Google AI Challenge: Languages Used by the Best Programmers | Hacker News Reader

Google AI Challenge: Languages Used by the Best Programmers

r-chart.com

36 pointsEzGraphs26 comments

Too many colors, too close together in hue, and the hue adjusts not by how the languages are related, but by alpha sorting of the language names.

Terrible info graphic! I'm sure I could do much better in 10 minutes, and I'm a rank amateur. Someone has not read Tufte.

Pretty sure he just typed them into R and pasted the result into his blog. How much time should someone spend on a blog post being provided for free?

Agreed, you can publish anything you like on your "provided for free" blog, but if you want anybody to care you should definitely put some effort into that. See http://news.ycombinator.com/item?id=1963612

At least five minutes more to create a better graph.

Sometimes it is ok to have ugly color combinations just for claritys sake. I guess the author might not agree.

The contrast in the combined graph makes it impossible to distinguish a single language.

Agreed - that's why some of the interesting stuff is broken out further down (and data and code is available for download for data mungers).

Suggestion: Lose the shading, and use a wider palette on the lines.

It would be nice if it was a "live" graph where you could toggle languages on/off.

I wonder why does the author refer to C programmers as 'hippies'? I sometimes use C, and I have been called a hippie, but I have never been called a hippie because I use C. But perhaps I am not hip to where this originates.

It's ironic since (as someone who was a C programmer for ages) we used to be seen as the uncultured pragmatists. Yeah, C is kinda crude but it's fast. When did that get turned into Hippydom? I'da thought all those dynamic languages were more hippyish with their (almost)anything-goes approach.

> I sometimes use C

Ah - but - had you entered this, would you have choosen to use C?

My anecdotal experience (Given a problem, and C, I struggle to get things done) is that lots of people like the idea of C, have fond memories of it and occasionally play with code snippets in it. But when those people go into getting-things-done-mode they choose C++ or Java as their primary.

But there are some die-hards who think C remains a superior general-purpose lanuage to C++ and Java, who probably don't like OO, and who stick with C.

I thought the writer called them hippy to highlight this spirit of contrarianism. If that's true, though, it's a poor choice of word because it doesn't suggest a connection with either tradition or asceticism. 'Monk' would better capture that.

I'm amazed that Python had about 20 times as many entries as either Perl or Ruby. I realize they each have their own niches and that Python tends to be the current language of choice for a lot of general-purpose weekend hacking, but I wouldn't have expected the difference to be that overwhelming.

Python seems to be pretty popular for AI. This is a shot in the dark but could have something to do with the popularity of "Programming Collective Intelligence". It uses and teaches Python and is a top seller in the AI category on Amazon http://www.amazon.com/gp/product/0596529325?ie=UTF8&tag=...

I have really enjoyed going through the examples in that book, which was the first time I did any AI (although one might argue it's light on the AI side) type stuff since doing some Lisp stuff way long ago.

Norvig's Artificial Intelligence: A Modern Approach also has Python code examples.

Oh wow, do you know what edition started doing that?

These results are potentially biased by the fact that starter packages were originally only available for C++, Java, and Python (at least when I took a look at it back when the competition started)

What a horrid graph. I wish I could read that.

Was kind of interested if PG's assertion about Java being for "average" programmers and Lisp (and some other languages) being for "good programmers..."

It seems like there is a trend that way - and C and Haskell are also for "good" programmers.

Actually, the results showed that there were a much greater range of Java programmer abilities than most other languages. There were more Java programmers in the top 10 than any other language, so asserting that Java is just for 'average' programmers is rather shown up for the baseless snobbery that I rather suspect is PG's basic attitude.

Maybe I'm misrepresenting Graham's remark concerning averageness of languages, but I the 'average' attribute was not assigned to the capabilities of the individual programmer. The point was (among others) that languages with a good "market position" attracts all kind of programmers, from challenged to geek. On the other side you have languages who attract more geeks because the i-have-to-pay-my-mortgage-which-language-should-i-learn group is underrepresented. Which makes for a good filter if you want to hire people. Just like "participating in open source projects" can be a discriminating factor.

The last point could be argued about, but I don't recall language bashing in said essay.

No Prolog, how sad! I suspect there wouldn't be a particularly high frequency of developers using it anyway but it would be nice to know what proportion do given a larger study.

According to the TIOBE Index[1] about 0.3%.

[1] http://www.tiobe.com/index.php/content/paperinfo/tpci/index....

Interesting how Python goes from 20 in the top 100, to 0 in the top 10. Java, C++, and C# seem to scale all the way down to the top 10.

Lisp won, but maybe the question is, is there something wrong with Python?

It is impossible to tell which color is which in this graph.

Oh that is awesome. Anyone know if there are plans for another tournament?

http://ai-contest.com/forum/viewtopic.php?p=8080#p8080

Hopefully not as long as this one (3 months).

Reply on news.ycombinator.com