Objects Have Failed

The article says: "In the face of all this, it's fair to say that objects have failed." The author appears to define "this" extremely broadly, including: "address[ing] the computing requirements of the future" (whatever those might be); dealing with "software evolution" and "wholesale changes" (again, not sure what that means), although that's claimed to be addressed better by open source (huh?); making programming easy; and lots of other things. Well, OK, objects also haven't cured cancer. I'm not sure what the point is, except that the author sure is pissed about Lisp being less important than it used to be.

Objects were certainly oversold -- what computing trend isn't? But I think they succeeded at the problem they were intended to solve. Objects are basically encapsulation and a sensible bundling of behavior and state. It's just the next step after structured programming. Before structured programming we had line labels, and goto statements, and every time you had to do branching or looping, you had to express the control structures in these primitives. Now you write if statements and while loops. It's better, it's not controversial. It didn't solve world hunger, but that doesn't mean it failed. Structured programming succeeded in the sense that it is just accepted and is no longer even noticed.

Objects are an improvement of that sort. Objects have state. That state can only be manipulated by a few pieces of code (methods). This way of organizing code makes it easier to structure and debug software. It organizes coding patterns that everyone was reinventing anyway and no longer has to. It's a modest goal, and objects have succeeded in meeting it.