interesting implementation. I really liked the idea of just building a restful interface that people could plug into however they wanted to.

the best way to scrape HN would be for each user to submit an item's HTML or its cache of it when idle and when called upon. It could do a pre-check of the item, the time, size, # comments then do the whole page if called upon. You could keep a tally of who's submitting what. The system could go open-source as I'm sure other sites could use it. Perhaps the plugin just digs into a brower's cache at set periods of time. There might be copyright issues.