The claim that the RMS Titanic was "practically unsinkable" may have been more a marketing tactic than a commentary on its engineering, but its prelaunch reputation of being impervious to the perils of the high seas has lingered for the past 100 years.

It is dangerous to cast engineering projects in such absolute terms—of course there had to be some combination of conditions under which the ocean liner would have failed. As elegant and grand as it was, however, the Titanic—like any other ship—was far from unsinkable.

At nearly 275 meters long with a gross weight of about 42,000 metric tons, the Titanic was the largest ship ever built at the time. It featured 16 major watertight compartments in its lower section that could be sealed off in the event of a punctured hull. Yet the luxury liner sank less than three hours after colliding with a massive iceberg in the North Atlantic, despite some estimates that it should have been able to stay afloat for as long as three days after an accident at sea.

The watertight compartments proved to be a fatal design flaw—one that James Cameron illustrated well early in his 1997 film recounting the fateful April night in 1912 when the Titanic sunk, taking about two thirds of her 2,200 passengers into the icy waters with her. The 90-meter gash in the Titanic's hull caused the ship to take on water near its bow, flooding six of the compartments. When enough water had penetrated the hull breach, the ship pitched forward at an angle that caused water from the individual compartments to spill over their bulkheads, inundating the front of the ship and sending the Titanic like a torpedo to the ocean bottom almost four kilometers below. Had the bulkheads been higher, or watertight at the top as well as the bottom, the water rushing into the hull might have been distributed more evenly, giving passengers more time to escape.

Ironically, builders of the Titanic were given a preview of how their ship might react to a hull breach several months before it even left port. On September 20, 1911, Titanic's sister ship, Olympic, was broadsided by the British warship HMS Hawke, which ripped away metal plates and riveted joints, leaving an 11-meter opening in the starboard side of the Olympic's hull. The collision caused the flooding of two of the Olympic's lower compartments, but the ship was able to make it back to port, perhaps contributing to the unsinkability myth.

Engineering and design are an important part of any construction project, but they are part of a larger system that includes the people that will manage and use the project's end product, whether it is an ocean liner, suspension bridge or spacecraft. Scientific American spoke with Henry Petroski, a professor of both civil engineering and history at Duke University and author of To Forgive Design: Understanding Failure, about the folly of believing a design is infallible, the Titanic's fatal flaws, and how even the best-engineered technology fails when a larger system breaks down.

[An edited transcript of the interview follows.]

It has been a century since the Titanic disaster. Will engineers ever be able to build an "unsinkable" ship?
The short answer is no. And anyway, it seems the claim that it was unsinkable didn't come from engineers but rather from advertisements for the Titanic. The ship had a lot of design features—such as the watertight compartments and their bulkheads—that may have led people to believe that it wouldn't sink.

Any design, whether it's for a ship or an airplane, must be done in anticipation of potential failures. In the case of the Titanic, the engineers would have been asking themselves: "What if we have a hole in the hull?" Well, water's going to come in. "How much water?" That depends on how big the hole is, so you have to make those calculations. You can always imagine a bigger hole or some worse condition.

What were the Titanic's greatest design flaws?
Probably the fact that the bulkheads didn't go higher, so that they weren't truly watertight and didn't actually compartmentalize water between the bulkheads. Other design elements meant to ensure passenger safety weren't adhered to. Although the ship was designed to carry enough lifeboats, it wasn't at the time of the accident, for example. That would be unheard of today. They had radio, which they called wireless back then, for calling other ships, but it was seen more as a novelty at the time, and ships turned them off after hours.

The Titanic also failed to incorporate a crucial safety feature available long before its maiden voyage. In the 1850s there was a British ship called the SS Great Eastern designed by Isambard Kingdom Brunel and built by John Scott Russell that featured a double hull. A double hull is a similar concept to bulkheads. Water comes in but you keep it from overtaking the interior of the hull. Generally speaking, the distance between the hulls is not that great, so the amount of water that gets in won't be that great. The debate over double hulls goes on to this day. After the Exxon Valdez oil spill there was a question as to whether all oil tankers should have double hulls.

Given the expense and careful attention to detail that went into building the Titanic, can you speculate as to why it did not have a double hull?
This is a point of debate. Looking at the long view, even though double hulls did exist at the time there would have been pressures on shipbuilders not to build double-hulled ships simply because they would be more expensive. There would also be a tendency among ship owners, operators and captains to argue that they know how to steer their ships so that they don't puncture a hull, that a double hull is an unnecessary expense. As long as there were no accidents it would be hard to argue against that reasoning. Then when something like the Titanic happens there's a reaction to that failure, thinking that if there was a double hull the ship wouldn't have sunk.

What parallels can you draw between the Titanic and the Costa Concordia, which ran aground in January off the coast of Italy, killing 30 or more people?
There's going to be a lot of pressure to operate cruise ships differently. They're not going to be going close to land as that one did. There are probably also going to be calls for design changes. I understand it took them quite awhile to pump the fuel out of the ship, so making that easier will be an area they'll focus on. The Costa Concordia also listed so badly they couldn't get the lifeboats down on one side. That's something that will be looked at.

The Costa Concordia situation is analogous to the Titanic's sinking. In the case of the Titanic, ships were moving through the North Atlantic at a regular rate at that time, and the fact that they almost never hit icebergs was interpreted as: "We don't have to hit icebergs, we know how to avoid them." This raises the level of confidence almost to overconfidence that nothing's going to happen because nothing has happened, which is not logical, of course.

What should the shipbuilders have learned from the Olympic's accident months before the Titanic struck the iceberg?
This should have been a warning. There's so much written about the Titanic, and it's hard to separate what's fact and what's fiction. My understanding is that the way the Titanic was designed, the emphasis was placed on surviving a head-on collision. The idea of a side grazing was apparently not anticipated, although it's hard for me to imagine why.

How much of a factor do the materials used to build a ship play into the safety and reliability of that ship?
Part of a design is to specify the materials. They were building the Titanic and its sister ship the Olympic at the same time, so they needed a lot of rivets. It turned out those who supplied the rivets were overwhelmed by the shipbuilder's demands. This led White Star to look for a broader range of providers, so the quality control wasn't as good as it would have been if they were dealing with a single supplier. Some of the rivets in the ship were steel, like the hull plates were, but some of the rivets were made of iron, and not the best iron. There were rivets that were put into the ship that were not sound. Getting the right materials continues to be an issue for large building projects even today.

How do engineers take into account poor judgment or misuse when designing something as massive and complex as an ocean liner?
It depends on whether you look at the ship as a structure made of steel and rivets and so forth or if you look at the ship as part of a larger system that transports people overseas over long distances through waters where there are icebergs. Not only does a structure need to be robust, it also needs to be robust under operational conditions where people make mistakes and things go wrong. The systems approach was very important to understanding what happened to the Titanic.

For a more recent example, look at the space shuttle program. Before the Challenger accident, there were 24 successful space shuttle flights, although quite a few of them actually had leaks in the O-ring, which turned out to be the weak link in the design and fatal for the Challenger. But it was rationalized away in the sense that, true there were leaks, but the shuttle missions were completed anyway. The system—which includes both the equipment and the people in charge of operating that equipment—tolerated those risks, and for awhile we were able to bring the astronauts home safely. Because no one can foresee all of the conditions that a design will be subjected to, managers, whether it's a ship's captain or NASA engineer, need to know and respect the limitations of their equipment.