\t: matching tabs in TSV and indented text

The reason to write \t rather than typing a tab into the pattern is legibility. A literal tab in a pattern is indistinguishable from a run of spaces in most editors, survives a copy and paste as something else entirely, and disappears entirely when a formatter runs over the file. The escape is two visible characters that always mean the same thing.

The tab separated file is where this matters most. Splitting a row on /\s+/ looks reasonable and quietly destroys the data, because a run of consecutive tabs collapses into one separator and every empty field disappears. The row "a\t\tb" has three fields, the middle one empty, and splitting it on whitespace returns two. Splitting on /\t/ keeps all three.

For indentation you usually want horizontal whitespace only, which JavaScript has no shorthand for. Other engines offer \h; here you write [ \t] by hand. That distinction is what separates trimming the ends of each line from flattening the blank lines between paragraphs, since \s would have eaten the line terminators too.

Inside a character class the escape behaves normally and combines freely, so [\t ,;] is a perfectly readable set of common delimiters. When you are handed a file of unknown provenance, a class like that plus a count of how many fields each row produced is a faster diagnosis than guessing at the format.