I’m terrified of making one of these ‘big picture’ mistakes. I’ve long been weary of the fact that there is a surprising amount of engineering going on without scrutiny or peer review. And over reliance on computer modelling and increasing complexity makes it less and less easy to grasp the big picture.
Just yesterday I made a trivial error crunching some numbers for when a construction site could turn off their dewatering. Quite the opposite of this submarine, my structure would be 1000s of kips too buoyant and could potentially float up out of its excavation due to hydrostatic forces. A mistake like going unnoticed could easily be a career-ender.
The mistake is not that one might make an error, everyone will make errors. The mistake is the system that enables an individual contributor to make such an error without any accounting for these normal errors to occur. Second, blaming the person is just scapegoating and a further failure of a flawed system trying to cya.
Exactly. If all it takes is one person to misplace a decimal or fsck up a calculation and your entire end product is fucked... brother that ain’t you, it’s the system that is fucked.
For any system that matters, you have to prevent a single (or multiple) failures from sinking it. Any new failure that happens should be evaluated in a blameless manner and solved by improving the system so mistakes of that class cannot happen again.
I don't know the specifics for thread parent, but on many construction projects there is only one engineer, and even that engineer may be a sub to an architect. In my experience, architects are offended at the suggestion that they check the work of engineers, and engineers are offended by questions from clients. So, entirely due to the emotional preferences of the "one person", that one person is really totally responsible for a great deal.
Makes it impressive how SpaceX can iterate so fast on designs and yet have few of these types of errors. It seems they understand their failures before they happen.
Hire people (on good money) with the personality type that likes to find problems. If the team is all composed of "Mr Right-Now" and no "Mr Right" personality types then it will move very fast and have lots of issues. A diverse team helps.
TDD tries to insist tests are developed first, partly to ensure they're not skipped.
Reinforce the message: "Measure twice, cut once".
Good leadership and top-down support to make changes are important. If they won't care enough to stump up the money and follow through to check it's happening, don't expect much change.
And over reliance on computer modelling and increasing complexity makes it less and less easy to grasp the big picture.
More importantly, this also diminishes order-of-magnitude/approximation thinking, which is a very important source of "common sense". I've graded physics/electronics exam questions, and it's astonishing/amusing just how far-off some of the wrong answers are. Some memorable examples:
- Asked to calculate the current-limiting resistor for an LED in a low-voltage circuit, some students gave answers in the megaohms, others in the miliohms (correct answer was in the low kiloohms.)
- A question about the engine power of a dragster resulted in wrong answers of over a GW (roughly a million horsepower!) and under 1kW. (Correct answer was around 1MW.)
I find that students routinely divide instead of multiplying, especially if they are asked to solve for something and then calculate the result. For example, $f=ma$ gets transformed to $a=fm$ instead of $f/m$, yielding a large error unless $m$ happens to be near 1kg. Checking units is good way to check for incorrect formulae, but I've found that students who mix up division and multiplication are also uncomfortable with units. The use of calculators cannot explain lack of understanding of units, so I think the problem is more basic.
I've known school teachers, and am somewhat familiar with curriculum. Both are basically fine, and neither explains the problems I see with students in university.
It seems to me that the root problem is that students are never told, in serious terms, that getting the right answers is important. The way we signal importance is with grades. Students who have received passing grades for incorrect work through many years of school can hardly be expected to show attention to detail in university.
America (et al.) may be throwing away a generation of talent, by pursuing what seems to be a policy of unconditional approval. Luckily, there is some hope: my foreign-trained students hand in correct work, with appropriate digits and units. Quite often, they also show how they have checked the work. I don't think it's because their first language was French, or German, or Mandarin; I think it is because they went to schools where learning was still the focus, as opposed to the congratulation for simply being, not accomplishing.
>I think it is because they went to schools where learning was still the focus
Are you sure that's the case? Or is it perhaps because selection bias is bringing you only the smartest and best foreign students.
I went to a public university. The criteria to get accepted was much lower for Americans. Foreign students had multiple layers of competitive selectivity they had to make it through before matriculating there.
It’s interesting that it’s been common to see this exact same opinion (a lack of conscientiousness on the part of students) expressed in Germany about German education over the past 10+ years yet you cite German students as a good example.
In Germany it’s commonly stated that things were better previously and that there’s a ongoing decline.
My understanding in the German case is that it’s not supported in any measures of student attainment - grades, earnings after 10 years etc.
It is not necessarily a contradiction, it is possible that German education is now clearly worse than 10 or 20 years ago, but still not as bad as the current situation in grandparent's country.
I think it’s regression to the mean. Parents with lots of education see their kids struggling and don’t understand that they themselves were positive outliers back then.
Also that rich kids think riches come for free. Boondocks to boardroom in 3 generations, and three generations back. Rags to riches to rags. Shirtsleeves to shirtsleeves. Clogs to clogs. Rice fields to Rice fields.
Perhaps all are true as you said, and that this is civilization's path towards always using the calculator for any calculation and uncritically trusting the output.
> It’s interesting that it’s been common to see this exact same opinion (a lack of conscientiousness on the part of students) expressed in Germany about German education over the past 10+ years yet you cite German students as a good example.
