The software that helped launch the big data era was named by a two-year-old. The word means nothing at all, and that was the point.
Doug Cutting's young son had a stuffed yellow elephant, and he called it Hadoop. Cutting told CNBC that people in software are always hunting for names, and he had been saving this one for the right project.
The project came in 2006. Cutting had been working on Nutch, an open-source web search engine. Its developers were rebuilding two ideas that Google had described in research papers. The first, on the Google File System, appeared in October 2003. The second, on MapReduce, came in 2004. Together they showed how to store and process huge amounts of data across many cheap computers.
In early 2006 that storage and processing code was split out of Nutch into a project of its own. Cutting, who was then working at Yahoo, gave it his son's word. Hadoop 0.1.0 was released in April 2006. It went on to become a standard tool for what came to be called big data.
Why a nonsense word? Cutting explained his reasoning to CNBC. Names for software should be meaningless, he said, because what a program is used for tends to drift. A name tied too closely to one job could end up being wrong. A toddler's invention carried no meaning to go stale.
Naming projects after family was already a habit. Cutting's earlier Lucene search library was named the same way. Lucene is his wife's middle name and her maternal grandmother's first name.
The elephant stayed on as the project's mascot and logo. The original toy had a rougher career. CNBC reported that it was retired from play after Cutting saw it crammed into a toilet paper roll and shot across the backyard.
The projects that grew up around Hadoop include a machine learning library called Mahout. In English, a mahout is a person who keeps and drives an elephant. The word comes from Hindi and Urdu and is recorded in English from 1662.