Feedback trumps revenue (but not entirely)
One of the challenging questions about Max has been when to port to other IDEs. IntelliJ is a very attractive platform–customers are already guaranteed to be willing to pay for development tools, there is a reasonably large installed base, and it’s just cool. NetBeans has a different set of tradeoffs–more users (at least according to Sun) but yet another set of APIs to learn.
A thought experiment clarifies the situation. Suppose Max was now supported on all three IDEs. Now what? Well, the kind of experiment I did today to present recent test results would take three times as long. It seems pretty important to support all three IDEs from a single code base as much as possible, to reduce overall costs, but that is also going to require engineering investment that doesn’t translate into new features.
Max is still in the “what would convince people to spend money?” phase. Yes, I charge for Max. Otherwise I wouldn’t get an answer to the question. The price is artificially low to prevent money from being a barrier to adoption (lots of people, myself included a lot of the time, simply won’t pay for software, especially not development tools, but that’s a different piece of information). Success is defined for the moment by the number of validated ideas per unit time, not absolute dollars. Anything that interferes with validating ideas goes out the window. Anything that supports it stays in.
Fast forward six months or a year. Max now embodies a set of ideas that are clearly worth money to a significant percentage of developers (or development organizations). Instead of 0.001% (<- not an actual measurement) of developers being willing to pay $2/month, Max can convince 5% or 10% or 20% of developers to pay $5/month or $20/month (again, all made up numbers). The pace of innovation has slowed. The revenue per platform is higher. Now it makes perfect sense to port.
Trim widget with history
BTW, here is the new trim widget in version 1.2.8. Each vertical bar represents a test run. The most recent run is on the left. Here it looks like some trouble getting a test to run followed by classic TDD rhythm.
