Comments on: Spaghetti? Ravioli? Farfalle? http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/ Dagfinn Reiersøl on PHP, agile development, Ruby and other addictive substances Fri, 23 Jul 2010 20:17:39 +0000 http://wordpress.org/?v=2.9.2 hourly 1 By: dagfinn http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/comment-page-1/#comment-21 dagfinn Tue, 17 Jul 2007 17:52:11 +0000 http://localhost/wordpress/?p=1339#comment-21 <p>My intent with the word "design" in the graphics may not be clear in this context. It's just to say that you get mileage from the tests even before you start implementing. This is not an actual "design phase"; it's the process of thinking about the API that you have to do when writing the tests first.</p> My intent with the word “design” in the graphics may not be clear in this context. It’s just to say that you get mileage from the tests even before you start implementing. This is not an actual “design phase”; it’s the process of thinking about the API that you have to do when writing the tests first.

]]> By: Paul M. Jones http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/comment-page-1/#comment-20 Paul M. Jones Mon, 16 Jul 2007 01:41:55 +0000 http://localhost/wordpress/?p=1339#comment-20 <p>From your graphics, it appears you're saying there is no design occurring in test-later development. Is that an accurate reading on my part?</p> <p>If not, it won't be the first time I've read something wrong. :-)</p> <p>But if so, the graphics would seem to be in error -- design can certainly happen without testing. I would say that design "before" and design "during" implementation are at least two options that are not represented. (Whether or not such design practice is better or worse than TDD is a separate discussion. ;-)</p> From your graphics, it appears you’re saying there is no design occurring in test-later development. Is that an accurate reading on my part?

If not, it won’t be the first time I’ve read something wrong. :-)

But if so, the graphics would seem to be in error — design can certainly happen without testing. I would say that design “before” and design “during” implementation are at least two options that are not represented. (Whether or not such design practice is better or worse than TDD is a separate discussion. ;-)

]]> By: Travis Swicegood http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/comment-page-1/#comment-19 Travis Swicegood Mon, 16 Jul 2007 01:17:28 +0000 http://localhost/wordpress/?p=1339#comment-19 <p>Far enough... You're right, my definition of a bug is something that's been committed, not test cases that have yet to be implemented :-)</p> Far enough… You’re right, my definition of a bug is something that’s been committed, not test cases that have yet to be implemented :-)

]]> By: dagfinn http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/comment-page-1/#comment-18 dagfinn Sun, 15 Jul 2007 22:57:42 +0000 http://localhost/wordpress/?p=1339#comment-18 <p>...when you say "code written using TDD". You're talking about bugs in production code. What I was referring to as "bugs" are the mistakes you made as you write the code. You write something that doesn't work, the test fails, and you fix it. So there are "bugs" that exist for a few minutes at most, and are eliminated before the coding session is finished.</p> …when you say “code written using TDD”. You’re talking about bugs in production code. What I was referring to as “bugs” are the mistakes you made as you write the code. You write something that doesn’t work, the test fails, and you fix it. So there are “bugs” that exist for a few minutes at most, and are eliminated before the coding session is finished.

]]> By: Travis Swicegood http://blog.agilephp.com/2007/07/15/spaghetti-ravioli-farfalle/comment-page-1/#comment-22 Travis Swicegood Sun, 15 Jul 2007 19:56:01 +0000 http://localhost/wordpress/?p=1339#comment-22 <p>To be fair, you should reduce the size of "debugging" in test first. ;-)</p> <p>There are only two times when there will ever been a bug in code written using TDD: 1) when the original tester made a logical error while writing the test (the tool is only as good as the one implementing it), 2) code was added without proper test coverage (i.e., this if will rarely be hit, so I don't need to write a test for it). </p> <p>We've had a few bugs crop up in SimpleTest recently. One was just code that didn't have proper test coverage for whatever reason, the other wasn't technically a bug as no expectations had been put in place on how to handle that fringe case.</p> To be fair, you should reduce the size of “debugging” in test first. ;-)

There are only two times when there will ever been a bug in code written using TDD: 1) when the original tester made a logical error while writing the test (the tool is only as good as the one implementing it), 2) code was added without proper test coverage (i.e., this if will rarely be hit, so I don’t need to write a test for it).

We’ve had a few bugs crop up in SimpleTest recently. One was just code that didn’t have proper test coverage for whatever reason, the other wasn’t technically a bug as no expectations had been put in place on how to handle that fringe case.

]]>