As a linguist, Larry was also acutely aware of the difference between prescriptive rules of grammar vs how language is actually spoken by real people, and in terms of informing Perl he always came down firmly on the side of the latter. And now perl is the gold standard for "regexes" even if they don't encode regular grammars anymore.
Agreed on Larry Wall understanding the difference between linguistics and computer science, languages for humans to use vs languages for computers to use.
I think his linguistics background helped him push the boundaries of what people expected from computer programming languages rather; being able to break out of the PL design mindset is part of what made Perl groundbreaking. If I recall correctly, the syntax and semantics of Perl are intertwined to the point that parsing Perl without also evaluating is provably undecidable. From a purely academic standpoint, that sounds like a nightmare, but it's hard to argue that Perl didn't have a significant impact. Many people can learn all the rules and conventional wisdom in a field, but being able to understand the underlying principles well enough to ignore some of them and produce something of value is much more rare.
> If I recall correctly, the syntax and semantics of Perl are intertwined to the point that parsing Perl without also evaluating is provably undecidable.
Not that big of a deal. Lisp has a drastically simpler syntax, but can behave is the same way.
This thing that is different from other languages is having a static lexer and parser, versus having a fynamic one. In Perl you modify it indirectly, and in Lisp you can modify it directly.
The fact that this happens has nothing to do with Perl's syntax complexity, but everything to do with dynamism in parsing.
That's hilarious, somehow I had never seen that one before! I love how it immediately makes it much more visceral how wild things can get without needing to get into the weeds of how programming languages work.