Jared's Weblog

Main Page - Books - Links - Resources - About - Contact Me - RSS

Jared Richardson Most Popular
Help! I've Inherited Legacy Code
Testing Untestable Code
Continuous Integration... Why Bother?
Mock Client Testing
The Art of Work
Technical Idiot Savants
Targetted Skills Aquition vs Free Range Chickens
The Habits of Highly Effective Developers

Blog Archive

Fri, 24 Feb 2006

Diving Deep: Introducing a New Technology or Language

Bruno Melloni passed on his experiences introducing a team to a new language or technology. It's good enough to pass on, so here you go.

So, with Bruno's permission...

Thank you! Ship It! rocks.

For the last 15 years I have been following the principles you describe in the Ship It! book. And for 15 years people told me that I was nuts, even though they brought me success after success in my projects. It is unbelievably satisfying seeing those principles finally recognized in print. Thank you!

I'd also like to suggest an extension to Tracer Bullet development that I used in 1996 when I had to rewrite a system that had no documentation, no original source code, no past experts... just the system's output as a model, the target language was Java, I was the only developer that knew Java and all of the other team members only knew C++:

After the skeleton, interfaces and stubs are created, do a code deep dive - from UI to database - on a small slice of the application. Or in other words, write the full and final code of that slice before even bringing in the rest of the team (or done by a team with lots of code reviews to optimize the result).

A deep dive, when done with readability in mind serves as a template for junior and new developers. As a consequence those C++ developers were fully productive writing Java code within a week! And because method names, variables, and even logic were chosen to make them easy to understand, we were even able to show code to the business experts who could then validate our assumptions.

The net result was that in 6 months we successfully duplicated a system that nobody knew how it worked, made it run 10 times faster (yes, on interpreted Java!), discovered bad data that had not been detected in years and cleaned it up, and the company went from outsourcing 80% work to insourcing lots of work because of the new excess capacity. As far as I know, the system is still happily in operation after 10 years... All of this thanks to the Tracer Bullet and Deep Dive combination of approaches.

Bruno Melloni

Thanks Bruno! I like the idea of Diving Deep. I suspect many others will too!

By the way, this note came to me via the Pragmatic Publishers's feedback page about Ship It! Feel free to pass on your experiences via that form or on the Ship It Mailing List.



posted at: 08:54 | path: | permanent link to this entry