4 ms·

Programming like Kent Beck

85 points by simenfur 14y ago

discreteevent 14y ago
Its refreshing to see that he focuses on integration testing. Most programmers I know understand that in any kind of engineering every decision is a tradeoff. So testing every line of code means you are trading off speed and flexibility for very little gain in some cases (80/20?). XP introduced the idea that there isn't a tradeoff. You take things to the extreme. So as a methodology its not a good one to follow. But as a force it was absolutely beneficial. There were plenty of moderates arguing against the whole waterfall thing for years and getting nowhere. XP recognised (conciously or not, I don't know) that the zealots who enforced waterfall weren't the brighest so there was no point in putting subtle arguments to them. They would only move to something that was a bit boneheaded. And so XP as a force actually managed to move a lot of the industry more twoards the centre.
- route66 14y ago
  http://leanpub.com/leprechauns http://leanpub.com/leprechauns In the downloadable sample there is a nice analysis which refutes the "zealots who enforced weren't the brightest". In the short: there were no zealots. Nothing was enforced, the term "waterfall" comes up much later and it was meant as being iterative from the beginning. So one might ask: how come that this folklore is being trumpeted all around like a truth?
  - discreteevent 14y ago
    I know that waterfall was originally meant to be iterative. What I meant by "the whole waterfall thing" was the practice of it in industry. And by "zealots who enforced weren't the brightest", I meant mainly management in enterprise development who arent' bright enough to understand what development is and so want to treat it as a repeatable process like manufacturing which is much easier to 'manage' than any kind of creative process.
  - ExpiredLink 14y ago
    Each hype needs a straw-man. At least, when someone criticizes 'Waterfall' and praises 'Agile' you know s/he is clueless.
andrewcooke 14y ago
wasn't beck the person who couldn't solve some basic problem recently? it was mentioned here. [google...] ok, here it is - re-formatting some text (in a very basic way - just clipping lines) - http://blog.8thlight.com/uncle-bob/2012/04/20/Why-Is-Estimating-So-Hard.html http://blog.8thlight.com/uncle-bob/2012/04/20/Why-Is-Estimat... the hn discussion is at http://news.ycombinator.com/item?id=3880522 http://news.ycombinator.com/item?id=3880522 (not sure what my points is; the article here seems quite reasonable - very good advice, in fact, imho. i guess it just stuck in my mind as being so odd.)
- KentBeck 14y ago
  Yes, that was me. I get stuck sometimes. In this case the key insight is that the recursive algorithm is much easier to get to work than the iterative one. There was also the added complexity of two mammoth egos attempting to collaborate.
  - andrewcooke 14y ago
    thank-you for removing the dissonance there (and i am so glad i added that disclaimer ;o)

85 points by simenfur 14y ago

Its refreshing to see that he focuses on integration testing. Most programmers I know understand that in any kind of engineering every decision is a tradeoff. So testing every line of code means you are trading off speed and flexibility for very little gain in some cases (80/20?). XP introduced the idea that there isn't a tradeoff. You take things to the extreme. So as a methodology its not a good one to follow. But as a force it was absolutely beneficial. There were plenty of moderates arguing against the whole waterfall thing for years and getting nowhere. XP recognised (conciously or not, I don't know) that the zealots who enforced waterfall weren't the brighest so there was no point in putting subtle arguments to them. They would only move to something that was a bit boneheaded. And so XP as a force actually managed to move a lot of the industry more twoards the centre.

route66 14y ago
http://leanpub.com/leprechauns http://leanpub.com/leprechauns In the downloadable sample there is a nice analysis which refutes the "zealots who enforced weren't the brightest". In the short: there were no zealots. Nothing was enforced, the term "waterfall" comes up much later and it was meant as being iterative from the beginning. So one might ask: how come that this folklore is being trumpeted all around like a truth?
- discreteevent 14y ago
  I know that waterfall was originally meant to be iterative. What I meant by "the whole waterfall thing" was the practice of it in industry. And by "zealots who enforced weren't the brightest", I meant mainly management in enterprise development who arent' bright enough to understand what development is and so want to treat it as a repeatable process like manufacturing which is much easier to 'manage' than any kind of creative process.
- ExpiredLink 14y ago
  Each hype needs a straw-man. At least, when someone criticizes 'Waterfall' and praises 'Agile' you know s/he is clueless.

andrewcooke 14y ago

wasn't beck the person who couldn't solve some basic problem recently? it was mentioned here. [google...] ok, here it is - re-formatting some text (in a very basic way - just clipping lines) - http://blog.8thlight.com/uncle-bob/2012/04/20/Why-Is-Estimating-So-Hard.html http://blog.8thlight.com/uncle-bob/2012/04/20/Why-Is-Estimat... the hn discussion is at http://news.ycombinator.com/item?id=3880522 http://news.ycombinator.com/item?id=3880522 (not sure what my points is; the article here seems quite reasonable - very good advice, in fact, imho. i guess it just stuck in my mind as being so odd.)

KentBeck 14y ago
Yes, that was me. I get stuck sometimes. In this case the key insight is that the recursive algorithm is much easier to get to work than the iterative one. There was also the added complexity of two mammoth egos attempting to collaborate.
- andrewcooke 14y ago
  thank-you for removing the dissonance there (and i am so glad i added that disclaimer ;o)