Connections - Intertwingled: Information Changes Everything - page 71

Of course, we often lack the luxury of space-time. This limits collaboration and the scientific
method. When thorny questions arise, folks love to suggest A/B testing. Sometimes that's a
great idea, but often the complexity and connectedness of the system make it unfeasible. It's dif-
ficult to isolate variables, and we can't always judge long term efficacy based upon the initial re-
sponse. Users adapt to change over time. Also, creating dual designs that integrate into the
whole takes a lot of effort. Eric Raymond argues that forking is a taboo of open source culture
and almost never happens.
There is strong social pressure against forking projects. It does not happen except under plea of dire ne-
cessity, with much public self-justification, and with a renaming. l xxxvii
The right to fork is an important freedom of open source, but it's also a last resort. Both child
projects have fewer designers and developers; and once done, it's impossible to undo. While it's
pretty to think so, we truly can't explore parallel universes. It's simply not possible to put all
your energy and resources into both sides of an A/B test. In reality, we often must satisfice with
imperfect information. This is where strategy can help.
To outsiders, a company's actions and acquisitions can appear chaotic, but behind the scenes it
may all make sense. For instance, Walt Disney built a sprawling empire, but each asset fit stra-
tegically in his map. lxxxviii He knew how the parts made a whole and never made decisions in the
dark.