This is the first time I've ever heard anyone say, "I wrote an app that didn't use a database and it worked out for the best in this instance." I would counter with suggesting that, for any app that you expect is going to either (a) grow quickly, or (b) be unsupported by you in the future likely should have a database, otherwise you will have trouble.
What I'm taking away from the article is the ol' KISS prinicple.
I don't see why building on an RDBMS necessitates added cost, though. I can think of at least two production-quality products that are free, and depending on the application, using the structure provided by database tables and SQL queries can actually be simpler than using a mishmash of hashtables and XML files.
I don't want to make my own system for storing and retrieving records (juggling flat files?). I don't write my own server (no matter how little the project), and I don't make up my own database system.
The SQLite docs are clear, and there are a billion tutorials for interacting with SQLite in every programming language. It's easy to use, and it leverages the work of others so I can focus on the things I want to do.
This article is ridiculous. The lazy hack and the next lazy hack coughed along until they did it right.