People should please take away from this paper the central claim that they are offering up a fun, easy, and startlingly practical way to implement a parsing library in your favorite language / environment. It is but a few hundred lines in some functional languages. Scarcely importantly more in other languages. It is (per their plausible claims) a nice alternative to trying to muck up some dubious approximation using your regexp engine.<p>The "YACC is dead" claim of the original paper, and of this follow-up, is not so much "this is so good, everyone will stop using YACC". Rather, its that people don't use YACC when they ought to because it is inconvenient -- but this derivative-based parsing hack is convenient in important ways that YACC fails to be.
<p><pre><code> Over the past year, even in the little time we've had to
work on the paper, we've learned a lot more about parsing
with derivatives.
In the week after the community found it, you all taught
us ten times more than that. Thank you!
</code></pre>
Why isn't all computer science research done like this?
Claim strikes me as a bit premature given the amount of Yacc based code there is in the world busy producing away day in and day out. If the algorithm is what it is said to be and I've no reason to doubt that, then at some point we will all be blessed with a Yacc replacement. Since I've always found Yacc to be a PIA, I certainly won't miss the old regime, but since there is nothing to use at the moment, I'm going to hold off a bit...
I nearly don't understand anything of this. It sounds cool, though.There is so much math and lisp...<p>But I want to learn and if possible even help. What do I have to learn to understand this and maybe implement it in my own languages?