The policy is:
If you're scraping, your useragent string must include a contact email. (And ideally a URL of the project)
I'd like to be very bot-friendly, but that requires bots be friendly back.
I'd also like to get up a URL ($LJHOME/bots/ ?) which explains:
-- the rules (user agent, rates)
-- what we provide in machine-readable format
-- who to contact for other access
Then when we block a new bot that hasn't read the rules, the block message will include:
"You've been banned because you seem to be a new bot. Please read the bot rules at $LJHOME/bots/ and contact us to get unblocked."
Please pass word in relevant communities: lj_clients, lj_research, etc. (I don't follow everything.)