IEEE Spectrum Ranks the Top Programming Languages

An anonymous reader writes Working with computational journalist Nick Diakopoulos, we at IEEE Spectrum have published an app that ranks the popularity of dozens of programming languages. Because different fields have different interests (what's popular with programmers writing embedded code versus what's hot with web developers isn't going to be identical) we tried to make the ranking system as transparent as possible — you can use our presets or you can go in and create your own customized ranking by adjusting the individual weightings of the various data sources we mined.

Comments

Not a ranking of what is the best language

[Meshach]

But more a ranking of what are the most talked about / downloaded / popular.

Re:Not a ranking of what is the best language

[Samantha Wright]

...And somehow, HTML is both on the list and lower than Java. Anyone got any theories?

Re:Not a ranking of what is the best language

[NortWind]

Java is often the top rated language, so it is not surprising that HTML is lower rated. I don't know if they are counting HTML 5 as HTML.

Re:Not a ranking of what is the best language

[Anonymous Coward]

HTML is a markup language, not a programming language. Jscript is the typical programming language associated with it.

Amazing that people still mix these up.

Re:Not a ranking of what is the best language

[rubycodez]

and HTML5 plus CSS is Turing complete. you lose

Re:Not a ranking of what is the best language

[Samantha Wright]

...is that really a property of Javascript, though, or just your computer? Do the ECMA standards prescribe memory or reference size limits?

Re:Not a ranking of what is the best language

[Mister Liberty]

HTML is not a language. Sorry.

Re: Not a ranking of what is the best language

[JoeJohnson2175]

HTML could be ranked lower due to the myriad of other languages which produce HTML output, rather than the programmer manually writing it.

Re:Not a ranking of what is the best language

[ericloewe]

By that logic, no programming language is Turing-complete.

Re:Not a ranking of what is the best language

[Dragonslicer]

HTML is not a language. Sorry.

All these years, I guess I was wrong about what the L stood for.

HTML may not be a Turing-complete programming language (I haven't looked to see how much HTML5 added), but it is a "language".

Re:Not a ranking of what is the best language

[phantomfive]

Just keep buying more RAM. Not a problem. Like, every year, buy another gigabyte, and it's practically infinite.

Re:Not a ranking of what is the best language

[Zontar The Mindless]

I believe you intended to say "HTML isn't as popular a beverage".

I hate it when poor delivery spoils what would have been a good punch line.

Re:Not a ranking of what is the best language

[Samantha Wright]

Nice catch, Captain Obvious. That was my point!

Re:Not a ranking of what is the best language

[jeepies]

Correct - real computers are not Turing complete. The term is misused. Real computers fall into a class of devices called Linear Bounded Automations because they have finite memory.

See here: en.wikipedia.org/wiki/Linear_bounded_automaton

Re: Not a ranking of what is the best language

[ultranova]

For example, is it possible to communicate using only HTML tags and nothing else?

Sure. You're limited to communicating the logical structure of a page, which is meaningless without the actual contents, but then again, all communication is meaningless without the context.

Re:Not a ranking of what is the best language

[ultranova]

I'm not sure Javascript is literally Turing complete, because it can't simulate an infinite tape.

Can't it? Since we're being technical, we must differentiate between Javascript specification and any particular Javascript implementation. Does the specification have anything in it that enforces finite pointer size, or could you, in theory, use BigInts? For that matter, can BigInts be implemented in a way that doesn't run into trouble when memory space gets exhausted, at least for the purposes of acting as pointers?

Re:Not a ranking of what is the best language

[jellomizer]

Figuring out the best programming language is just an opening to a flame war.
1. Different languages have chosen a different set of trade offs as to meet the problems they solve. Speed to run vs speed to code. Compiled vs interpreted. Verbose descriptive command vs quick to type but cryptic commands.

2. Different platforms. Are you coding for Windows or Linux perhaps for Apple. How much do you want to take advantaged of the platforms features?

3. You tend to favor what you know. Why do you think most of these top languages are C like. They are just variants on what you know.

Having had the choice to choose a language for a project there are a lot of factors. To say this language is superior then the rest is silly because the other languages were made for a reason.

Re:Not a ranking of what is the best language

[skids]

HTML5 is enough of a language that it is supplanting Flash and Java, and I think this is what they are referring to. Declarative animation can get you a long way and if all you are using ECMA for is to fill in a few gaps you can arguably classify it as a language for the purposes of this survey.

