BTW, for fun, see this recent job for a Haskell expert in India - http://www.simplyhired.co.in/job-id/jh4gosrgqr/haskell-exper... -- turns out (I asked) they're developing tools to translate COBOL into more modern languages, since you can't hire COBOL programmers anymore.
Well, at least he used some line graphs at the end of the article, showing ratings over time, and not just a pie chart with status at some date.
Looking at the line graphs allows us to intuit the "first order" derivative of the popularity graph. COBOL has a negative derivative, Scala has a positive derivative. There might be more people employed maintaining COBOL, but which field is shedding jobs, and which is hiring, if only a little?
I'm a little bit biased, because I have actually used Scala a tiny bit at work. I found a niche that it is actually quite good at: one-off scripting on top of existing Java application code.
Scala may compile slow, but when you are running a 200 line script to do some research, having concise code with a few functions passing around a few lists and tuples, which allows you to "just twiddle a bit, hit 'up arrow' and rerun" is a win.
I've never developed an application in Scala, and I probably never will, but I love it for scripting.
Oh, and I have worked on RPG a bit, too, but that was in the early 90s. Actually, not so much programmed in RPG as worked on a product to translate RPG into C on various OSs. Whether or not the clients continued programming in RPG using a cross compiler, or took their C code and ran, I can't say. YMMV :-)
Except that services like Twitter are used massively in the real world and do use Scala.
They are both languages in the tails, although different tails: Scala currently only has early adopters, Cobol is mostly used in systems that are in maintenance model. Scala has the potential to get a bigger slice of the pie, while Cobol's glory days are past.
Well, for true hackers[1], it is not about getting shit done. It is about how shit is done or, to be precise, how proof-of-concept shit is done. And getting enjoyment out of the process itself.
So if the point of this article is to say that we shouldn't buy into hype then I agree. However to try to claim that the Tiobe represents anything close to an accurate portrayal of programming language popularity is nonsensical. also this idea that web hype and the "Real World" don't match up is flawed. Most things get web hype because they are doing things in this real world.
Agreed. TIOBE is only a part of gauging a language's popularity. TIOBE looks at the results of search engines regardless of the recency of the search results. So older languages will obviously have more sticking power. I like to look at current job offerings, google trends, and book trends along with TIOBE when gauging a language's popularity.
TIOBE might not be accurate but I think we can say it is "close to accurate" based on the fact that it agrees with known trends such as the recent iOS related rise of Objective C, or the long slow decline of Perl.
From wikipedia: "Popularity is the quality of being well-liked or common, or having a high social status."
I doubt that Scala is less liked than COBOL, however, due to the age of COBOL, it is very likely that COBOL is more common. Don't really see the point of the article. Either it states obvious news or it kinda claims that of all programmers out there, most will choose COBOL over Scala (which I doubt heavily and there is no data to back that up - but that is probably the more interesting question on popularity).
Anyway, I did a bit more than a Hello World with COBOL in university. I know companies are searching like mad for COBOL programmers and you could earn a shitload of money... But I won't do that stuff for the rest of my life!
First of all Cobol is actually quite widely used for legacy systems.
As as been covered frequently Tiobe is a poor indicator of language popularity especially for Scala. e.g. capecoder.wordpress.com/2012/04/09/language-popularity-its-not-about-search-engine-result-counts/
I did a short stint of Scala programming and I didn't like it.
The last straw was when I spent a few days debugging something involving actors that the author of the code didn't understand. It took 15 minutes to reimplement this flawlessly with Java's ExecutorService and that handled all the tough problems like error handling.
I learned a lot from Scala that I carried into my Java programming, but I wouldn't start a project in Scala.
> That must be a accident of the way TIOBE is put together yes? No!
I had at least expected that the author did his 5 minutes of research to discover that there are indeed issues affecting rankings of languages like Scala, Haskell, ...
BTW, for fun, see this recent job for a Haskell expert in India - http://www.simplyhired.co.in/job-id/jh4gosrgqr/haskell-exper... -- turns out (I asked) they're developing tools to translate COBOL into more modern languages, since you can't hire COBOL programmers anymore.
Of course, the translation tool is in Haskell. :}