This a great resource, I use it regularly as a reference and it would work well as a textbook too.
It's also comprehensive, from traditional statistical methods, linear models & context-free parsing right through to modern neural network-based approaches.
There's been a lot of progress in the past 18 months (mainly the transformer/self-attention model) that I don't think is reflected, but that's probably because it's so recent. I assume this will warrant a mention in future editions.