That fits the Dunning-Kruger effect. They are better than others at what they've been paying attention to improving. Not sure how to square that with their perception of decline though; maybe others who haven't been paying attention to the issue have declined worse, or just started out worse.
I don't think it's an issue of the focus on grading/getting the right answers, more a lack of focus, when teaching units, on getting coherent results, in regards of the units: for example not multiplying a distance and a time to get a speed and so on.
>in regards of the units: for example not multiplying a distance and a time to get a speed and so on.
From my own personal experience it was exactly units like velocity d/t and pressure m/a or things like volume d^3 that I made the most mistakes on while I learned. Especially when it came to having to convert between different units within an equation. The only thing that really helped was practical experience working with real world units and geometries.
I think one of my bigger problems was being unable to clearly visualize what it was the units actually represented. It wasn't until I started having to calculate distances, speeds and areas on maps, culvert diameters and volumes, water pressure through tunnels and actual 'real things' that I really started to understand and stopped making errors with units. But it took quite a while and it was only after 5 years or so at my last job of working with units and measurements every single day that I really became very comfortable with working with them.
Your stereotypes are clouding your judgement. One could just as well/poorly say that the Mandarin students do well because they are rewarded for cheating all through school and continue to do so in college.
An old engineer (marine) who was a mentor of mine, lamented the end of the slide rule, for exactly this reason.
A slide rule can tell you the numbers; it can't tell you the magnitude, you have to know that going in. He claimed this taught engineers an intuition for figures that calculators don't.
There's should be a way build a calculator that did that, for for pedagogical purposes. For example, you could do everything in floating point and zero the exponent after each operation. I wonder if that's the same as the effect of slide rules? (Goes to check wikipedia...)
Maybe, but my parents came from the era of slide rules, and when I asked what the answer was to a calculation, I would be told "the answer is ... but I can't tell you where the decimal place goes"
We had this unspoken of law back in school where you better off write any answer than leaving the question empty. This is quite possibly drives some of these answers (at least here in my case)
The education system has some major problems with it... That said, a simple fix for this particular issue is to assign negative marks for a wrong answer (compared to zero for leaving it blank) - some multiple choice tests do this to minimise guessing.
I had a few teachers that gave negative points for wrong answers to prevent guesswork. The idea being that empty answers were feedback to them on what they might not have taught well.
Psychology and political majors can get away with that but it's not how safety/life-critical should engineering work. They original company failed to review, at multiple levels, fundamental calculations including density. They should have to recheck other calculations, such as hull strength and those related to propulsion as well.
To be fair, until I got an electric car, I would not have known intrinsically that it takes about 10-30 kiloWatts to keep it moving under normal freeway conditions or that 120kW is a pretty solid acceleration.
For some reason, we all know the right order of magnitude for a car in horsepower. But most don't know that 1hp is roughly 750w. It's odd because we all know roughly the power rating of a microwave or a lightbulb or a kettle. Amazing how use of one obsolete unit can totally mess up having a unified mental model for power.
Speaking of having a lack of a unified mental model for power , when it comes to the closely-related problem of energy (for society) there are a dozen or more commonly-used units in use, extending to "a cubic mile of oil". I believe that partly as a result of this (and partly due to the magnitudes involved), even those who follow the literature can have a lack of intuition when reading different papers and articles. And that has consequences, e.g., how many MW-GW capacity replacement per day for an energy transition.
yeah unless you have one of those cars that actually tells you the current output, it's pretty hard to estimate. to be even close to correct, you have to know the whole power curve and understand how the throttle is mapped (usually not linear). turbos add another confounding factor.
I don’t know why you’re getting downvoted. 1kW is a microwave, and 1GW is a nuclear power plant. It seems not unreasonable to expect people to understand this.
Maybe, but the fact that a high end commercial microwave is about the same amount of power as my car rolling along at about 30mph(45kph) on a level road would have surprised my intuition initially. I would have pegged a car for MUCH higher power output.
So, while I probably wouldn't have pegged a dragster at GW, I might have pegged it at 50-200MW assuming I put my car at an incorrect 200kW on average as opposed to top end (assuming a dragster is at least 100x a bog standard car wouldn't be out of bounds). In addition, assuming that a dragster has an instantaneous (or short lived) power output on the same level as a diesel power plant (designed to be continuous) probably also isn't out of whack.
My estimates are still hilariously out of whack--simply because my errors are compounding in the same direction.
I have had this happen at other times.
RF level strength loss due to antennas is one of those. I was hunting a problem and saw the signal but something like -90dbm down. I mumbled that "Wow, did we get our antenna match that wrong?" An experienced RF engineer looked at me and laughed "If you have an antenna that can actually kill a +20dbm signal to better than -90dbm, you should patent it. Check the firmware. Power amp probably isn't turned on." He, of course, was completely correct because he knew that killing an RF signal more than 50dbm or so gets really hard.
A car driving 100 kph (62 mph) on a highway, uses 2 or 3 ml (0.07 to 0.1 US liquid oz) of gasoline per second. That sounds also like about the size of a fire that you can use to heat up food.
"Maybe, but the fact that a high end commercial microwave is about the same amount of power as my car rolling along at about 30mph(45kph) on a level road would have surprised my intuition initially. I would have pegged a car for MUCH higher power output."
Well, I think most people don't have an intuition for whatever a "high end commercial microwave" is. I would expect a car on a level highway to be using on the order of 10 hp. I'm sure my microwave uses far less.
