• (disco)

Em... huh?

• (disco)

TL;DR: A guy had a long interview process. In the seemingly last part, he was asked a custom Interview2.0-esque question that he couldn't answer. It is implied he wasn't hired based on that alone. Stay tuned for part 2 - with the point/funny bit.

• (disco)

Great, another company that thinks that asking you to guess an answer using Fermi Estimation (FE)(1) tells them anything useful.

Hint: no, it doesn't. If it is known that your company asks FE questions - this information is available to the whole world seconds after the first candidate gets home - candidates will prepare themselves, and while preparedness is, perhaps, something you'd like to see, you don't really learn anything else about them.

(1) For those who don't know, FE is a technique where you derive a simple equation for the result you're seeking (e.g. the flow out must equal, subject to certain losses, the flow in (rain), so average rainfall times area of basin gives volume of rain, and therefore the flow out) and then estimate (no, you wildly guess(2)) the various parameters (rain amount, area of basin), hoping that the various wildnesses will cancel out to a certain extent (e.g. this one has N zeroes too many, that one N too few, etc.)

(2) Lesson for the incautious: "estimate" is a fancy synonym for "(informed) guess".

• (disco)

Well, after a bit of reading, I though it is impossible that they asked such a question, so I was wondering if it's a trap. What if the answer is ZERO? And - yes - the Atlantic Ocean is an ocean, not a sea.

• (disco) in reply to Steve_The_Cynic

Rainfall shmainfall. I stopped halfway through the article and gave it a go, using width of river at mouth x depth at same x flow rate, plugging in what seemed reasonable (and easy to work with) guesses of "1 kilometre" (it's a big-assed river), "10 metres" (about 30 feet) and "5km/hr" (moderate walking speed, about 3mph).

That gave me an annual outflow of 4.38e11 cubic metres, which translated into 'Merican is about 1.2e14 gallons -- about one order of magnitude less than the correct answer of 1.9e15 gallons. I didn't think that was too shabby.

As it turns out, I underestimated the width of the mouth by about an order of magnitude: according to Wikipedia, it's 15km wide (if you're measuring the main channel).

• (disco)

Well, now I know. 1900 trillion. File and forget (or remember).

Oh, and the other answers are "three socks" and "the last day".

• (disco)
Carlos had been **paging** everything he knew about Amazon Web Services into his brain
He should have infiniscroll’d that instead.
• (disco)

Been there, done that: how many calories are stored at your local supermarket. The only information given is that "supermarkets store some seven days of stock". We're seven days away from a famine!

This was particularly hard for me since I've never have had to pay attention to my calories intake.

• (disco)

Reminds me of one that I got. "How many manhole covers are there in SF city?" Eff if I know and what do I care? Then I found out there entry level paid in the low 40s and an internship a couple cities away paid in the mid 50s and I didn't care that they found someone else for the position. Big companies FTW.

• (disco) in reply to SheriffFatman
SheriffFatman:
Rainfall shmainfall. I stopped halfway through the article and gave it a go, using width of river at mouth x depth at same x flow rate, plugging in what seemed reasonable (and easy to work with) guesses of "1 kilometre" (it's a big-assed river), "10 metres" (about 30 feet) and "5km/hr" (moderate walking speed, about 3mph).
It is indeed a big-assed river, but (as you found out) it ain't anywhere near that small. ;)
• (disco) in reply to Marcin_Kozlowski

Next up, in part two: Experience the thrill and suspense of Carlos' journey all across Spain seeking out former other interviewees of that company in search of the burning answer of how much water does flow in the Amazon river.

• (disco) in reply to Steve_The_Cynic

Indeed. My woeful underestimate is made even more embarrassing by the fact that I live a few miles away from a river wider than that.

• (disco)

"How much water flows annually from the Amazon River into the sea?"

My answer: [spoiler]All of it.[/spoiler]

• (disco) in reply to Sizik

duh!!! NONE, the river flows into the OCEAN

• (disco) in reply to maslan

How much do you want to flow into the sea, and how much of a budget are you going to give me to make it happen?

(i tend to roll with the ocean/sea confusion as i've yet to meat a manager that actually got the distinction

• (disco) in reply to maslan

The question didn't say "a sea", it said "the sea", and "the sea" is a common way to refer to the entire interconnected body of liquid which covers 70% of the planet, which is composed of sea water and the surface of which is at sea level.

• (disco) in reply to SheriffFatman
SheriffFatman:
"the sea" [...] is composed of **sea** water and the surface of which is at **sea** level.

