Screen scraping's original purpose was to get information out of mainframe computers and into client server systems. The new breed of screen scrapers use HyperText Markup Language to pull information off one Web site and deposit it into another site or a database, even if the two sites are not connected or if one site would prefer not to share information with the other.