Air resistance is proportional to speed squared, meaning (air resistance component of) energy per distance is ~ speed², and energy per second is ~ speed³.
45 km/h vs. 90 km/h is almost an 8-fold difference of power to sustain. Air resistance is not the dominant factor on lower speeds, so I'd guesstimate it at 6-7x.
At 130km/h, with 33.6 MJ/l gasoline energy content, at 7l/100km, you get: 9.1l/h -> 300 MJ/h -> 84kW. At 30% efficiency you get 25kW or 33hp. That's about a dozen microwaves.
At 45km/h, air resistance power use should be about 1/24 of 130km/h. Guesstimating total consumption as 1/10, you get around 2kW, or a single (240V 10A) microwave/kettle.
Going at a steady 30 mph on a straight, level road is not really something people do in my part of the world, so I can't say that I have experience to know what sort of gas mileage you'd get.
Edit: More research suggests that gas mileage should be similar between 30 and 60 mph, it kind of plateaus, but I guess that means power required would still go up proportionally. 10 hp is something I was remembering from reading something about driving at around 60, so I guess 30 mph should take 5 hp.
In about the 80s, cars were rather more efficient at 50mph. Depending on brand, market segment, and transmission, I'd believe they are roughly as fuel efficient at higher speeds. My car (2.0L turbocharged hatchback) uses about 7-14kw to maintain 60-70mph on level ground at around 30-33mpg. The unfortunate uncertainty in my figures is due to poor memory
It's a meaningless question if the person answering it simply doesn't have some qualitative frame of reference to gain intuition from...and gaining intuition on EM phenomena grounded in reality isn't easy; there'd be a lot more EEs and physicists out there otherwise.
Besides, if you're walking into a time-constrained exam of toy questions, I fail to see what incentive a student may have to corroborate toy answers against a reality that it may or may not be an approximation of.
A lot of household appliances have power measured in watts (microwave ovens, electric heaters, etc). And megawatts and gigawatts are commonly used in discussion of electric power plants. My electricity bill is in kilowatt-hours, and that's what my electricity meter says on the front as well. Isn't that enough background knowledge to guess that an electric car would likely be more than a kilowatt (my microwave oven is more than that) but less than a gigawatt (a decent sized power station?)
No, not offhand. If someone, at some point, had said that kiloWatts and Horsepower are the same order of magnitude, I might have been a little better about that.
Gigawatt would have set off my "Huh?" detectors as I have heard that in correspondence with power plants. However, given that diesel generators sometimes power buildings, it would not have set off too much worry.
It's only 3% of the total but the characterisation of the error as a "decimal point in the wrong place" suggests that something was 10x heavier than it should be.
And how would you propose one gain intuition on what individual pieceparts of a build-to-print submarine ought to weigh, especially given the scale (and relative weight) of the thing in error can't easily be gauged by any individual? 10x error is hardly familiar territory if the thing in question is, say, trace heavy metals in a custom alloy.
If I gave you volumetrically equal pieces of 6061 and 7075 aluminum--whose difference in density is ~4%--would you be able to intuitively sniff out that something was off? What if those pieces were several tons a piece and your calibrated industrial weighing equipment had a reasonable +/-5% margin of error?
Sure if the industrial weighing equipment leveraged is, say, a weigh-in-motion system with an acceptable GVW tolerance of +/-10% as described in NIST HB 44[1].
In any case, this toy detail was intended to snuff out the grandparent's presumption that 10x error in any arbitrary parameter is obvious. The difference between 6061 and 7075 is offered as a simple thought experiment in that they're both common alloys in industry, yet difficult to discern one from the other despite the latter having a zinc composition that's 20x greater in the best case (6061 = 0.25% max; 7075 = 5.1-6.1%; 20 ~= 5.1 / 0.25) as specified by ASTM B209[2].
Fact of the matter is a contractor was responsible for a build-to-print submarine. A particular drawing specified something-or-other with an off decimal value that escaped human detection during multiple engineering design reviews. Contractor had something-or-other manufactured to the exacting detail specs in said drawing and no doubt was able to provide supporting documentation in the form of material certificates of conformance and lot testing results...but all the precision instrumentation and manufacturing quality control in the world amounts to nothing if the basis of truth is fundamentally flawed...and it certainly didn't help that the only effective opportunity to detect catastrophic design error (not that it's a build-to-print contractor's responsibility to identify such subtle faults) was spot validation of a system-level key performance parameter that wouldn't have been possible until the entire system was fully assembled.
I believe in this case the issue lies at the planning yard instead of the building yard. Building yards are usually not filled with engineers evaluating the design criteria sent from the planning yard.
A feel for reasonable values is rarely an explicit curriculum objective. Perhaps in quantitative introductory biology.
When a popular physics text had an ideal-gas-law chapter problem, with numbers describing solid argon, for multiple editions and many years, it was because everyone, checkers, instructors, assistants, and students, was mindlessly plug-and-chugging, not because there was a desire for students to validate simplifications or reflect on the reasonableness of results. There's like more than half a century of professorial letters to the editor, across multiple sciences, complaining of their graduate students lacking a rough-quantitative feel for the field.
