Yeah, that's probably the most dangerous statement in the article. There are some excellent tests that show that heaps don't perform better for inserts. And they certainly don't for deletes or updates.
I choked on my coffee when I read the OP's statement - enough to actually drop out of lurker mode and say something. 🙂 Considering the amount of time I spent when I started in this position dealing with index management to speed up database performance in an OLTP database, I was really, really surprised to see that in what appears to be an academic treatise.
-Ki