The Building Boom
History is merely a list of surprises. It can only prepare us to be surprised yet again.
—Kurt Vonnegut
Of course, it's not as though nobody was working on Build tools and systems until the last year or so. In fact, there was a lot of development going on: Jim Weirich's Rake, to cite one example. An ant-like Build language written in Ruby. (Martin Fowler has pointed out the perhaps obvious but still interesting fact that Build tools such as ant and Rake are in fact languages, but languages that use a different computational model than the developer probably is using in the code they Build: dependency-based programming. Thus, you are programming when you use these tools, and in a less familiar paradigm.)
But for the most part, these tools have been elaborations on make, which has served the UNIX world for decades, plus ant in the Java world, and in the .NET world, what—Nant and MSbuild? In general, Build tools have not been venture capital-attracting commercial endeavors. And as recently as a year ago, it was possible, in the right setting, to say, "Everyone uses ant," without being contradicted ("Tips for J2EE Development," February 24, 2006, AWProfessional.com).
Now, though, a lot of companies have seen dollar signs in the Build market. And there will be more: "Success feeds on itself," Ousterhout says. "For years, software development managers resigned themselves to the idea that Build problems were just...a limp we can live with'." But when serious Build tools like Ousterhout's and BuildForge pointed the way, there were two effects. "Development teams are starting to realize that this is a limp they don't have to live with [and] entrepreneurs are seeing the commercial success of the leaders." Selling your young company to IBM, as BuildForge did last year, is one such sign of success.
Which means that even more companies will enter the Build tools market, creating new, innovative products.
So it goes.
—Kurt Vonnegut