Re:Not a ranking of what is the best language

[surd1618]

Haskell is hypothetically Turing complete, because of lazy evaluation.

Re:Not a ranking of what is the best language

[ericloewe]

For very large values of "large", a large program will still require more memory than available.

Re:Not a ranking of what is the best language

[DamianJPound]

Probably because HTML/CSS (markup in general) is so easy to learn that it isn't a big deal if you know it, unless you're a graphics designer, then it's probably a huge benefit.

Re:Not a ranking of what is the best language

[rubycodez]

fallacy, don't need infinite tape, just one long enough to solve whatever problem is given.

buy more memory you cheap-ass 8D

Re:Not a ranking of what is the best language

[rubycodez]

no doubt much older than you, kid

GIGO

[BasilBrush]

Whether with programming languages or with studies it's the same: Garbage In, Garbage Out.

Select mobile, and you'll find Objective-C listed 16th, 6 places after MATLAB, and two places after Visual Basic. Which is clearly nonsense.

We already have tried and tested (back to 1989!) rankings for this. http://www.tiobe.com/index.php...
And Objective-C is currently number three across the board, never mind just mobile.

Re:GIGO

[rubycodez]

hardly anyone uses Objective-C

you are very funny judging a list of languages to be poor because your pet niche language isn't in top ten

Re:GIGO

[BasilBrush]

Unlike you I supplied a more reputable source of data. Objective-C is number 3.

And a long way above Ruby! ;-)

Re:GIGO

[Anonymous Coward]

Select mobile, and you'll find Objective-C listed 16th, 6 places after MATLAB, and two places after Visual Basic. Which is clearly nonsense.

No, Objective-C is number 6 on the mobile ranking. The default on the website is click-to-hide, so you've just hidden all the stats for mobile, while showing the aggregate of web, enterprise and embedded.

Re:GIGO

[rubycodez]

I only use Ruby for personal project, it's fun.

your data source is flawed, Apple biased. Sure at my job are lots of Macs, not a one used for Objective C development but instead java, python, php, etc. the bulk of the world uses other langauges to move money and move data on the web. certainly not objective-C though 20 years ago it was popular for some trading apps.

Re:GIGO

[Herder Of Code]

I have a hard time believing so much objective-c is being written. Maybe the list assumes that all ios apps are in objective-C which is false.

Re:GIGO

[iggymanz]

here are more realistic non-Apple biased list, your Objective-C pushed way done into the fringe where it belongs. Anyone who has been around in IT knows Objective C doesn't even come up outside of Apple development (and really there aren't many of those compared to finance, engineering, healthcare and web developers in the world

http://langpop.com/
http://blog.codeeval.com/codee...

Re:GIGO

[mark-t]

If you consider people who specifically write software for Apple hardware "hardly anyone", then yes.

Re:GIGO

[mark-t]

Not all are, obviously... but most are.. And even if "most" included only just more than half, that would still be a lot.

Re:GIGO

[Forever Wondering]

In the link they provided explaining how they do their rankings, they mention Google search is one metric and also mention that it's what tiobe uses [with a link to tiobe's page]. They're trying to be more transparent and use multiple metrics vs just one or two. Maybe it's time to have an alternative to tiobe. If both indexes, done with different methodologies, provide similar results, this would tend to bolster the validity of each.

Re:GIGO

[BasilBrush]

Rubycodez, it's a bit of a giveaway you have a sock-puppet account when all messages have an inability to use capitals or the English language, and have a maturity age of about 11.

Re:GIGO

[BasilBrush]

Maybe it's time to have an alternative to tiobe. If both indexes, done with different methodologies, provide similar results, this would tend to bolster the validity of each.

But as I pointed out cursory examination proves it's garbage. I've already donr Obj-C. How about MATLAB being 5 places above HTML? Complete drivel.

Re:GIGO

[iggymanz]

you are funny, now address the issue of millions of developers for other popular languages EACH (-- there are some caps for you) while objective-C has few hundred thousand.

Re:GIGO

[angel'o'sphere]

Every point of measurement is data. Hence an anecdote is as well, and bottom line I consider them the most important data points.

Re:GIGO

[phantomfive]

Select mobile, and you'll find Objective-C listed 16th, 6 places after MATLAB, and two places after Visual Basic. Which is clearly nonsense.

lol you're reading the results wrong. Read the instructions where it says 'click to hide'. Objective-C turns up in 16th place if you disable mobile.

