Bosworth: Intelligent reaction, not intelligent design

Bosworth: Intelligent reaction, not intelligent design

Summary: Adam Bosworth, former Microsoft and BEA development guru and now a Google engineering vice president, gave a presentation at the salesforce.com user event that offered a much improved articulation of the notion of continuous improvement and innovation I wrote about my recent post on Microsoft making some of its Office interface improvements available now rather than in 12 to 18 months when the new version ships.

SHARE:
TOPICS: Salesforce.com
4

Adam Bosworth, former Microsoft and BEA development guru and now a Google engineering vice president, bosworth3.jpggave a presentation at the salesforce.com user event that offered a much improved articulation of the notion of continuous improvement and innovation I wrote about my recent post on Microsoft making some of its Office interface improvements available now rather than in 12 to 18 months when the new version ships.

He started by describing the old model (from five years ago) as a form of state socialism, with the Mandarins from the best schools locked in an ivory tower on a nice campus setting coming up with five-year plans, formulating big APIs and then having it built from scratch. That's a bit of an exaggeration, but you get the point. Ivory tower think, often out of touch with reality.

He described the new model as follows:

  • Run like mad
  • Try things out, watch, learn
  • Iterate
  • Learn form the customers in real-time
  • Real applications, APIs to follow

"I call this intelligent reaction, not intelligent design," Bosworth said. "You have to learn and change with the times, following customers in real-time."  He advised getting applications into the hands of users and then when they start screaming for APIs, provide them, which is the model the eBay, Google and now salesforce.com are following. "It’s not a science, it’s a living, breathing way to react to customer needs," Bosworth went on. "Don’t obsess about a grand plan. It doesn’t survive an encounter with reality. It’s not the clash of cymbals, but the inexorable evolution of natural selection."

The old model isn't dead, but what you can do iterating lots of releases quickly based on learning from users will ultimately be better for customers. Another element is a dynamic marketplace that spawns custom solutions and has a trust model (rating, certifying, peering under the hood) and reputation system that involves both the company and its customers, Bosworth said. 

Referring to salesforce.com's Appexchange, Bosworth, a long-time friend of the company, said, "I am sure some of it is wrong. Some things that salesforce has to do for you [customers] to trust the apps we don’t understand."

Topic: Salesforce.com

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

4 comments
Log in or register to join the discussion
  • Sounds like evolution

    Maybe hyper-speed evolution! Things that work get better. Things that don't work die off. The way everything works in nature! Except for airline bankruptcies . . .
    Roger Ramjet
  • Think about how evolution works

    Trys millions of possible ideas out, almost all of which fail. You can't run a business like that.

    With human intelligence we can do things that evolution cannot do, that is try things out as thought experiments before we actually do things. This is what planning and designing are all about.

    What is being described here just "code like hell" dressed up in fancy names.
    jorwell
    • My take as well...

      Basically it sounds like what is being extolled is live rapid prototyping. Make it happen now, we'll clean it up later. Sounds like a recipe for maintenance headaches.
      Zinoron
  • Several pointed it saying it is live coding madness.

    what is the difference now? We buy a product and it takes months of work to make it actually make it do 3/4 of what we want. Endless patches from the vendor to fix things that should have never been broken in the first place. I prefer the KISS principle the best.
    xshakes