Those are creepy coincidences. *shudder*

• (disco) in reply to SheriffFatman

Another Fermi approach: the Amazon basin covers about half of the South American continent, so 2% of the surface area of the planet (7 continents, world is 70% ocean). The circumference of the planet is 40000km, so (approximating pi as 4) the Amazon drains 8 million square kilometres. Guessing net annual rainfall (after evaporation) at 1m gives 8 trillion cubic metres, equal to 1800 trillion gallons.

Yeah, Fermi estimation is unreasonably effective.

• (disco)

Correct answer: "Why should I waste company time and resources guesstimating at something I can look up on the internet in 5 seconds?"

That's a paraphrase of Einstein, in case that didn't sell it for you.

Followup: "Why would I want to work for a company that wastes time asking such dumb questions?"

• (disco)

All can be explained.

The subordinates (who clearly knew what Kyle was going to ask) had a system worked out where those who they actually wanted to hire for reasons unrelated to actual job skills (e.g. nepotism) were told in advance of the last interview exactly what was going to be asked and how to solve it satisfactorily. The receptionist didn't get the memo that this guy wasn't going to be one of those who knew the right answer, and thus took it upon herself to remind the hapless candidate what he was supposed to remember..

Kyle may well have known about this system, but also knew it provided a legally defensible means of not hiring qualified candidates for reasons of "cultural fit" and other similar nonsense.

• (disco) in reply to EatenByAGrue

Or Kyle wanted someone who could think on their feet and not panic when facing a problem with a large unknown at the heart of it, and who could come up with a halfway reasonable working assumption on which to proceed until more data was available ...

• (disco) in reply to SheriffFatman
SheriffFatman:
Or Kyle wanted someone who could think on their feet and not panic when facing a problem with a large unknown at the heart of it, and who could come up with a halfway reasonable working assumption on which to proceed until more data was available ...

well... if I was the interviewer asking the question and used the phrasing that the one in the article used then I would myself be a resource in front of the ineterviewer and would have looked up and memorized most the the salient facts about the amazon river.

continuing the assumption that i would ask such a question i would then expect the interviewee to do one of three things to earn an advance in the interview process:

1. Use a Fermi Approximation
2. Ask Me questions to gain additional information 1 (at -1 point)) Consult their pocket oracle of all information and ask for the answer from Siri or Google or.... Cortana? i forget the name of the windows phone assistant....

.... crap. the interviewee was one i interviewed and declined and then the anonymization changed it to a C-l;evel exec rather than a developer and added the secretary, and all the other details....

umm... if so, sorry to whomever submitted this story... i probably should ahve done a better job setting up the question to make it clearer what i expected

• (disco)

Amazed that people go to interviews without doing some research into what sort of questions interviewers are asking these days. These questions are all the rage: they don't care about the answer, just about whether you can work out how you would calculate it. You make a guess at any data you need, and if it's wildly wrong they do not care, because obviously you'd look it up if you were answering the question in a real-world situation (though you may score bonus points for saying how you'd find it).

If you somehow happen to be able to rattle off the right answer you fail - they want to know if you can reason, not what the actual answer is.

If you are clearly capable of getting to an answer (e.g. you mention some of the relevant values you need to calculate it) but throw a paddy over the fact that you don't actually have the data, you fail because you clearly did no preparation for the interview or you would have understood the purpose of the question.

• (disco) in reply to CarrieVS
CarrieVS:
These questions are all the rage: they don't care about the answer, just about whether you can work out how you would calculate it

that about sums up the point of all the questions i ask in interviews when i do them. I care how you think, not what you memorized. I can teach you the institutional knowledge you'll need. i can't teach you how to think, not and have you be an asset to the company in a reasonable amount of time.

• (disco) in reply to accalia
accalia:
i've yet to *meat* a manager that actually got the distinction
Freudian slip? And what kind of *meat* do you get from a manager when you *mete* out whatever sort of justice you find appropriate for someone you will *meet* once, probably?

;) ;)

• (disco) in reply to Steve_The_Cynic

I'm told it's quite similar to pork.

• (disco) in reply to Steve_The_Cynic

originally a typo, but i did notice before posting and decided to keep.

not that with my usual lack of proof reading on this forum you'd be reliably certain of this. ;-)

• (disco) in reply to accalia
accalia:
(i tend to roll with the ocean/sea confusion as i've yet to **meat** a manager that actually got the distinction

I tried to find a funny image to make light of the spelling error, and all I could find was this:

http://greenasathistle.files.wordpress.com/2007/04/meatface.jpg

• (disco) in reply to CarrieVS
CarrieVS:
These questions are all the rage: they don't care about the answer, just about whether you can work out how you would calculate it.

They are actually good questions. A bit overused, but very good questions. You want to give the interviewee a question that they likely cannot answer and see how they would go about solving the problem. Much better than the stupid FizzBuzz question that they likely got on a prior interview and have memorized the answer to. You want to give them something that they actually have to think about. And if they just sit there and stutter, they will likely take a while to get started on any problems that you give them that is out of their comfort zone.

A winning answer for me would be some pseudo-code that said something like:

``````(depth in the middle/2) * width of river * speed of flow * 24 * 365.24 =
``````

There would also need to be some conversions in there that would vary depending upon the units involved, especially for speed of flow to make it work with the units used to find out the cross-sectional area of the river, etc. If they figure the cross section in meters and the speed of flow is in meters per second then they would have to perform the calculations to get from meters per second to meters per hour.

The front page is really starting to suck lately. The interviewer obviously did not care that he could not calculate the annual flow with no data to work with. He just wanted him to show that he would be able to if he had the data to plug in and that he was able to think and reason off the cuff. The only WTF here is the over-dramatized writing style.

Honestly, the guy blew the interview when he asked:

"Uh... I understand what you're asking, I'm just not sure how I'm supposed to figure it out without any, you know, data. Like the average annual rainfall in the Amazon Basin, for example."

Because that is not going to get you anywhere. You will have a highly variable figure downstream in your calculation...evaporation. Right off the bat he was very far from the actual problem. The interviewer wanted to know how much water flowed from the Amazon to the sea and the interviewee is looking at the problem from back in the rain forest. Go to where the problem is.

• (disco) in reply to Polygeekery
Intercourse:
`(depth in the middle/2) * width of river * speed of flow * 24 * 365.24 =`

that assumes a rectilinear river. surely a more accurate estimate would be to assume an ecliptic cross section?

``````(π * (width of river) * (depth of river))/4 * (speed of flow) * (calculated unit to convert flow per unit time to flow per year)
``````
• (disco) in reply to accalia
accalia:
that assumes a rectilinear river. surely a more accurate estimate would be to assume an ecliptic cross section?

Meh, if we were wanting to pick nits, then sure. I would gladly accept a solution that assumed a rectilinear cross section calculation would be close enough. There are a lot of correct answers to this in my eyes. Average depth * width would work pretty well also. Measuring depth every X' and averaging that would be another acceptable answer. The iterations are almost infinite on this.

If we really wanted to pick nits, your calculation would not account for disruptions to laminar flow that would cause slower speed of flow along the shores and bottom than it would in the center at rate of fastest flow. I would not care that they missed a detail on a problem in a field that they do not have expertise in. I would just want them to get started.

"A good plan violently executed today is far and away better than a perfect plan tomorrow." General George S. Patton

• (disco) in reply to Polygeekery
Intercourse:
your calculation would not account for disruptions to laminar flow that would cause slower movement along the shores and bottom than it would in the center at rate of fastest flow.

granted, but then yours suffers from the same problem. ;-)

Intercourse:
I would gladly accept a solution that assumed a rectilinear cross section calculation would be close enough.

especially if we're doing a fermi aproximation anyway. ;-)

• (disco) in reply to accalia
accalia:
granted, but then yours suffers from the same problem.

Yeah, but I was not worried about absolute accuracy. ;) I just wanted to see positive movement on the problem towards a solution.

• (disco) in reply to Polygeekery
Intercourse:
I just wanted to see positive movement on the problem towards a solution

true.

also i guess i'm never getting that pedantry badge. that was about as pedantic as i get.

and i suppose that's a good thing.

• (disco) in reply to accalia
accalia:
also i guess i'm never getting that pedantry badge.

You will get there someday. - pats accalia's head -

• (disco) in reply to Polygeekery
Intercourse:
You will get there someday.
accalia:
and i suppose that's a good thing.

it would probably help if i tried harder.

or let Berk out of his cage in the back of my head. ... but i suspect if i did that Berk would overdo it a bit, also i just got him back in there from when he brok out a couple of week ago.

• (disco)

My question to the company would be the following:

If you have a search feature, and the results are formulated by a complicated mechanism that determines relevance. However, you have more results than you can show (limited to 15), and you cannot use a paginate feature. Is it acceptable to select your final set of results based on taking the topic name's hash value, and picking the greatest 15 hash values?

Because that's about as effective as the question you just asked me.

• (disco)

My professional response (and they would so nearly have earned the other one!) would be :

"You seem to have me confused with geographer... The algorithm for calculating the volume of water passing out of the river not being known to me, I would consult with someone qualified to provide such an algorithm, and then implement it.

However, as a developer, the algorithm wouldn't be "Look it up on Wikipedia", because the rate of flow might conceivably be dependent upon any number of factors. Similarly, I would implement the algorithm in such a manner (this is where OOP comes in) that the algorithm could easily be swapped out and replaced, should it transpire that it is actually in error, or should a more accurate / efficient algorithm come to our attention in the future.

