Last week I attended the Geek Dinner and CTO Breakfast, two unrelated technology events in the area.
My notes from the Geek Dinner — speakers were David Spann and Alistair Cockburn:
- A tech manager shouldn’t know too much (he won’t allow others to do the work) or too little (others won’t trust his judgment.)
- Ask two questions often (daily? weekly?) of your development teams: What went well? What could be improved?
- IBM studies showed that companies focused on process didn’t deliver results. Those “flying by the seat of their pants” delivered.
- Software development is like a cooperative game — like a reality TV show — in which you solve problems today, get past the current obstacle, and then repeat it tomorrow.
- Responding to change is better than following a plan
- Customer collaboration is better than contract negotiation
- A working product is greater than comprehensive documentation
- People’s interactions are greater than processes and tools
- A greater development model is to shut 4-6 people in a room with a whiteboard, give them access to the customer, and have them ship something every month or two
- The typical “iron triangle” is “time, spec, budget — pick two.” It’s a cop out.
- The VC model is you keep performing, we keep paying, bit by bit
- Don’t ever go a month or two without releasing something
- Develop a list of features you want, pick the one with the biggest bang per buck, then renegotiate the list (feature reprioritization instead of feature creep.)
My notes from the CTO Breakfast:
- Several people have had good experiences with Vonage for voice over IP.
- Smart phones will be full fledged computers within a few years.
- Bruce: Too many OS’s and platforms for mobile systems right now. When that consolidates we’ll really start seeing a lot of apps.
- Bruce recommends going to China and taking a tour. Impressive infrastructure. Lots of engineers.
- Can you build a mash up on parts you don’t control?
- Flickr’s API makes it easy to get photos in, provides a natural, social lock in. Data is the new “Intel inside” — the value of the app.
- Instead of exuberant VC funding of pre-bubble era, the new model for web companies is to sell out to Yahoo or Google.
- Infrastructure now much cheaper, methodologies much faster, outsourced work much higher quality. Show stringing now possible. (Low budget or no business model is the model!)
- One person paid $1200/month for an experienced CS grad in the Philippines who had worked at the IBM lab there.
- Phil: When there’s no competition you can use Java. But when development can be outsourced for $10/hour you need something more agile.
- Mystical force drives innovation, precisely at the time it’s needed. Necessity is the mother of invention.
- The business model should be to take existing technology and execute on it now. Focus on now not the future. (Though you have to watch the future)
- Matt: If you’re early you can fail. Established technologies with better execution makes more business sense.
- Google, Bluehost examples of great second movers.
- Google Ads with phone calling coming soon.
- $1200 offered to boot Windows on a Mac.
- GPS cheap, could be built into cameras with compass, tagging every photo with location information.
- Use cases, Q&A, architecture, testing most important skills to learn now.
- Scott: New Mac Book Pro ripped and encoded DVD in 80% of real-time as opposed to 200% time with Power Mac.
- The up and coming $100 laptop and a steady income from Mechanical Turk might make a big difference to the poor in other nations.