Databases are the workhorses of the information age. Like Atlas, they go largely unnoticed in supporting the digital world we’ve come to inhabit. It’s easy to forget that our digital interactions, from commenting and tweeting to searching and sorting, are in essence interactions with a database. Because of this fundamental yet hidden function, I always experience a certain sense of awe when thinking about databases, not unlike the awe one might feel when walking across a suspension bridge normally reserved for automobiles.

The database has taken many forms. The indexes of books and the card catalogs that once stood in libraries are both databases of a sort, as are the ad hoc structured text files of the Perl programmers of yore. Perhaps most recognizable now as databases proper are the sophisticated, fortune-making relational databases that underlie much of the world’s software. These relational databases, with their idealized thirdnormal forms and expressive SQL interfaces, still command the respect of the old guard, and appropriately so.

But as a working web application developer a few years back, I was eager to sample the emerging alternatives to the reigning relational database. When I discovered MongoDB, the resonance was immediate. I liked the idea of using a JSON-like structure to represent data. JSON is simple, intuitive, human-friendly. That MongoDB also based its query language on JSON lent a high degree of comfort and harmony to the usage of this new database. The interface came first. Compelling features like easy replication and sharding made the package all the more intriguing. And by the time I’d built a few applications on MongoDB and beheld the ease of development it imparted, I’d become a convert.