered Jeff Bowman CC BY-SA 3.0