I have seen a fair few parser generators over the years, but it has been a long while since I have looked at anything that has been newly developed.

What improvements have been made to make them better? The problem domain seems pretty well defined and even 20 years ago the things that were changing felt like polishing off a few rough edges caused by earlier resource constraints.

I don't want to be dismissive and say "Why make this?" as a implied suggestion that it shouldn't have been made.

Nevertheless, Why make this? I assume there are good reasons for doing this that I am not aware of, what are they?