I used to work in a school system where numerical answers weren't required in physics. Usually, the only reason you'd calculate a numerical answer was to help you check it for reasonableness. If you calculated a wrong numerical answer and didn't explain it, then your grade would be lower than if you had just stopped at the formula.
If a question asked you to compare two machines to see which produced the most power, you'd get a top grade writing a symbolic inequality and pointing out in words that some factors are common so they don't affect it, but one of them has an x^2 while the other only has an x, so that for large values of x, that one produces more power, while for small values of x, it's the other one. You'd get a lower grade for calculating two numerical results and identifying the bigger one.
Of course it depends on several situation, but during an exams people are under stress and mistakes happens.
What changes is the attitude of whoever is grading them.
I know one of my professor wouldn't give me full marks, but if the reasoning behind the computation was correct I would still get something. On such cases I simply wrote a note next to result saying that I understood that the results doesn't make sense and that there must be some computational mistake around but I didn't had the time to find it.
Other professors wouldn't accept nothing short that the correct answer, in such cases I know that whatever note I would leave would be pointless.
What do you expect students to do? Write "I don't think that my answer is correct"? Put no answer in? Or put in a plausible answer that isn't backed up by the calculations?
Wait, is the engine power measured in something else than KW nowadays?
I think that the HP -> KW switch happened something like 20 years ago, at least in my country.
I still see Japan use PS, which is known as Metric Horsepower. I think it comes from the German words for Horse Power, "Pferdestärken". It's 1 PS == 0.98632 HP
Vehicle dynamics literature does seem (as an outsider) to involve a slightly weird blend of imperial and SI units, research nowadays is metric but old research being cited in textbooks (i.e. RCVD has a lot of imperial)
All cars in europe have their power defined in kW on the registration certificate, hp is not mentionned.
It will make more sense in the future with electric vehicle.
Hp is used when speaking about cars, but it is a very misleading unit (bhp, metric hp, customary hp etc..) pretty much the same mess with imperial and us gallons.
Sure, my registration certificate has a kW value, but I have never in my life heard anyone speak of kW output of an engine(except for electric cars) and I'm a rather passionate petrolhead.
And apparently most of the countries use the SI unit instead of the HP. From Wikipedia:
Most countries now use the SI unit watt for measurement of power. With the implementation of the EU Directive 80/181/EEC on January 1, 2010, the use of horsepower in the EU is permitted only as a supplementary unit.
At least here in Germany, it's a bit like screen sizes: sure, the manufacturers and sellers have to state the SI unit, but everyone prints the other unit too, and people basically only talk about PS in cars and inches in screens.
This is a kind of problem where just knowing a lot of shit is a partial inoculation. If you know the population of New York City, and the weight of a Boeing 747, and the number of languages spoken in the world, and random shit like that... you catch these errors all the time.
Here you notice that engineering involves a lot of general culture, and making a good engineer starts at birth.
When I was a baby I asked my dad to draw power outlets, we had Legos, we’ve visited dozens of factories, I did a dozen programming languages before 18, I’ve always gathered data about things (Ok a nuclear plant is 1GW, a plane has 20t of fuel, etc), and I’m an average engineer.
Meanwhile my bigger sister has technical but never much about making things work, we went to the same engineering school but she ended up project manager because she can’t do the engineering job. So she manages people.
> there is a surprising amount of engineering going on without scrutiny or peer review
I've worked on testing DoD satellites prior to launch.
There were comment-level spelling errors [with "no real impact" to actual testing or functionality, honest] which passed through multiple reviews involving 3-4 people each review before being caught -- and even though they were in comments, required a version-uplift for the change.
And these were comments for Humans. Imagine if I or someone else buggered up on actual executable software functions.
[Hardware bugger-ups are a different similar stories]
In this case, reliance on computer modelling could have detected the error. Knowing the mass and volume of the sub, modelling the buoyancy is simple. Keeping the model updated and verifying that it can still float should be required for a project of this nature.
Unfortunately most commercial engineering software is probably too general to know to check whether the sub floats or not automatically. That would require reading the engineer's mind. That check would need to be at least turned on by someone, if not coded into the software itself. And my experience is that few even consider turning these checks on.
If you're writing the code yourself, then it should be easy to have all sorts of automated checks. Another HNer recommended using assertions to check that the sub floats when you want it to and sinks when you want it to: https://news.ycombinator.com/item?id=22396621
Important calculations like these ought be done in at least two independent ways. For example, a detailed version accounting for every nook and cranny and a reality check version that takes things in the large just to make sure both results pretty close. If they're not, then have someone else try again. Having said that, there's the famous (pretty simple) SAT question that all test takers got wrong, even the test setters. Oh, except there was that one guy who wrote "the right answer is not here." (It was a circle rolling around a bigger circle.)
There was a big story in Sweden in 2008-2009 when a bridge collapsed due to a faulty drawing, killing one person and badly injuring two others. Two were construction workers on the bridge (killed and injured respectively) and a person in a car passing under the bridge (injured).
The engineer had specified a 7 mm thickness of a part of a metal beam that was supposed to be 25 mm. She argued that the manufacturer should have realized that the drawing was a preliminary draft because it was undated, it was missing other information, and was "clearly" underdimensioned, but it was still manufactured and put in the bridge without anyone noticing until the collapse.
