Push vs. Fetch
Back in what I like to think of as the olden days (let's say, the 20th century, give or take), email clients always bore the responsibility for checking to see if there were any new messages on the server waiting to be downloaded. You might click a button to do this manually, or set your client to check on a recurring schedule (every minute, every half hour, or whatever). Indeed, this is the way most desktop email programs still work. This process—asking whether there are new messages, and downloading any that are there—is called fetch or pull, emphasizing that the activity is always initiated from the receiving end.
Fetching works just fine, but it involves two kinds of inefficiency:
If there's no mail waiting, the transaction is ...