These lines of text have very strict rules limiting what they are allowed to be.

If you try to process source code with the unix tools built with the "just lines of text" model, you won't get very far.

If you use a parser with rich data types -- you can do so much more.

Unix is not just a kernel -- Unix includes the "application programs" and criticizing their weakness is criticizing Unix itself.