HackerRank Tries To Calculate Which US States Have The Best Developers (venturebeat.com)

An anonymous reader writes: Palo Alto-based HackerRank, which offers online programmng challenges, "dug into our data of about 450,000 unique U.S. developers to uncover which states are home to the best software engineers, and which pockets of the country have the highest rate of developer growth." Examining the 24 months from 2015 through the end of 2016, they calculated the average score for each state in eight programming-related domains. (Algorithms, data structures, functional programming, math, Java, Ruby, C++, and Python.) But it seems like low-population states would have fewer people taking the tests, meaning a disproportionate number of motivated and knowledgeable test takers could drastically skew the results. Sure enough, Wyoming -- with a population of just 584,153 -- has the smallest population of any U.S. state, but the site's second-highest average score, and the top score in three subject domains -- Ruby, data structures, and algorithms. And the District of Columbia -- population 681,170 -- has the highest average score for functional programming.

California, New York and Virginia still had the highest number of developers using the site, while Alaska, Wyoming and South Dakota not surprisingly had the least number of developers. But maybe the real take-away is that programmers are now becoming more distributed. HackerRank's announcement notes that the site "found growing developer communities and skilled developers all across the country. Previously, the highest concentrations of developers did not stray far from the tech hubs in California. Hawaii, Colorado, Virginia, and Nevada demonstrated the fastest growth in terms of developer activity on the HackerRank platform..." In addition, "we've had a noticeable uptick in customers across industries, from healthcare to retail and finance, with strong demand for identifying technical skills quickly."
Their conclucion? "Today, as the demand for developers goes beyond technology and as there is more opportunity to work remotely, there's a more distributed workforce of skilled developers across the nation, from the Rust Belt to the East Coast... Software developers aren't just attached to VCs, startups or Silicon Valley anymore."

Comments

State Doesn't Matter

[Anonymous Coward]

The best developers all attended the tremendous Trump Iniversity. If that's not on your resume, don't even *think* about applying to work at my company.

Re:State Doesn't Matter

[pem]

I wasn't able to attend it before it went under. Sad.

They can't

[phantomfive]

Hackerrank can't measure the best developers because the best developers don't waste time on hackerrank.

Furthermore, even if we assume that some of the best developers do spend time on hackerrank, the questions that hackerrank provides don't measure developer skill particularly well. A lot of them are more like tutorials.

Re:They can't

[h33t l4x0r]

The Hackerrank challenges are actually a good measure of how unemployed developers are because I can't imagine doing them if you have anything else going on. Seriously, they are fucking tedious.

Re:They can't

[plopez]

I hear the whine of someone who doesn't live in Wyoming.

Re:They can't

[Dutch Gun]

I trust your opinion on this matter completely just because of your name.

To be honest, even though I consider myself a competent programmer, I completely suck at the clever little programming challenges a lot of programmers seem to be good at. I think my brain just isn't wired that way. I'm pretty good at doing normal programming work, which typically involves a completely different skill set, in my opinion, along with plenty of time to experiment and reason out how to solve problems that come up. I tend to dread interviews, because they almost always measure and depend upon skill of the former type, and not the latter.

Re: They can't

[F.Ultra]

That's because in the real world you get a problem and have to create a solution while questioners like these are more: solve this problem using this specific technique.

Re:They can't

[s122604]

Being able to do ridiculous nonsensical problems with b-trees, it's the most important skill any developer can have... That and "rewrite this paragraph backwards, with every word ending in a consonant in pig-latin"... type of things..

Re:Easy

[Anonymous Coward]

In California it is possible to get paid even if you're incompetent. It draws the Java programmers here like flies to shit.

Simple

[war4peace]

The one with most Indians.

Restated: HackerRank Fails To Accurately Calculate

[MangoCats]

Garbage in, Garbage out.

Re:Easy

[AHuxley]

East Coast.
Second and third generation trusted CIA, NSA, FBI, Army staff.
Some good US education centres still based on merit.
Near the NAS, CIA for winning big contracts and secure face to face meetings.
A few more ms closer to France than from CA to France.
The access to the more direct EU/US communications links.
Near the all the interesting embassies. Diplomatic, banking and UN areas of the east coast.
CA is more the science and consumer hardware design centre of the USA.

Re:Easy

[Jake Griffin]

It's a pity that "Araksas" came in 34th. Poor Araksas.

Re:Easy

[Jake Griffin]

Correction: 39th

Re:Easy

[tempmpi]

Proper abstraction can actually help efficiency, e.g.: the templated c++ sort is a lot faster than the C qsort function because the compiler can optimize the code for each datatype and inline the compare function, while the C qsort has to use indirect calls to the compare function via function pointers and can't do static optimizations based on things such as size of the elements.

Efficiency is also often a question of using a proper algorithm. Most of the time that O(n) DP-algorithm coded in Python is going to be much faster than a O(2^n) bruteforce algorithm handcoded in assembly.

Re:HackerRank is useless

[tempmpi]

In the most popular challenges in Hackerrank speed is not an issue. You only need to be a speed champion if you are competing for the top 5% or so. Otherwise speed is not an issue but figuring out the right algorithm and being able to write it down without making tons of mistakes that require long debug sessions. Taking your time to think properly about the problem and then carefully writing correct code is actually the way to go, as debugging can easily use a lot more time than coding itself. Many of the easier tasks are just a few lines of code in C++, Python or Java without any code golfing.

Re:hackerrank does not correlate with good develop

[tempmpi]

I disagree. That someone is doing well in competitive programming (CP) type of questions does not tell you that that person is a good programmer. However, it also doesn't measure pointless stuff and proper knowledge of data structures and algorithms is a skill every good developer should have. CS101 can sometimes be pointless because many people cannot transfer this knowledge to any other problem. Just being able to tell that Quicksort is average O(nlog n) and worst-case O(n^2) is not useful if that doesn't mean anything to you. A bigger CPUs is not going to fix that O(n^3) algorithm that worked fine when the developer unit tested it with 10 elements, but somehow struggles when trying to run it on 100k elements in the production system.

Re: Easy

[F.Ultra]

Because qsort is the only possible sort algorithm in C...

New York would be higher

[DragonIV]

Except that I bring down the average. A lot.

Re:Easy

[__aaclcg7560]

Then I make a shit load of money cleaning up their mess.

That's the very first lesson I learned when I started working in IT Support. If you're the guy who loves to clean up messes, there's always work to be done.

"Best" is a semantic argument

[Tenebrousedge]

Hackerrank can't measure the best developers because the best developers don't waste time on hackerrank.

This is an insupportable statement.

the questions that hackerrank provides don't measure developer skill particularly well. A lot of them are more like tutorials.

And a lot of them are tricky dynamic programming problems with constraints designed to force you to find the correct algorithmic solution because any other will murder your CPU. I mean, perhaps you could sketch out a solution for this one? I was having a bit of a tough time finding the efficient solution, personally.

The simpler problems are of course far more tedious but I think it's probably not a bad way to pick up a functional programming language or two.

The more I read your post, the more silly I find it. You seem to be either unaware that there are difficult problems on this site, or you seem to doubt that other people can come up with hard programming challenges. Or, if you do not find any programming problems to be challenging, you must admit that it's possible to design challenges for lesser mortals. I may be singling myself out as the chaff rather than the wheat by saying so, but I disagree that HackerRank does not provide a reasonable measure of programming aptitude. However, even assuming that HackerRank is a terrible measure -- what better one did you have in mind?

Re:"Best" is a semantic argument

[phantomfive]

So you are one of those best programmers who is messing around on hackerrank. Good job being a top programmer.

If you want to find where the best programmers are, to begin with you'll do better with a random sampling, or even by attempting to attract the top programmers (like top coder, or ACM programming contest). Hackerrank markets to people who are looking for a job or are trying to build up their skills. So sampling needs to be improved, and that's even before talking about how to measure 'top,' which is not clearly defined in itself.

I do agree with you that hackerrank is a good way to practice new programming languages. It provides ways to practice in the language easily. Please note I wasn't saying that hackerrank is bad, just that it's bad for this particular purpose (for reasons I've elucidated).

Re:"Best" is a semantic argument

[Tenebrousedge]

So you are one of those best programmers who is messing around on hackerrank. Good job being a top programmer.

I'm honestly not at all sure how you read that into my post. Was it the part where I asked for help with a problem or the part where I implied that you might be so far above my level of skill as to find all of these problems easy? I mean, I don't think that there are many people who fit that description, but I'm certainly not going to assume that a longtime slashdot contributor is not one of them. I did in point of fact assume that I was responding to someone with greater skill and experience than myself, and thought that I was being fairly explicit about saying so: "this may put me in the chaff rather than the wheat", i.e. that I find this difficult may mean that I am simply a poor programmer. Did something I say contribute to the opposite impression?

(I've been talking to recruiters lately, and every time I hear someone intimate that I might count as a "senior" developer I die a little inside. The worst thing would be if it were true.)

So sampling needs to be improved, and that's even before talking about how to measure 'top,' which is not clearly defined in itself.

Right, hence the subject line, and hence why your first statement was unsupportable. Even if you could account for all programmers' time we still don't have an objective measure.

What you're saying is literally correct, that we do not know whether their sample is representative, however, the definition of "top programmer" does not naturally exclude participation on HackerRank. It's not like we yank your Geek Card for solving a challenge there. Statistics is not my strong suit, but I suspect that there would be a substantial overlap between top programmers as measured by HR and as measured by some other method.

That said, I'm a bit confused. You seem to be in favor of timed programming competitions, are you aware that HR does that, or do you find some other flaw with how they run them?

Re:"Best" is a semantic argument

[phantomfive]

That said, I'm a bit confused. You seem to be in favor of timed programming competitions, are you aware that HR does that, or do you find some other flaw with how they run them?

It's not whether or not they have programming contests or not, because programming contests are not necessarily a measure of top programming ability. They measure one aspect of programming ability (my current favorite programming contest is defcon CTF, although again that only measures certain aspects of programming).

The important thing is whether they brand or advertise themselves in a way to attract the top programmers. Topcoder is definitely attempting to do that, whereas hackerrank is branded to attract job searchers. They just aren't trying to attract the country's top programmers so sampling will be a problem for them.

Re: hackerrank does not correlate with good develo

[cordovaCon83]

That's when you write some really hacky code but it's okay because somebody else signed off on it.

Re: hackerrank does not correlate with good develo

[tempmpi]

You should be able to implement something easy in 30 minutes even in an interview situation. And most Hackerrank are not about implementing linked lists or trees, but actually using data structures such as trees or hashes, so its perfectly fine to use the data structures (and algorithms) included the standard libraries of C++, Java or Python. Take something such as Marc's Cakewalk. It easily can be implemented using a few lines of code (e.g.: 10 lines of non-golfed python). It shows if you are able to spot the trivial greedy algorithm and are somehow able to sort a small amount of numbers, but the number is so small that even bubblesort would work well.