Having done all that, I would then set about publicising my resume, because I doubt this question is in any way relevant to selecting a competent developer, unless of course you would care to explain... If not, I would prefer to work somewhere where the management were used to working with actual software developers."

• (disco) in reply to Polygeekery

So, to your mind, the reason the the interviewee for a programming position blew the interview didn't know much about hydrology?

I get that questions like this give you an insight into how an interviewee handles difficult questions with inadequate data. However, I don't think it makes a lot of sense to count them down for not being well versed in something far outside of their field, like not understanding the significance of evaporation on river outflow. Besides, he did say, "for example", which would tend to imply that wasn't the only variable he was concerned about.

• (disco)

In my very first interview out of college 14 years ago, the interviewer asked me this exact same question, except it was the Mississippi river, not the Amazon. I squirmed in dumbfounded shock as to why the hell I was even asked that question. And it wasn't just 30 seconds. He sat there with a smirk on his face for at least 5 minutes, refusing to accept "I have no idea" as an answer. He was such an asshole about it. Fuck him. All it did was reinforce the insecurities I already had about being qualified for ANY programming job I was interested in, since the only qualification that I had that matched the listed requirements was the BS in CS. I left that interview feeling very angry, insecure, depressed and convinced I'd never get a programming job.

To this day, my blood boils thinking about that experience.

• (disco) in reply to calliarcale
calliarcale:
So, to your mind, the reason the the interviewee for a programming position blew the interview didn't know much about hydrology?

No, but it would expose the exact line of thinking that this sort of question is designed to expose, he was thinking about the problem far away from where the solution actually lies.

To give it a parallel to the sort of issues that he might see, he might end up looking for a problem in the application logic when a performance issue actually lie in poor schema or indices on the database, etc. The interviewer wanted to know how a person would deal with solving the problem of how much water flows from the Amazon into the sea (or ocean, to be slightly more proper as was stated above). You could start calculating that with how much rain falls on the Amazon basin, but then you would need to load a hell of a lot more variables in to your equation such as: evaporation, absorption into the ground water supply (aquifers, etc), other various forms of water retention, evaporation in to the atmosphere, etc.

He was given a problem and then started looking for the solution thousands of miles away over millions of square miles of surface area. He failed, but could have redeemed himself had he realized where the solution to the problem actually is, near the mouth of the river.

• (disco) in reply to calliarcale

Signs of the times.

I can see the next interview for Presidency.

"I can see that you have valuable leadership experience, but I want to know.... can you organize a riot/protest?"

• (disco) in reply to Initech
Initech:
He sat there with a smirk on his face for at least 5 minutes, refusing to accept "I have no idea" as an answer. He was such an asshole about it. Fuck him.

Yeah, if that was his reaction and resolution, you would not want to work for such a shitty manager anyway.

"I can see that you have valuable leadership experience, but I want to know.... can you organize a riot/protest?"

that would lead to a more interesting political landscape if it were true.

I can see the next interview for Presidency.

"I can see that you have valuable leadership experience, but I want to know.... can you organize a riot/protest?"

Meh, as a libertarian, I would approve of this. Our political landscape needs a few more revolutionaries in it. :smile:

• (disco) in reply to Polygeekery
Intercourse:
Our political landscape needs a few more revolutionaries in it.

true, but we need to be careful. we don't need a frothing at the mouth radical revolutionary in charge of the biggest nuclear stockpile in the world (debatable. it depends on whether you believe the numbers that Russia says they have)

EDIT (and for that matter the numbers we say we have... )

• (disco) in reply to Polygeekery
Intercourse:
He was given a problem and then started looking for the solution thousands of miles away over millions of square miles of surface area. He failed, but could have redeemed himself had he realized where the solution to the problem actually is, near the mouth of the river.
If you want to know where someone would look for performance bottlenecks, then ask them where they'd look for performance bottlenecks. If someone answers this question by starting with average rainfall, and your takeaway is that he is likely to look for performance issues in all the wrong places, you are a fucking idiot.
• (disco) in reply to Polygeekery

No, that's basically begging the question (the "correct" meaning of "begging the question").

Sure, the mouth of the river can support x gallons per minute, but do x gallons per minute consistently flow? Or does the flow rate vary?

Is the flow mostly dependent on rain water, or does the flow come from snowy mountain tops where it would be greater in the months following winter?

There are so many variables.

All you've done is eliminate variables to give a simpler model, which could be way off if there is enough variance in the flow rate over the year.

• (disco) in reply to Initech

There are ways to determine if a person can find an answer to a problem outside of their comfort zone, without looking at the edge of the unknown universe to be sure you're outside their comfort zone.

Oh I see you don't have networking experience. Where would a bottleneck in a network be. Let me give you a little background information about the problem.