Changing Nature Of Software Rollouts

Jeremy Zawodny blogs about Jason Fried's sppech at Web 2.0 Conference. Jason notes that traditional software development is expensive, resource-intensive, and born of a Cold War mentality and advises to "think about one downing, instead of one upping, and underdoing competitors"–beating them with less. Protoype the app quickly. Don't build a ton of features, build a few really great deep features and then move on from there. Don't worry about up-front functional specs and fully documented requirements. Faster development and better products are the result. Say no by default. New features must beg to be added. Make sure there's a real demand for 'em. It's an agile development model.The hidden costs of "new." When you add a feature, it's re-training, updating docs, howto, training, may be the terms of serivce, and so on. Zdnet notes that he suggests 30 hours per week per person, which "forces you into building better products and being creative with your time." And, if you have less time, you have less time to think about abstractions, such as functional specification documents, which Fried characterized as a waste of time. "Instead, build the product and start from the user interface customer experience first; then wrap with the technology," Fried said. "The interface screens are the functional specification." Release a major update 30 days after launch. Have stuff planned for this, just keep in your pocked for the 30 day update. Builds momentum and good will with your users.
My Take: Use it while you build it. What a dreamy world - I would like to be part of this world.Of course well expressed and quite timely - These ideas may have appeal in the design and development of small utility like products – no doubt that these would influence all forms of software development- but enterprise implementations and industry strength software development have to respect the classical rollout – press releases like this not withstanding. Though I certainly note that multiple month rollout plans are becoming less and less and increasingly enterprise rollouts are becoming faster and faster. The direction is right but large scale transformation of the likes outlined by Jason in the enterprise may not happen that fast.