It's true that different index codecs are being designed with different tradeoffs (index size vs postings lists access time and cost), but the work is very incr

It's true that different index codecs are being designed with different tradeoffs (index size vs postings lists access time and cost), but the work is very incremental IMHO, no huge advances to speak of. Also, can you talk about the top-k challenges you mentioned? Priority queues are not optimal enough ?