Bambi Meets Godzilla

Round two of a fight Steve had already tried to walk away from. The original "Anti-Anti-Hype" post went up on the O'Reilly Ruby Blog, enraged the Python world, and came down voluntarily inside a day; this longer, cooler-headed replacement is framed as a reply to one aggrieved commenter. What it adds is testimony. Steve was around for the Smalltalk collapse and had talked with some of the key players — hallways echoing "with roars of protest," two commercial vendors, a crowd of about-to-be-ex-Smalltalk programmers — and by 2005 he thought he recognized the early symptoms in Python, at that point his own second-favorite language. The evidence he actually musters is strikingly small: one line in the Python Cookbook, a single word in a Jython book, a long-deleted FAQ answer, and a dinner-table riddle about waiters in Paris and Tokyo that does more of the persuading than any of the citations.

Read it with the scoreboard handy. Steve rules here that Python "has failed" in the commercial marketplace (his own hedge: "so far"), and stakes the next five to ten years on Ruby becoming "a Java-like phenomenon." One of those bets looked sharp through the Rails boom; the other produced some of the wrongest sentences he has ever published, since Python went on to become the most widely used language on the planet. What outlasts the forecasts is the mechanism — small repeated acts of unwelcome, compounding one turned-away newcomer at a time into a market outcome. Steve labels his whole theory "just my own opinion, endorsed by nobody," a level of hedging the flame war that followed never gave him credit for.