SICP is a great book to learn some neat programming from, but it wasn't really a classic text in the same sense as the rest of the list, which is mostly original work or something close to it.
Personally, I think surveys of a field like Russel/Norvig's AI textbook, SICP, or CLRS are more worthwhile to spend time reading compared to original literature.
Perhaps the author thought it was subsumed by the MIT or PLT Online collections mentioned in the introduction? It's in both of those, but I'd agree that it deserves special mention.