But Delicious's database is already public (if you take out private fields on the user table and the private links). Even just the links + tags without any user info would rock for semantic Web usage.
I don't remember the robots.txt rules for sure, but doesn't that mean they don't allow crawlers at all?