score:44
They still do it that way:
LCU Replacement in Preliminary Design, Anticipating 2022 Fleet Debut
If the beaches are heavily defended, the Navy is supposed to bombard them prior to the landing. Occasionally the military and naval planners have been known to make deadly mistakes.
As pointed out in the comments, the debouching troops and equipment hit the beach much more quickly when going out the front. If there is too much fire from that direction, they may leave the "door" shut, and may go over the protected side. But this slows down the process, and reduces the equipment load.
Amphibious warfare provides a list of amphibious operations for modern wars. My father came ashore this way at Anzio.
Image: Into the Jaws of Death: Troops from the U.S. 1st Division landing on Omaha beach as part of the Allied military campaign to free France from the stranglehold of Nazi Germany
Upvote:49
The LCVP or Higgins boat was designed with one primary goal: to land a large number of troops quickly. Therefore it was important that:
Given this, the solution, inspired by observations of Japan's Daihatsu-class, was to install a large bow ramp so the passengers (which could be jeeps) quickly disembarked from the front into shallow water, but the rest of the craft was still in deep enough water to be able to drive back out.
So why not use rear exits, especially since it's a common feature of modern IFVs like BMP, Merkava or Bradley? The difference is that the LCVP isn't amphibious; disembarking from its rear means troops could end up in deep water and potentially drown in full combat load which made them sink, while not giving them a great firing position and also blocking the LCVP's return trip. With amphibious landers like the BMP, the vehicle can drive onto dry land giving its disembarking passengers great cover.
It's debatable whether having rear exits and forcing troops to swim to the front is better than being able to land more troops simultaneously and overwhelm defenders, but what happened at Omaha beach is the result of multiple failures, and a situation where the LCVP's design simply wouldn't have made much difference either way.