Open Profpatsch opened 6 years ago
I'm aware of Marpa but haven't been directly influenced by it.
But buried in their paper is a solution to the zero-length rule bug. And this time the solution requires no additional bookkeeping.
Have zero-length rules been a problem in Earley?
Yeah, that's been very tricky to get right, especially when you want to do semantic actions as well. This library uses the approach from the linked paper. But maybe it can be improved still.
The author of Marpa claims it to be an improvement over Joop Leo’s ’91 paper.
Has Marpa influenced this library’s implementation?