The format string contains conversion specifiers introduced by %. %s is the most general, calling toString() on whatever you pass. %d requires an integer, %f a floating-point number with optional precision (%.2f for two decimals), and %n emits the platform-correct line separator (use it instead of \n for portable output). String.format returns a String; System.out.printf writes directly to stdout using the same syntax.

A width number after % reserves at least N characters for the value, padding with spaces if shorter. By default text is right-aligned; the - flag flips to left alignment, which is what you want for variable-width strings in a table. You can combine width with precision (%-10.5s left-aligns and truncates to 5 chars). Use this for log lines, CLI output, or anywhere a fixed-column layout makes scanning easier than CSV.

Without an explicit Locale, String.format uses Locale.getDefault(Category.FORMAT), which means the same code can produce 1,5 on a German JVM and 1.5 on a US one. That is fine for human-facing UI but disastrous for logs, file names, or JSON: a Locale-sensitive number that a downstream parser cannot read. Always pass Locale.ROOT (or another explicit locale) for machine-consumed output. Number-heavy code that mixes both should standardise on Locale.ROOT and only switch to user locale at the rendering boundary.