The reason the story was extra covered in engineering magazines at the time was because the prosecutor argued that the responsible design engineer should go to prison, and three others (the owner of the design firm, a manager at the construction firm, and a coordinator who led the project) should get suspended sentences. No one from the manufacturing firm was prosecuted.
The result of the trial was that the designer got a suspended sentence and a fine, and the design company was fined 1.5 million SEK. All others were freed.
Most people agreed that the lack of process was to blame (the design should have been checked by someone else at the design firm, they should have clear markings on preliminary drafts, things shouldn't be manufactured without a formal handshake, and so on), but it's still an interesting question where the line should be on personal responsibility.
Some forms of negligence in the workplace should be criminal, especially if it results in people dying, but it's pretty harsh if an individual engineer is personally responsible for putting a 7 instead of 25 (possibly due to copy-paste of work in progress) in a drawing or in a piece of code. It's hard to draw a clear line of the responsibility between the individual engineer and the company that they work for, which should have processes to catch these things.
I come to think of a other highly publicised case in Sweden, where two employees were to clean a lime kiln and got hot steam and hot limestone dropped on them, one died and one severly burned. This case I think showed more obvious neglect from responsible managers, but no prison sentence was given here either.
Two days ago I discovered that Node-based server that I wrote converts user money balance from a number to the string somewhere completely silently, and after that money added to it is treated like string concatenation instead of integer addition. Made me shiver when I realized what was happening, good thing we're not live yet.
(It's not real money, I'm a game developer, and would obviously never use JS's floats for real currency. But it still would be a product-killing bug).
One time I had a shipment that was a few million JPY and just out of habit I put the currency in as USD, fortunately, the system said "hey dummy, that duty sure seems high" at which point I noticed the mistake and was like ahhhhhhh that could have been really bad. I was paranoid for like a week after that.
I can't imagine being part of an error like this or that one Mars mission, oof.
In this case, the result seemed odd, so I checked it by stepping away from the spreadsheets and FEM software, and simplifying the problem down to a quick hand calc. But if I hadn’t done this exact exercise on dozens of buildings, I might not know what ‘odd’ looks like.
Luckily, in this field, you can usually come up with a quick check that gets you within +/- 20% of the more sophisticated answer. These kind of sanity checks should be baked into the design process, but unfortunately not everyone seems to do them.
> These kind of sanity checks should be baked into the design process, but unfortunately not everyone seems to do them.
Could you please writeup an essay on this and publish it in an industry journal? This should be required, it should be a social and cultural norm across all engineering professions.
The story of saving the Citicorp Center to uncontrolled oscillations from an incoming storm is fascinating.
Apparently somebody in the calculations made a mistake in the very beginning and nobody paid attention to review the calculations.
This is common in projects of such scale, where everyone is so specialised that no one has a general idea of the whole project and could thus have raised a "something feels wrong" feeling.
What about building a huge skyscraper in New York City and later being told by an undergraduate student that it's got a critical design flaw that could make the whole thing collapse?
> LeMessurier realized that a major storm could cause a blackout and render the tuned mass damper inoperable. Without the tuned mass damper, LeMessurier calculated that a storm powerful enough to take out the building his New York every 16 years.
> In other words, for every year Citicorp Center was standing, there was about a 1-in-16 chance that it would collapse.
They dropped a factor here. Not every storm strong enough to damage the building without the damper would cause a blackout first.
Exactly! There is no real owner/leader of such project. The clerks just process their checklists. If a clerk finds some calculations, it’s all good, because it was done the right way according the checklist. Clerk has with high probability no knowledge to exam results of the calculations.
I just experienced same situation with documentation of the project. There was none at very beginning, clerks were mad, there must be one according checklist. Documentation was written later, absolutely useless document, but it was here and it was good enough for the checklist.
Yes totally agree, the clerks/integrators/chief engineers/etc. handle a lot of data and it is easy to get lost. That is why all the checklists must also contain bounds. Upper and lower limits of the calculated/measured physical quantities in question.
Any good tricks for catching these sorts of errors?
I'm a theoretical fluid dynamicist. When doing math I'll typically check the dimensions, order of magnitude, bounds, functional dependencies (anything missing or something there that shouldn't be?), overall trends of an equation, and limiting cases. For solutions to equations it's useful to also plug the apparent solution back into the equation. Another check ideally comes after waiting a while so that my mind is fresh. Sometimes what I did seems dumb by then.
Coding up the math is useful too. I've found that I can never add too many assertions. I might have caught the error described in the article as I often use assertions to check bounds.
Peer review isn't as reliable as I'd like. When it comes to math, most people seem to "rubber stamp" it. I can recall reviewing a paper that had math problems that were fairly obvious to me. For some reason I got the reports from the other reviewers, and I was the only one to notice...
Intuitively, only store numbers in their natural dimension. The same thing regularly happens in finance where interest rates (1.5% ie 0.015) is often stored as percent (1.5) in databases. And of course you never know what conventions people take, particularly in these days of super low interest rates. Then you have spreads which can either be stored in basis points (10bps), percent (0.1%) or natural (0.001). And currency of amounts, etc.
All table columns should also have unit in names, whether in charts, tables, database or variable names.
> All table columns should also have unit in names, whether in charts, tables, database or variable names.
