Hacker News new | past | comments | ask | show | jobs | submit login

I have a theory that commercially successful software development methodologies are like diets: they have to be almost impossible to follow. This ensures that when you fail to lose weight/achieve bug free software, you blame yourself for not following the rules exactly, rather than the rules for not working.

At the very least, I think it's common to judge methodologies by "how well does it work when it's done right?", when we should be using a three-part test.

1. How much value is there if it goes right?

2. How easy it is to get right?

3. How gracefully will it break down if it's not done perfectly?

Methodologies don't have to shine on 2 or 3. NASA-style clean-room coding is the gold standard of hard-but-effective; you have to suck it up and commit entirely with a great team, but the results are unbeatable. Waterfall was "good, bad, bad", but used by lots of less-exceptional teams with hideous results.

Agile, I think, improved #3, but not #2 so much. It's commonly sold by appealing to how much it can potentially help, but I think the more honest pitch is "it's tough to do right and it'll sting if we get it wrong, but it hurts a lot less than getting waterfall wrong."

Then maybe there should be one rule (as was my most successful dieting effort): Is what you're doing helping make or save money?

I know such a simple rule is probably causing a lot of you to cringe, but I can very easily fit just about any activity of value you or your team embarks on into that question, and it's a rule the entire company can get behind.

That's probably a very apt metaphor, but there's an upside too: many diets actually do at least temporarily help people in spite of their irrational basis, just because $random_diet causes the dieter to observe and reflect on their eating habits more carefully in general. Even a non-sense diet like "eat only foods colored yellow or red" probably does a lot of people a lot of good, compared to their usual state of being completely unconscious of their awful habits. Same goes for software development practices.

I think with this analogy you're presuming that people are exposed to a variety of foods and there is some positive correlation between action and awareness. However, for someone who only ever goes to McDonald's to eat, they'll just switch to french fries and ketchup for your nonsense diet.

I've definitely seen a "your cholesterol going up only means you need to Paleo harder" post on a forum.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact