Comment by kevin_nisbet | Hacker News Reader

Comment by kevin_nisbet | Hacker News Reader

Parent Full thread

While in general I agree with you.

I'd just like to mention one counter argument, which the rigidity of the code in some cases will help to protect against developers from using cheaper materials, or new materials that on paper seem better, but in reality may not be better.

An example I've heard of but am having trouble searching for the exact name, is in condo buildings here in Canada they started using a new piping material, to do the water deliver inside the units. The problem was, while I think in theory the material was better, it has the property that when it fails it fails catastrophically, due to either a manufacturing or installation defect. So instead of just having a small pinhole leak, the piping will split when compromised, and you have many thousands of dollars in damages to multiple units. Buildings with this material can no longer be insured in Canada.

So I don't know that I would really trust giving developers a wide latitude in selecting materials, even if they sound great on paper.

I don't know what the solution is, because I agree we need to be more flexible, and have a way to introduce new and better technologies, but we also have to be diligent, in ensuring that the new technologies and processes work the way we expect them to, and have the desired effects.

In Canada we use what is called limit states design and it handles the sudden failure vs gradual failure already. Essentially if you want to use a member or a material that suddenly fails you must design it to fail one tenth as often. In practice engineers go even more overboard because nobody wants blood on their hands and gradual failure greatly decreases the odds of that from happening.

In terms of construction materials in condos failing (glass and piping) I actually put the blame for those two failures on the individual testing companies. Engineers in Canada don't understand statistics properly because most aren't taught materials testing properly in university. But we don't live in a utopia either. It is expected that things will go wrong sometimes and that we'll have to update our building codes to address those shortcomings. My issue with Canada is that we don't have a (or at least I'm not aware of there being a) structured materials testing code.

Sure we'll regulate materials and connections for our main structural elements, but there will always be that last mile where someone will want to do something weird (like build part of the building under railway or if someone over bent some reinforcement bar and you need to authorize them rolling it back) and you're pretty much out on your own once you come to those scenarios. It's doubly unfortunate because most of the engineers that just look things up in tables are basically helpless because they don't remember or use the basics day-to-day. So they end up being extremely conservative - unnecessarily wasting money and material for all of us.

I don't practice in Canada, but my understanding was that the CSA is an analogue to the ASTM codes here in America (and internationally apparently, if you believe their name change). ASTM codes are very thorough when it comes to materials testing. I believe Europe has a similar standard.

I don't understand how your third paragraph's thrust follows from your second paragraph - what does materials testing have to do with site specific (railway) or field changes (bent rebar)?

How long did you practice in Canada? Your viewpoint of engineers meshes well with the opinion that I've heard from a lot of junior level engineers who are just making the adjustment to a mid-level position but are still interacting with the lower level staff who are, as you say, typically helpless. They are supposed to be - they are still learning.

Only about 3 years before I got fed up and left. I'll fully admit I didn't get that involved with materials testing; and perhaps the firm I was with was substandard in this regard; but I don't think so. When I looked into the falling glass in Toronto I learned that they only tested a very small number of fasteners. I don't recall the number, but it was trivial statistics to prove that for the number of glass panes going up in Toronto they didn't have a large enough "n".

The two examples I gave were two examples that I dealt with personally. I was extremely dismayed at the rigour the firm I was at used. To test the bent rebar I think we used a sample size of 6 and then tested to failure. For the railways example they just used the weight of the train. Then when I reviewed the designs and brought up that the train could apply the breaks and thereby increase the downward force they just multiplied everything by 2.

In my experience the low level staff was useless, with a couple people that knew what they were doing. The medium level staff had two groups of people, the people that still knew advanced math and the people that got good at AutoCAD, and the senior people, while good at sales or general guidance; had basically completely forgotten all but the most basic structural engineering principles. I've literally had to explain crushing vs bending moment to a 20 year structural engineer before. I've (accidentally) designed a structure that was already designed by a senior person that forgot to put it in the tracking system. I used one sixth the steel and mine could handle more load.

I will grant you, however, that I may have just been at a substandard firm. We had some large projects, but we weren't designing new skyscrapers or mega-structures.

