Dave Abrahams

This is done, right? Should I close it?

> This project is a huge sink of time I hear ya. Parsing is a deep deep well; I put my Earley parser implementation on ice some time ago (the...

> In gearley, an uint32 is associated with each token and you're meant to create your own side table to access whatever information you may need. As long as there's...

> Internally, each belong to a different data structure during the parse. Oh that's an interesting idea. By the end I was memoizing prediction sets and that was a real...

Care to explain? I didn't care much either but I can't see the relevance of that statement to how much grammar transformation one does or whether you use CNF. In...

If I may ask, what's the motivation for breaking rules down so there are no more than 2 symbols on the RHS?

Most of the item storage is described [here](https://github.com/hylo-lang/Lotsawa/blob/28d7a224ed19d03f4d819fecf895e6af21fc9368/Sources/Lotsawa/ItemID.swift#L9-L20). There's [another 32 bits ](https://github.com/hylo-lang/Lotsawa/blob/28d7a224ed19d03f4d819fecf895e6af21fc9368/Sources/Lotsawa/Chart.swift#L120) for the index in the chart where the mainstem of the item can be found. Mind explaining...

> So, instead of the bocage node ref id, you have a mainstem. > Also, you have the LHS symbol which I don't need. It's only the LHS symbol when...

I'll try to help; keep in mind that I don't know your code, though. Meantime, maybe it would be good for you to open an issue for things like this...

See https://github.com/swift-emacs/swift-mode/issues/193, where the discussion has progressed, for examples and more detail.