So yeah, you're right, that is clearly nonsense. :)

Re:GIGO

[phantomfive]

Maybe it's time to have an alternative to tiobe.

There's an alternative.

Re:GIGO

[mark-t]

Of course... but that's still not most of them.

Re:GIGO

[Forever Wondering]

I agreed with your original first paragraph [but forgot to mention it--sorry].

We need multiple such ranking lists just like we need multiple style guides. On the latter, some are better than others, but when they all converge on a given point, that's when it's more likely to be a valid concept.

I just looked at the latest tiobe and it appears to better match how I would have [being a programmer] ranked some of the languages. Spectrum will no doubt [have to] tune their methodology, based upon the drumming they're getting in this slashdot topic.

But, comparing languages can be apples/oranges or more like wrench/screwdriver. Although a master index can have benefit, subdividing it by use cases may be better. For example:
- OS/embedded/realtime: C/C++/Obj-C
- Web server backend: perl/Java/php
- sysadmin: perl/python
Even still, there are overlaps/hybrids in these as well.

Re:GIGO

[RedWizzard]

Whether with programming languages or with studies it's the same: Garbage In, Garbage Out.

Select mobile, and you'll find Objective-C listed 16th, 6 places after MATLAB, and two places after Visual Basic. Which is clearly nonsense.

We already have tried and tested (back to 1989!) rankings for this. http://www.tiobe.com/index.php... And Objective-C is currently number three across the board, never mind just mobile.

The filters are meaningless because they just hide the languages that are not classed as being used in that space, they don't actually measure usage in that space. When you hide all but mobile they're still ranking the languages by overall use, not use in the mobile space. So C# is at 4th despite it having almost no use in the mobile space.

Re: GIGO

[theCzechGuy]

I wouldn't call that "hardly anyone" and I wouldn't call Obj-C niche language, but it does seem other languages have an order of magnitude larger userbase.

Re:GIGO

[Half-pint HAL]

Anyone who has been around in IT knows Objective C doesn't even come up outside of Apple development (and really there aren't many of those compared to finance, engineering, healthcare and web developers in the world

But the GP was talking about mobile development, and a heck of a lot of mobile development is Apple development. Have you ever heard of a mobile Matlab implementation? I hadn't. It exists, but I only know it does because I googled* it right now.

(* using Bing. because something I downloaded yesterday changed my search page.)

Re:GIGO

[BasilBrush]

Every point of measurement is data. Hence an anecdote is as well

A point of measurement is a datum, and thus that's the best an anecdote might be. However the other side of anecdotes is that as verbal or written stories there's at last as much fiction as fact amongst them, so they aren't really credible even as a datum.

Re:GIGO

[Xest]

"I said data. An anecdote is not data."

Yes it is, but that's besides the point, because don't worry, I know what you meant, you meant that an anecdote isn't as important a piece of data in your view, as TIOBE.

But here's the problem, TIOBE's methodology is so less than useless that I'm not sure it is better than an anecdote. Go read how they figure out their rankings - they take results from pages where content is mostly user generated and weight the result based on the amount of traffic that site gets. This means that user generated content sites like Wikipedia, and Blogger are being used to determine language popularity such that you can massively boost the ranking of a low down language by spuriously creating blogs and Wikipedia entries so forth including mention of it. There are many problems with this alone including the fact that they're reliant on the quality of the search functionality of the site in question to help with the weightings and we all know how pointless even Google's search results rapidly decline to after only a couple of pages.

It then gets worse. They apply arbitrary exemptions, they for example preclude Objective-C from searches on C to ensure it doesn't pollute the results but they don't filter out things like C sharp, meaning there are many thousands, possibly millions of results about C# polluting the rankings for C. They try and also give arbitrary confidence ratings, but the methodology for determining these seems to be completely made up and unpublished.

"There's well over a million apps on the iOS app store. The overwhelming proportion of which use Obj-C."

Most apps seem to get written using things like C# (i.e. in Unity, or MonoGame) or Javascript and HTML5 in tools like PhoneGap. The amount of native development on both iOS and Android is an absolute minority of all development.

When someone points to TIOBE to try and back up a claim you might as well replace "TIOBE" with "The Daily Mail" or "Fox News" or "The Register" because they're all equally full of shit.

