Hacker News Reader

Do you scrape wikipedia as questions are asked or are you working from a static data set?

The questions are pulled live, but I save old q's so I don't have to crawl the same page twice.