For someone who is self-admittedly not knowledgeable about the testing requirements, you seem very certain of your conclusions about this Toronto falling glass problem. Testing of components is usually by the manufacturer and it is their responsibility to provide a product that meets the requirements of the design. This is not a problem from the design side and is very difficult to prevent without the engineer being onerous with his requirements to a point that no engineer is really willing to go to.

I'm not familiar with the specifics of your rebar example so I can't comment. Your example with the train makes no sense - the design loading for railway is codified in the design manual (AREMA in the USA) and includes dynamic forces. Braking forces are applied longitudinally to the track so unless you were in a curve there is no downward force. I find it hard to believe that your boss agreed with a fictitious force and then just multiplied everything by 2 to get around it.

Your opinion on your colleagues is concerning to me and is probably more indicative of your lack of experience than the other staff's incompetence. Your experience reads like someone suffering from 200th hour syndrome, I wouldn't be surprised that if you stuck with it another 3 years you would have realized your initial impressions were way off base. At worst, it sounds like you may have been working at a firm that did commodity work and didn't attract top tier talent. If you are as good as you seem to think you are then you should have jumped ship when you got "fed up".

I don't intend for this post to sound dismissive but it will probably come off that way.

As an aside, knowledge of advanced math is not necessary for structural engineering in my opinion, nor is it common.

"Braking forces are applied longitudinally to the track so unless you were in a curve there is no downward force." Why is that? Is it because the cars behind are pulling on it and keeping the usual forward weight transfer from happening?

Think of a motorcycle doing a "stoppie" i.e. read wheel is in the air under braking, all weight is on front wheel.

This is hard to describe without being pedantic and without being able to draw but I will attempt.

A motorcycle performing a stoppie experiences rotation because the inertial force couples with the braking force to create a moment about the front axle of the bike (this isn't technically correct language but you get the gist). While this idea holds true for the train, we have to take into account the differences in mass and contact between the two systems. Train cars typically ride on more than 2 axles and this provides stability from front or rear tipping. Train cars are also typically very heavy meaning that the braking force is not sufficient to 1) move the center of forces of the system ahead of the front axles and 2) tip the car. Increases in load because of this are, therefore, not sufficient to double the load on the front axle as you would see with a stoppie.

In general I agree with the idea that is put forth; however, it is important to note that what we are discussing is the BRAKING force. The inertial forces that result in differential axle loads is not a braking force (certainly, it is a result of braking in this case but this force also exists when a train begins pulling). These loads are DYNAMIC loads and are already considered in another part of the analysis. Dynamic loads also include consideration for bumps, etc. Because of this, the code is explicit that braking forces are applied only longitudinal to the track so that the forces are not counted twice.

Thanks for responding, and I'm very happy that there are people like you out there; but trust me when I say that despite my poor recollection from the time I practiced, my fundamental point is not wrong. Most engineers I've worked with in Canada are not to be trusted with advanced design. If you disagree I'd like to really talk to you about it because I felt like I was surrounded by people that had no idea what was going on and I would really like to be proven otherwise.

I'm not competent to explain why Tacoma Narrows failed, but it obviously wasn't up to design load. The existence of large sheets of glass falling off buildings and endangering people from multiple different installations strongly indicates that someone botched something.

What is 200th hour syndrome?

The 200th hour syndrome refers to when a pilot has reached their 200th hour in the air and their confidence in their abilities is enough that they begin to get careless:

"Enough experience to be confident, enough to screw up real good." is how a nice TV show put it.

You're right, one reason engineers stay conservative is if the new ideas have problems that show up later, the engineer gets the blame. Hence he sticks with "nobody ever got fired for buying IBM".

For example, when my house was built I wanted something better than fiberglas insulation. After research, I settled on icynene foam. The contractor refused to use it, because he'd be to blame if it went wrong, it would be an enormously expensive retrofit. He finally agreed after I formally agreed to accept the risk and not blame him.

15 years in, and the icynene has been great. No troubles at all.

Reply on news.ycombinator.com