Petroski, Henry (1992). To Engineer is Human: The Role of Failure in Successful Design. Vintage. Chapters 14-17.
Chapter 14 was a bit confusing. Petroski introduces two cases where engineers attempt to obviate failure in the form of fatigue however these engineers fail (the example he uses are the Comet and the Alexander S Kielland). Is there more to this chapter? How do engineers benefit from failing at preventing failure?
On p. 222, Petroski states: "Success ultimately leads to failure, aesthetic failure, functional failure, and finally structural failure". This makes little sense to me, seeing that a success likely means that engineers have identified proper techniques in construction and design to avoid these types of failure. What does he mean by that quotation?
If fiction novels are a better way of "expounding on engineering design and structural failure" (212), then why didn't Petroski write a fictional novel? Wouldn't more non-engineers be more prone to pick up a fictional novel and learn about engineering that way than by reading this book? (Basically, what is the advantage to this type of literature for this subject matter?)
In chapter 17 and in the Afterword, Petroski makes a point to show the reader that the blame for structural failures should be shared by others, such as those who ignore the engineers warnings and misuse the product/structure, but if he has done his job in writing the book should he need those chapters at all? Shouldn't the reader understand these things implicitly from the rest of the book? Or is he just patronizing the reader?
On page 215 Petroski says, "He cannot bury his mistakes in the grave like a doctor." Is he implying that if a doctor kills someone its okay because the dead person can't complain about it?
What was the point of the last chapter? Was Petroski trying to establish a viewpoint that he believes there is a limit to what man can design?
In chapter 14, Petroski states that engineers must get to the bottom of a failure in order to improve on the technology but he also argues that when an airplane crashes, "the final report only concludes a "probable cause" because there is no knowing when future evidence...will provide a "cause behind the cause."" What is his argument? Does he think it is NOT possible to get to the bottom of a given technological failure or that it IS possible and must be done.
If failure is the primary way to learn then why do engineers have such a hard time rejecting their first designs?
Which list, McKaig's or Blockley's, do you think is a better characterization of the causes of structural failures?
In the beginning of chapter 17 Petroski discusses how people misuse technology, how can engineers prevent misuse from happening, or prevent catastrophic accidents if it does?
Chapter 16 "The Connoisseurs of Chaos" seems to attempt to explain the difficulties in diagnosing the causes for structural failure in complex structures. Primarily, the difficulty seems to be the large number of things which can go wrong with a structure. It states that there is difficulty in diagnosing the primary causes of a structural failure, in general principle, rather than with respect to a single incident. Why is it so difficult to identify chief causes of failure?
When Petroski says that fatigue failure can be obviated by making sure that no cracks are present in a new structure, what does he mean by for tens of millions of cycles of safe-life can be guaranteed
?
How does a slide rule work?
Is it really customary engineering practices to keep flying planes
that randomly explode in the air just because they can't figure out why
it is that they're exploding?
Why was the Tacoma Narrows Bridge designed differently than the George Washington and Bronx-Whitestone bridges? After attempting to follow in the example of the Brooklyn Bridge, Niagara Bridge, and Cincinatti Bridge, why wouldn't the engineers of these three separate newer designs bounce their ideas off of one another so as to prevent the failure that occurred?
"..it is critical to identify exactly how a structure may fail." pg195
I argue that even humans dont know "exactly" how a structure will fail. Identifying failure beforehand without simulation. Besides, it isnt that hard for computers to guess like we do. The overall tone of this chapter seems like computers create more risk. The author doesnt dwell a lot on the negative effects of digital calculators. For some reason, he takes the absolute pessimistic view that engineers will favor complexity over simplicity just because a computer can handle it. I think everyone knows that simplicity in design trumps complexity. I feel Petroski's criticism of computers is more about the negligence of its users.. not about the tools themselves. Am I right? If so, why does Petroski think today's engineers are dafter than his generation's. If I'm wrong, why doesnt Petroski raise the same concers about compasses, slide rules or digital calculators. I bet there were other pessimistic petroski-like figures who shunned those technologies as well because it made them seem obsolete.
In chapter 15, does Petroski advocate for or against the use of computers in design?
In chapter 15, Petroski warns engineers of the dangers of over relying on computers for calculations. However is this argument outdated? It seems computers today are much more accurate and precise than the obsolete “Commodore”. Or does Petroski argument still have relevance? Do current and future super computers, with super calculation capabilities, really necessary for the "everyday" engineer?
