To be fair there aren’t many questions on stack overflow about, “How can I design my hyperblock selection algorithm for improved performance on a block-structured data flow processor?”

There might be if you first go out of your way to answer questions about dataflow throughput by pointing people toward block-structuring and mention keywords like "hyperblocks" there. People already do this; I've noticed a good few times now that I've done a number of increasingly-specific searches as I understand a problem better, and each time the answer that actually guides me to the next step comes from the same person/place.