This index actually looks pretty decent and quite realistic. It's based on stats that, unlike TIOBE, don't lie (at least not quite so much) like job listings - i.e. what companies are actually hiring for, in other words, what business is really actually using, an actually useful metric because most people looking at such rankings want to know popularity for the purpose of employment. I wont say it mirrors exactly my experience and view of the job market, but also I recognise that my experience is always going to be much more localised than a global study. I know Python is popular but I find it odd to see it above C#, however I guess there's every possibility it's used more prominently in other countries and cities so I don't think the ranking's results are far outside the realm of possibility.

It's certainly a much more realistic index than TIOBE, there's absolutely no doubt about that whatsoever - it far better mirrors reality for those of us working in the field. TIOBE isn't really much better than just outright making it up.

Don't tell someone you've proved conclusively they don't know what they're talking about when you clearly have not the slightest idea as to what tools and technologies are used for the bulk of mobile development. Even with the popularity of iOS and the vast number of apps it's app store has, much of it is still not written in Objective C regardless.

Re:GIGO

[angel'o'sphere]

Oh, so if someone posts 100m data points they are credible and when one only mentions one single life experience it is not?
Wow ... I did not know that. How does the judicial system in your country deal with testimonies? Sounds like you live in a very wacky society that you believe such nonsense.

Re:GIGO

[BasilBrush]

Oh, so if someone posts 100m data points they are credible and when one only mentions one single life experience it is not?

Do you know what a straw-man is?

Java is the new COBOL

[techfilz]

You use it at work but nobody gets excited to use Java at home.

Re:Java is the new COBOL

[DoofusOfDeath]

I used it in my PhD thesis work. It was a productive language, just not a fun one.

For fun, I still get the biggest kick out of pure functional languages. It's nice to see that job advertisements for them seem to be on the uptick.

Re:Java is the new COBOL

[RedWizzard]

You use it at work but nobody gets excited to use Java at home.

I use it at home. It's not appropriate for everything but I find Java + Eclipse to be a very pleasant and productive environment so it's often the option I prefer.

Re:Java is the new COBOL

[ultranova]

For fun, I still get the biggest kick out of pure functional languages. It's nice to see that job advertisements for them seem to be on the uptick.

Haskell's fun, but the second you actually want to do something with it, you run headfirst into the lack of an UI. You either settle for standard input and standard output (which, unbelievably, contain nasty bugs on WIndows) or pick one of a dozen or so wrappers around C libraries, and deal with the resulting issues (and hope the damn thing'll be maintained). It's an unfortunate situation which I think does a lot to keep the language from picking up usage.

Just goes to show that UI design should be part of language design from the start, rather than be left for the third parties. Otherwise you just get a mess.

Re:Java is the new COBOL

[abies]

I started to enjoy it again since I started using xtend (http://www.eclipse.org/xtend/). I suppose that same holds true for people switching to Scala/Kotlin.
Basically, you get expresiveness/tersity of python (often even better/more readable), while preserving performance (sub-second executions aside) and all the library support. Java-the-language might be going direction of COBOL, java-the-platform can be still enjoyable.

Embedded Perl programming

[Anonymous Coward]

I do embedded Perl programming.

Re: Embedded Perl programming

[ArcadeMan]

A whopping 1024 bytes of RAM? You live in pure luxury my friend. Try to work with only 128 bytes.

Re: Embedded Perl programming

[jd]

When I were a lad, we had to program in 96 bytes! In the snow! Uphill! Both ways!

Re: Embedded Perl programming

[narcc]

128 bytes? That's child's play.

C++

[Anonymous Coward]

C++ is the ultimate programming language. You can make any kind of application for any platform with the best performance. It also offers great selection of features, of which you can pick the ones that you need for your project. There is also a good ecosystem of libraries available, and premium compilers among other tools. Memory unsafety and lack of garbage collection might be problems though.

Re:C++

[phantomfive]

I've seen a lot of C++ in embedded. Anything that uses ARM allows C++, for example. Personally, I prefer to use C in that use, but C++ is fine.

The biggest problem with C++ in embedded is getting a compiler for it on obscure chips. Since a C++ compiler is significantly harder to write than a C compiler.

Re:C++

[_Shad0w_]

There's actually a standard for writing in C++ for (embedded) safety critical systems, created for the JSF. It exists partly because they were finding it increasingly hard to recruit engineers who knew Ada (or had any interest in learning it).

Re:C++

[serviscope_minor]

The biggest problem with C++ in embedded is getting a compiler for it on obscure chips. Since a C++ compiler is significantly harder to write than a C compiler.

