What separates text files from binary files, considering they are stored in binary

I use as follows:

Plain text

For configuration - usually using YAML or .ini. Deprecated by me for most uses except when a text file is the desired result (e.g. print to text, save to text etc.)

XML

For configuration and transportation of data; e.g. export, format via XSLT etc. Good as a portable file format (e.g. SVG). Excellent manipulation tools and filters.

Databases

Main data storage from inside app/webapp. Use it all the time as storage of choice. It's reliable, robust, and you get a lot built in (transactions, referential integrity, cascading delete/update, indexes, speed). Best used with a layer, or ORM (IMO).

Single file archive (e.g. .zip)

Suitable for storing related multiple binary streams compactly, e.g. ROM images for an emulator. Best for things that don't often or never have to be updated. It is heavyweight, slow and hard to manipulate;

Binary

Only where a database is not available for storing app data. Easiest with serialization (C++). A highly tuned binary format will outperform everything else for both speed and size.

Using fscanf() in itself probably explains most of it. fscanf() has to interpret the passed-in format string, and then has to scan the input stream from the file, trying to match the specified pattern. That's actually a huge amount of work. read() just has to read in the specified number of bytes from the file and doesn't have to do any parsing of the input. By contrast, fgets() does a little more work than read() since it has to watch for newlines, but it does a lot less work than fscanf().