This post impressed me a lot. Steve have improved one famous algorithm for a special but very common case by combining it with one of the coolest datastructures out there. And (applying a pinch of poetic freedom here) yielded results that surpass those of the academics and giants of the field in versatility and power. All with a very terse piece of code that fits within a blog post. Very inspiring!
> And (applying a pinch of poetic freedom here) yielded results that surpass those of the academics and giants of the field in versatility and power.
I (as well as a dozen other students) were taught an IR course where we had to implement a search engine from scratch. We used tries both to provide spelling suggestions and to build inverted indices for text. That was 7 years ago.
Not to discredit Steve, he's done great work and his article is a great resource, but it's not the groundbreaking discovery you make it to be.