I wonder about this. Sure a C++ parser is monumentally harder to write than a C parser, no questions. However, writing a C/C++ (one of the few time I believe this combo is justified) worth a damn is far, far harder than writing even a C++ front end. The optimizer on the best compilers is vast amounts of deep magic.

Not only that, apparently it is hard to write a C front end that doesn't suck massive donkey balls as many people who have tangled with obscure embedded chips will rant about over a pint or two.

Thankfully, quite a few vendors support GCC on new chips with new instruction sets which makes life much easier.

Though this leads to a curious thing. The massive, stinking pile of monumentally terrible 8051 compilers (for every embedded radio chip under the sun it seems) have sort of been grandfathered in, and since they kinda-sorta work just about well enough and there's no truly pressing need to provide any new compiler one is stuck with said donkey-ball sucking compilers rather than gcc.

Se we're stuck with: oooh so you want the second argument to this function being int a do you? lol no that's an internal compiler error.

Re:C++

[phantomfive]

I wonder about this. Sure a C++ parser is monumentally harder to write than a C parser, no questions. However, writing a C/C++ (one of the few time I believe this combo is justified) worth a damn is far, far harder than writing even a C++ front end. The optimizer on the best compilers is vast amounts of deep magic.

Well, I'd much rather write a C backend than a C++ frontend. Look how long the specification is, and it is full of ambiguities (some of them purposeful!). It would take me a long time to write a compiler I felt confident matched the specification, whereas C is small enough I think I could write a good compiler in a reasonable timeframe.

Re:C++

[serviscope_minor]

Well, I'd much rather write a C backend than a C++ frontend. Look how long the specification is, and it is full of ambiguities (some of them purposeful!).

I'm not sure there are any purposeful ambiguities. But, I'd still rather write the front end. I can more or less imagine how I could write a C++ parser that would pretty much work and not be appalingly slow. Of course I'm fanstasising slightly: I'm sure it would be harder than I expect. However, I know that a remotely competitive optimizer (e.g. half as good as GCC) is far beyond my capabilities.

C and C++ back ends are almost the same, though C++ has more in the runtime than C. I'm moderately sure that I could certainly write a parsed C/C++ to asm back end that would work. It would be very very slow, however.

It would take me a long time to write a compiler I felt confident matched the specification, whereas C is small enough I think I could write a good compiler in a reasonable timeframe.

Depends what you mean bu good :)

I certainly agree that going from nothing to a reasonably conformant, non crashy and somewhat correct C compiler would be vastly easier than the same for C++. But, doing that with one that has a runtime performance of even half that of GCC or LLVM would be I think far harder than writing a reasonably conformant, non crashy and somewhat correct C++ front end.

What makes me think that is a few people (e.g Walter Bright and a few PhD theses on parsing) have more or less written an OK C++ front end more or less single handed. I don't think anyone has written a high perfomance optimizer single handed.

Either way I wish hardware companies would stop pretending they have some proprietary advantage in their deradfully crappy software and stick to making good hardware.

Re:C++

[phantomfive]

Either way I wish hardware companies would stop pretending they have some proprietary advantage in their deradfully crappy software and stick to making good hardware.

Agreed.

Re:C++

[david_thornley]

Is a C++ parser that much harder to write than a C parser? The reasons C++ parsers are a pain is largely for the same reasons C parsers are a pain. There's more stuff C++ has to parse, but as long as that's fairly regular it shouldn't make the parsing much harder.

Verilog?

[whoever57]

I am somewhat doubtful that Verilog counts as a programming language. SystemVerilog, perhaps, but that isn't mentioned.

Also SQL -- yes, there is a distinct syntax associated with it, but is it a "programming language"?

Re:Verilog?

[ATMAvatar]

Yes. SQL is a turing-complete language. Whether you want to actually write full-blown programs with it or stick to the queries it was designed for, however...

Re:Verilog?

[Dragonslicer]

Also SQL -- yes, there is a distinct syntax associated with it, but is it a "programming language"?

If not, any of the variants of PL/SQL certainly are.

Re:Verilog?

[angel'o'sphere]

SQL is not a programming language but a querry language.

Re:Verilog?

[angel'o'sphere]

SQL certainly is not turing complete, but I wonder why people even bring up that matter?

Re:Verilog?

[eulernet]

If HTML is a programming language, then anything can be a programming language.

Re:Verilog?

[phantomfive]

Also SQL -- yes, there is a distinct syntax associated with it, but is it a "programming language"?

SQL99 is Turing complete. You need more than SELECT, JOIN, UNION though.

Re:Verilog?

