Re: [boost] Spirit C++ parser framework

From: Julian Smith (jules_at_[hidden])
Date: 2001-06-08 05:03:41

joel de guzman wrote:

> Shouldn't we do some rethinking? Why do comp-sci classes
> teach that bottom up parsing is `better'? Agreed, the *worst
> case* performance is exponential, yet what we should be
> concerned with is the *average case* and I see now that it can
> be much faster than bottom up.

Stroustrup writes some interesting things about parsing C++ in the D&E (page
68-69) - he originally wanted to use recursive descent, but was persuaded to
use yacc. He says `In retrospect, for me and C++ it was a bad mistake'.

- Julian
http://www.op59.net/