Oh yes. That would mitigate a lot of stupidity. Imagine if math in programming and Excel had to have both a) always visible/known units, b) unit propagation and c) unit checks. So if A is seconds and B is kilograms, the compiler/spreadsheet wouldn't let you write C = A + B.
In my experience reviewing engineering calculations, I end up doing a lot of work either by hand or in something that recognizes scientific/engineering units like MathCad or SMath and that also displays the equations.
I have marked up 100+ page reports, checking every number, in red pen and doing the math on a handheld calculator, checking the page references for all referred equations.
The best habit I had, from reviewing others' work, is to give my own work the same treatment.
This is the only way I got a long engineering report out in a reasonable amount of time: by catching my own mistakes by actually printing it out and going through it with a fine tooth comb.
It takes time, but not as long as you would think in many cases.
I've seen others just kind of skim a document and read a few words and see the shape of the equations, then sign off on it for non safety parts. Even then, you can easily lose lots of money on simple mistakes.
I've also seen inherited spreadsheets that went through several hands, with magical things added, that the past two people using it never checked.
Some people never learn to really go through and check.
Have someone else independently code up / compute the number from scratch and compare against it. Likely you'll come up with a different number, then figure out which one is right (you'll probably find both are still wrong).
If you're doing some sort of analysis, there is only one meaningful way to check it - have another equally capable person do it without looking at your work and see if they get the same answer. If (when) they don't, they can look into why. Maybe people in your experience don't do that for math or code, but it's a clear bright line standard, either you do it, because the task is important to get right or you don't, because it doesn't really matter.
I have seen where too many layers of checklists and verification just make people assume another layer will catch errors and then nobody checks thoroughly.
> Maybe people in your experience don't do that for math or code, but it's a clear bright line standard, either you do it, because the task is important to get right or you don't, because it doesn't really matter.
> I have seen where too many layers of checklists and verification just make people assume another layer will catch errors and then nobody checks thoroughly.
Yes, my experience is that few people are willing to do this. The important question then is how to find people willing to redo the work independently.
I haven't found anyone who seems interested in doing this or able to dedicate time to this.
"how to find people willing to redo the work independently"
Taking laborious manual procedures and automating them so as to flush out all the details that were overlooked or gotten incorrect is what I do at the moment.
Assertions is the way to go. In every domain there are invariants. Check for those invariants after any update to your model. For example - in the submarine case, the two invariants would be - "submarine can float" and "submarine can sink". No matter what your submarine design is, the final product should be able to float and sink as needed and therefore these assertions should always be true. If at any time either of these two invariants are violated, you've got a problem in your design.
Those errors are the ones that you catch by experience. That means making a comparison with previous done (and well finished) jobs. a When I refer only at theory I feel unsafe for the final result, so I try to search for similar jobs or asks experienced workers. I am sure that experienced workers also in this case of the submarine were complaining that was too heavy. Experienced workers base their "calculations" work on previous done jobs and this is really good to give an idea about if is going to work.
The real joke is that Navantia which is building this submarine actually stole the intellectual property from their French counterpart during a previous joint venture [0] and then managed to fail building a submarine with it.
Having worked in the field, the idea that someone can just misplace a decimal point is laughable. It highlights a complete lack of quality control and engineering know-how from Naventia.
The issue here is not the calculation. Mistakes happen. The issue is procedural. You can't hope to build submarines if you have no process in place to systematically catch this kind of error. That's true at every level from conception to manufacturing.
That they ended up having to go see the Americans because the French wouldn't help them say a lot about European defense integration however.
IDK why this keeps coming again and again. I'm sadly in my phone but this is actually false. The client asked for changes mid construction phase so they had to redesign it. S81 and following ones already include this changes.
I'm not able to read your link but it's probably part of this FUD that comes over and over. Navantia and DCNS had a joint venture, but the requirements of the French and Spanish clients diverged so Navantia decided to go its own way. What you can find is that DCNS tried to sue Navantia back in 09, but AFAIK nothing came out of it (at least yet).
Navantia didn't have experience designing submarines from scratch but they have plenty with civil and military ships. The fact that such "laughable" mistake was apparently just made in this project is interesting to say the least.
Not to mention that it's not the first time that Navantia faces accusations without proof.
IMO this looks like DCNS trying to block their competitor, that not only got its new founded expertise in submarines from you but also knows your commercial channels, relationships etc.
There are no French or Spanish clients. The Scorpene submarines were built strictly to be exported by the joint venture. Spain was expected to buy some, France wasn't.
As far as I know, France only uses nuclear propulsion for its submarine and doesn't export the submarines used by its navy.
> AFAIK nothing came out of it (at least yet)
It was settled. Navantia entirely gave up the export rights of the Scorpene. For a submarine built to be exported, that seems pretty damning to me.
> Navantia didn't have experience designing submarines from scratch but they have plenty with civil and military ships.
Submarines are a lot more complicated to build.
Most of the Scorpene engineering came from France. Navantia clearly wasn't ready to be left on its own.
> There are no French or Spanish clients. The Scorpene submarines were built strictly to be exported by the joint venture. Spain was expected to buy some, France wasn't.
Not french and spanish clients for Scorpene, but a Spanish client for Navantia is the Spanish Navy for example, and DCNS had their own.
> It was settled. Navantia entirely gave up the export rights of the Scorpene. For a submarine built to be exported, that seems pretty damning to me.
