James Robertson is taking on the curly brace nimby crew in support of Smalltalk [and lisp]. I wish him luck – our inability to expand our horizons is a major factor in slowing the evolution of software development. But he’s got a tough row to hoe ahead of him.
It’s hard to learn a new programming paradigm. Twenty years ago, I knew good Fortran programmers unable to transition to C because they couldn’t get a good handle on pass by value vs pass by reference. Ten years ago, I knew good C programmers unable to transition to C++ because they couldn’t get a good handle on objects. I’m willing to bet that there are good Java and C# programmers out there who will stumble over a transition to Smalltalk. And I’m sure that all of them would deny that there would be a problem before the fact.
There is a risk in adopting Smalltalk. And I think that Smalltalk would be better served by increasing the developer pool than by downplaying that risk. It’s not enough to make the language freely available. Provide tools and incentives to actually learn it.
The real question is whether it would be too little, too late? Would a Geoffrey Moore Crossing the Chasm marketing push still work? If you could somehow create a critical mass of people with both Smalltalk and domain skills, then you could really make some inroads.