Comments for http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml en-gb 30 Mon 22 Sep 2014 05:12:39 GMT+1 A feed of user comments from the page found at http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml Anthony Green http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=83#comment5 - GlenIn real world situations repeatedly compromising on test coverage and code quality leads to to the project requiring such technical debt it contracts septicaemia. More than once I've overheard developers state that they're waiting for a poisoned project to die so they can ditch the client. That to me is unprofessional. Have the courage to face your client down. We should adopt the oath 'epi dhlhsei de kai adikihi eirxein'. Tue 08 Dec 2009 15:28:03 GMT+1 Glen Mailer http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=66#comment4 As a developer, saying "compromise on features, not test coverage" is easy. But in real world situations, convincing clients and management that for whatever reason feature X will be delayed by some time because you don't have time to write tests is less than trivial.Maybe one day everyone will understand the processes and the reasons for them as well as the developers themselves, but educating clients on the actual benefits of unit testing isn't simple. Fri 27 Nov 2009 21:15:28 GMT+1 Craig Webster http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=50#comment3 @mathie That's a fair point. There are exceptions to every rule, but in order for them to be exceptions and not the norm it's good to establish the rule first - even if that is a sweeping generalisation.In my experience the only kind of technical debt that gets repaid is tangled but well tested code that was written in a hurry. Maybe it's not DRY or KISS or maybe it's particularly hacky, but it works and can be proven to work so it's okay. This debt can be removed with a reasonable degree of confidence that nothing is breaking thanks to the tests. If the code isn't tested then the debt tends to stay because no one has time to write the tests now, and eventually knowledge of how that code is used disappears. When that happens any change may introduce a bug so no change happens and the technical debt lives forever. Fri 27 Nov 2009 15:58:59 GMT+1 Graeme Mathieson http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=33#comment2 @craigwebster Well, if we're making sweeping generalisations, then if your team's culture requires you to cut back on tests to meet your deadline then you've got far, far bigger problems. :-)Even the best development teams build up a little technical debt over time and in my experience, that debt often creeps in near a tight deadline. There are valid reasons for introducing the debt as a trade off against other factors (not delivering a feature or delivering late, for example) and it's OK to make that decision sometimes.So long as you commit to paying up the debt again quickly, that is. ;) Fri 27 Nov 2009 15:36:41 GMT+1 Craig Webster http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=16#comment1 If you're dropping or cutting back on tests to meet your deadline you're doing it wrong. Compromise on features, not on quality. Features which didn't make it into this release will make it into the next. Tests which didn't make it into this release are unlikely ever to make it into a release. Fri 27 Nov 2009 14:29:43 GMT+1 Anthony Green http://wwwsearch.bbc.co.uk/blogs/webdeveloper/2009/11/testdriven-development-in-agil.shtml?page=0#comment0 "Cutting tests to meet a deadline is like taking up smoking to lose weight."http://www.oreillynet.com/pub/a/oreilly/oracle/utplsql/news/fulldoc.html Thu 19 Nov 2009 17:10:52 GMT+1