Hmm, do you have any source, I've tried but didn't find anything.
> Submarines are a lot more complicated to build.
> Most of the Scorpene engineering came from France. Navantia clearly wasn't ready to be left on its own.
That's true, although my point is that they are not some bunch of amateurs that misplaced a decimal. They've been designing ships for a long time, and building submarines at least from the 60s, so the decimal displacing is probably FUD, specially when the client changed requirements mid-construction phase, and the source of such mistake is Rafael Bardaji, who seems that has never been a Spanish Official but conservative and later far-right advisor, and honestly, is quite a fishy source.
So the spanish company was building a submarine for their own clients, they allied with the americans Lockheed Martin for electronics, and the french company parted ways avec le coeur brisé.
This things happen all the time, and nous aimons nous voisins du Nord plus qu'ils le croient. We enjoy all those silly jokes about spaniards in any case.
Dcns (now naval group) is owned by the french state, it's technology is french property.
All the french navy is built by this company, so it accumulates a lot of know how.
The Isaac Peral, the first in a new class of diesel-electric submarines, was nearly completed when engineers discovered the problem. A U.S. Navy contractor in Connecticut, Electric Boat, has signed a deal to help the Spanish Defence Ministry find ways to slim down the 2,200-ton submarine.
In case you were wondering (I was) - the entire sub is 2,200 tons.
Slightly off topic, but my dad worked on submarine engineering at that E.B. facility, and in related subcontractor / government roles, in the 1970s - 1990s.
He had lots of stories about major oversights making it WAY farther into the production process than you'd expect.
Yes. What the slide rule era didn't have was the internet.
EDIT: Here's a 1992 GAO acquisition study[1] on the AN/BSY-1 combat system whose development started in 1980 and was an absolute clusterfuck for the Navy in the decade that followed, as history would have it. I originally learned of this disaster from a former sailor (and now good friend) who participated in sea trials for one of the submarines that this system was installed in. By his account, the submarine had had its hull recut just to get the system installed and routed properly, and it never really worked as intended. The study vaguely alludes to this:
> As a result, the shipbuilder was eventually faced with rip-outs, recutting of hulls, and rerouting of cables.
Electric Boat was the prime and responsible for system integration. They've had their fair share of glorious fuckups to the tune of $82.4MM in late 80s dollars, so it's funny to read the hero part on how they've "helped other countries with their submarine programs".
In the right professional circles, stories like these quite literally abound. Now what's the probability of learning about them without the internet?
"Spain’s Newest Submarine is Too Big For Port" (July 2018)
"In 2013, ten years after the first boat was ordered, authorities detected a critical flaw in the design: The submarines were 75 to 100 tons heavier than anticipated. The submarines could dive but there was some question as to whether they could reliably surface again."
"The shipbuilder, Navantia, immediately suspended construction of the subs and called in U.S. submarine builder Electric Boat to solve the buoyancy issue. EB recommended the shipyards lengthen the S-80s by thirty feet, stretching it from 232 feet to 265 feet..."
"Now, fifteen years after the submarines were ordered there is a new problem: The first submarine, Isaac Peral, is too large to fit in its port at Cartagena. Authorities say the port will have to be enlarged and dredged to fit the updated design. It’s odd that Spanish authorities are suddenly discovering the problem considering the submarines were actually built in Cartagena and everyone involved has had five years to ponder the ramifications of enlarging the sub’s design."
shouldnt something be tracking the BOM and thus calculating the weight of everything being added in. It seems like catching this could even be automated.
The company I worked for had a fairly strict policy for that kind of thing. Design Review was an integral part of their engineering culture. Involved lots of heavy-duty folks, going over your design with a fine-toothed comb.
A junior engineer would present the design from their team, and everyone would take turns picking up brickbats and having at it.
Unburying this old fail just right now is really interesting. In the same week when the minister of defense Margarita Robles is about to close several big contracts including S80 submarines, F110 frigates and MGCS chariots with several companies.
I would not discard that the decimal event happened either only in the mind of the journalist or as a scapegoat history after a major change of specifications in the middle of the project. Honestly, smells a lot like FUD
On they plus side, they won't have nearly as much trouble as they otherwise would have had to make it go down.
On a boat like this 70 tons isn't all that much, but it is outside the margin of error in terms of what you would correct with the ballast tanks. This will be expensive and tricky to fix without impacting other systems.
In slide-rule world this did not happen. Slide rule gave you 3 numbers, but you were not always sure of the tailing zeros, so you checked the sanity of the scale all the time.
Of course it did, you just didn't hear about it because either it wasn't publicised at all, or there wasn't enough tracking to find out where the error had originally been introduced, or nobody shifted through literal tons of papers to find out, or if they did there was no internet and it'd only appeared as a note on page 7 in the "journal of overweight submarines", circulation 500 copies.
This sort of errors is pretty much why double-entry bookkeeping was invented.
Morocco constantly sometimes makes claims over territories that don't belong them. Usually small islands of no use to anyone. Probably a play on nationalist feelings, quite stupid. Despite this theater, Morocco and Spain have a benefitial economic relationship.
Something similar could be said about the UK. But they are actually more aggressive and Gibraltar harmful to the surrounding area (smuggling and off-shore societies).
