With normal, or forward DNS, a domain name is used to look up information associated with that domain, including the IP addresses of its nameservers, webservers, and mailservers, and any general text information the domain owner would like to publish, including authentication records for the domain's mail transmitters. With reverse DNS, an IP address is used to look up a specific name associated with that address.
Information in forward DNS is under the control of the domain owner. Information in reverse DNS is under the control of the network owner, or at least the lowest owner to which authority over a large block of addresses has been delegated.
In general, reverse DNS is not as reliable as forward DNS. This is due to the communication barriers between owners of small domains and owners of large networks, and the lack of incentive for network owners to keep reverse DNS records up to date. Often they simply run a script which assigns numerical names in sequence to each address in their allocation. These automated names have no connection with the actual use of the address.