Petroski notes that before computers became popular and available in
the engineering field, all engineers would had to do calculations by
using a slide ruler or other charts. He also points out that in general
engineering instructors prefer teaching about these technologies'
"common limitations" (190). Today, he says that computers can do
any calculation and predict the amount of length, width, strength
and weight of material (steel) to have a successful structure, but
he also points out that the computer cannot store "trial and error"
experience, so its answers can produce a design that does not line up with
“reality’s” capability. So, what are the cons and pros of using a
slide ruler versus a computer? Do you think society is better off using
computers to minimize marginal error? Or is society just using computers
for their fast calculations? Should society prefer more limitations or
unlimited possibilities of designing?
Petroski says that computers are a great source to use for engineering, but only if the engineer knows how to ask the right questions? What do you think he means? He gives us a list sample questions, but he does not directly define what a “right question” is (195).
On 194, Petroski says, "the computer can be instructed to calculate the sizes of the various components of the structure so that it has minimum weight since the maximum stresses are acting in every part of it. That is called optimization." What does the first sentence mean?
Petroski's main issue with using computers to model and predict
engineering failures seems to be that he cannot fathom their being
powerful enough to efficiently calculate all the complex variables
associated with a project. Considering the dramatic leaps and bounds
made in computing power since the book was published, do you think his
opinion has changed on the usefulness of computers, or do their other
"faults" still outweigh their benefits? Do you think the chapter still
has any relevance in today's world?
Petroski quotes David Pye
"The requirements for design conflict and cannot be reconciled. All designs for devices are in some degrees failures...."
This seems like an opinion and it can be generalized to be glass half empty kind of thinking. There is nothing factual about this statement.
He says that "if you vary the terms of your compromise -- say more speed, less heat, less safety...It is impossible for any design to be the logical outcome of the requirements simply because the requirements are in conflict and a logical outcome is an impossibility."
This is true if requirements are vaguely stated as "more heat", "more speed". If requirements are specified using real finite numbers, it is possible for a design to be "one of" the many "logical" outcomes of the requirements. The time taken to achieve the design can be greatly enhanced by fast calculation/simulation/modelling.
On p. 220, Petroski states that "though there is never any excuse for a faulty design, even under the worst constraints, there should be room for understanding." However, his entire book is claiming that failure is unavoidable; how is he reconciling these two ideas? Is he saying that a faulty design like the Hyatt skywalk is inexcusable, but fatigue failures should be seen as inevitable? If this is the case, if a failure is the result of both human error and other factors, how do we (in a litigious society, as Petroski notes) draw the line and place responsibility?
Would it be more accurate to say that the Kielland disaster an engineering error or an error in construction?
Petroski says that engineers should welcome making their failures
public (pg 223), but does the airing of their screw-ups in the
"morning newspaper or in the evening news" really help anyone?
Engineering failures on any substantial levels are incredibly rare and
by making them known to the public, doesn't this just give the
incorrect impression that well engineered objects we use on a daily
basis are not safe? Surely some significant discretion is warranted?
Petroski talks a lot about engineers and people needing to walk the line between overconfidence and paranoia, but he never says how to find this line. So how are people supposed to tell when they are being overconfident or paranoid about a structure? After all, if the Golden Gate Bridge had broke we would say people were being overconfident of the structure, but since it didn't we could say those who thought it would fall were just being paranoid.
According to Petroski, cracked knives are related to the wearing down of buses in New York City in that... (Complete the sentence/sentences)
Can CAD's ever completely replace engineers in terms of safety and design? How close are we to achieving this? Can they completely get rid of human error? Is it a good thing to have an "artificial intelligence" designing our structures?
Beyond his point that engineering is human endeavor and thus subject errors and failures that engineers can then learn from, what do you see as the message of Petroski's book? Is he saying anything more or is that the overall message?
In chapter 14 and 16, Petroski describes how some popular novels about structural failure can help raise awareness about the dangers of reckless engineering. But Petroski has also made the point that engineering disasters are very rare. Do you think popular culture is promoting a balanced view of the safety of structures? Or is it skewed?
It seems that McKaig has a grim outlook on what is needed to succeed in engineering terms. Why does he use the word "intelligence" to explain the ignorance that occurs and ultimately leads to structural failure? Isn't "information" more appropriate, as you can teach individuals of a variety of intellectual levels to construct, supervise, or maintain a structure?
In chapter 17, Petroski talks a lot about failures and gives several designs and perspectives that equate to such conflict of learning from experience. I think innovation should begin with experience with different things without inhibition, there should be no limits. Even with studying past failures, innovation/engineering should be about risk and just going with where it takes us to. Is past failures the best way rely on a good chance of success?
;