UK makes claims about the surrounding waters (which as far as I know, they have no claim to) and it's very bullish about enforcing that right. Given the nationalistic derive UK is taking and the change in status because of Brexit, it is very likely that tensions will rise.
Gibraltar has been UK territory for longer than the US has existed. By all means grumble about it, but if you grumble about morocco claiming Melilla in the same breath and don’t see the hypocrisy you really need to try to take an objective look at your views.
I invite you to reread my comment, as I do not talk about whether the UK having Gibraltar is legitimate or not. I simply pointed out that as far as I know and as per the treaty, it doesn't have surrounding waters. This is not something the UK agrees with and it is a source of constant tensions.
If you want to draw comparisons with Melilla, it was an abandoned military plaza that was simply taken by the Castilian crown in the 15th century. It is not a colony, unlike Gibraltar (according to ONU). Morocco didn't exist as a clear political entity back then, so I am not sure what claims would they have to it.
And my personal views are that if Spain could easily get rid of Ceuta and Melilla most spaniards politicians would sign right of the bat. They are just a sinkhole of money, source of diplomatic tensions and immigration problems.
Not OP, so not sure what they had in mind exactly, but Spain has several enclaves on Morocco's northern coast [1], including two cities of ~85k people each. Morocco is unhappy about this and has asked for sovereignty. This dispute has been armed relatively recently: https://en.wikipedia.org/wiki/Perejil_Island_crisis
To add some facts that seems to be forgotten now: Spain was going to give up its last colony in Africa in 1975 (Western Sahara), following UN's resolutions, when Morocco invaded it taking advantage of the dictator's terminal illness.
Think about the Strait of Gibraltar, Ceuta, Melilla, and Canary Islands:
- Gibraltar is a potential choke point for the mediterranean if something happens in the Suez Canal.
- 3 nations (UK, Morocco, and Spain) have territorial waters on the strait. All three are NATO members, but the UK left the EU last month and cannot be considered as an ally for it gibraltar politics anymore...
- Marocco is claiming more territry waters both in the ocean side near canary islands, and in small islands Ceuta and Melilla. See Perejil Island incident (https://en.wikipedia.org/wiki/Perejil_Island).
Also, Spain sells a lot of weapons. Submarines are a good addition to the catalog.
> the UK left the EU last month and cannot be considered as an ally for it gibraltar politics anymore...
I'm sorry, but this is fairly laughable - if you had any idea of the relationship between Spain and the UK over the last n decades, you'd know that Spain is the belligerant 'partner' in this relationship and has, for all but about 5 years around the turn of the millennium been apt to abuse the Gibraltar question as a temporary salve for whatever domestic political problem it is distracting its populace from.
It would be simpler to say that Spain is historically a naval power and has a lot of coastline, and leave it at that... .
- The nation of Morocco did not exist when Ceuta and Melilla became Portuguese and Spanish respectively, so I assume the only claim Morocco could have would be based on territorial integrity. They were also never part of the Spanish protectorate in Morocco, and are not recognised by the United Nations as colonies.
- On the other hand, Gibraltar has been classified by the United Nations as a Non-Self-Governing Territory subject to decolonisation. Furthermore, it considers Gibraltarians mere "settlers" and as such does not recognise their right to self-determination, and urges the governments of Spain and the UK to negotiate the decolonisation process between themselves.
Whether the UN's position is correct or not is whole different matter, my point is that given the UN's point of view, Spain's position on Gibraltar and Ceuta/Melilla is not contradictory.
Regarding the Canary Islands and "other African colonies", I'm not sure what you are referring to as the only inhabited territories Spain has in Africa are Ceuta, Melilla and the Canary Islands, and as far as I know nobody is claiming the Canary Islands.
This is not true. Norwegian investigations concluded that there were no technical defects involved in this incident, discharging Navantia.
Quoting your source: The watertight condition of the ship was guaranteed by the 13 watertight bulkheads. Seven compartments were damaged as a result of the collision but initially the ship remained afloat. No one intervened to break the chain of errors.
> The whole point of stuffing boxes was to contain water in damage like this.
That is not the point of stuffing boxes. They keep water out when the shaft is not turning and lubricate the the shaft with sea water when it is. A grounding can bend the shaft easily letting lots of water in.
The AIBN considers:the vessel's lack of watertight integrity to be a safety issue relating to Nansen-class frigates and therefore issues the following two safety alerts
I get the whole "home grown industry" and sovereign production of defense and all that, but just like when Australia tried to build its own submarines, really you're better off just buying them off the shelf from your allies who've already been ploughing billions into design and development of the things for decades.
Spain has an active defense shipbuilding industry (e. Eg scorpene submarine), why would they buy them from someone else ? And these kind of program secure a lot of jobs, when you buy from outside you secure those jobs abroad.
From what I read about american airplanes they have this days software that call home and send tracking data, plus imagine some diplomatic incident(like a tweet of some extradition case) and you now have to worry if your submarines will be remotely bricked ... so yeah if you buy important crap from others make sure you can control the software too.
Just yesterday I made a trivial error crunching some numbers for when a construction site could turn off their dewatering. Quite the opposite of this submarine, my structure would be 1000s of kips too buoyant and could potentially float up out of its excavation due to hydrostatic forces. A mistake like going unnoticed could easily be a career-ender.