[philip.paradis]

You're incorrect. Please see Cyclic Tag System and Turing Machine in SQL.

Re:Verilog?

[philip.paradis]

Indeed, Turing Machine in SQL is a really neat five part series on demonstrating Turing completeness in SQL. Fabien Coelho does a really nice job of walking the reader through the various stages.

Re:Verilog?

[jd]

I didn't see SystemC, either, but may have missed it. But the real question is what the stats would look like if you only included Wishbone-compliant usage?

Re:Verilog?

[phantomfive]

Nice

Re:Verilog?

[Tablizer]

Shit, that means SQL viruses/injections can have more power

Re:Verilog?

[phantomfive]

That's why I write all my data lookups in Prolog

Re:Verilog?

[angel'o'sphere]

This is not SQL but an enhanced/extended proprietary dialect.
The first sentences even mention the extensions :)
Better luck next time.

Re:Verilog?

[philip.paradis]

You clearly didn't actually read the material provided. Please read it, in its entirety, and let me know when you're done. Cheers!

Re:Verilog?

[angel'o'sphere]

Why should I? It does not add to the discussion.
On top of that, the definition of turin complete is not that you are able to simulate a turing machine, albeit if you can do that you are turing complete, hence the misconceptions in this thread I guess.

Re:Verilog?

[philip.paradis]

You should do so because you've proven that you're capable of not only once, but twice, commenting on a topic that you're unqualified to speak on. Will you continue to refuse to read the referenced materials, reply again, and hence continue to demonstrate your willful ignorance? Your initial statement was provably false, and you're simply unable to admit your error. That's pathetic.

Re:Verilog?

[angel'o'sphere]

What is your problem?
Standard SQL (1986/1999) is not turing complete.
The articles you linked are not about standard SQL but about modern derivates, no idea if they 'just yesterday' became an agreed standard (which is not adopted yet, and from which is unclear which vendor will when implement what of it). And I don't care. I'm not particular interested in using SQL for solving programming problems.
Your initial statement was provably false, and you're simply unable to admit your error.
No, it was not!
Google: is SQL turing complete
So simple, 100ds of articles showing you that I'm right.
But good luck with your insulting attitude, you might need that luck.

Re:Verilog?

[philip.paradis]

You should be asking yourself what your problem is. Clearly, you still haven't read the referenced materials; proprietary extensions are not needed. Here's something else to read while you're at it: SQL Standardization. There is no single "standard SQL." SQL standardization has gone through many iterations: SQL-86, SQL-89, SQL-92, SQL:1999, SQL:2003, SQL:2006, SQL:2008, SQL:2011. The SQL standard is presently maintained by ISO/IEC JTC 1.

So yes, your initial statement (which was "SQL certainly is not turing complete") was and remains provably false, and you're still simply unable to admit your error. Put down the shovel. Do you conduct your professional affairs with the same level of reasoning you're demonstrating here? Incidentally, appealing to the authority of "hundreds of articles" you claim will show you're right doesn't help your cause any, as the majority of such articles will be nothing more than "Bob's Blog" posts and will be equally based on ignorance. Please feel free to keep arguing, though I still recommend taking a break to read the originally referenced materials in their entirety.

Re:Verilog?

[philip.paradis]

As you are a practitioner of Aikido, I'm genuinely surprised by the direction this entire commend thread has taken. I am looking forward to your next reply, though.

Re:Verilog?

[angel'o'sphere]

So you still believe that 'ordinary' SQL is turing complete?
Wow ... that is astonishing.
But as you obviously keep insulting me ... I guess you don't want to learn anything.

Re:Verilog?

[angel'o'sphere]

There won't be any as you obviously only argue for arguing sake ...

Re:Verilog?

[philip.paradis]

To repeat myself: there is no single "ordinary SQL." SQL standardization has gone through many iterations: SQL-86, SQL-89, SQL-92, SQL:1999, SQL:2003, SQL:2006, SQL:2008, SQL:2011. The SQL standard is presently maintained by ISO/IEC JTC 1. Your original statement was "SQL certainly is not turing complete," and that is a false statement. Under the ISO standards, it is absolutely possible to create a Turing machine with SQL. Examples have been provided, including (but not limited to) one written "entirely in SQL:2008-conformant SQL." The degree to which any given database engine may adhere to ISO standards may vary, but by adhering to said standards, there exist code examples which demonstrate Turing completeness. You're only insulting yourself by continuing to refuse to accept reality, but if you're still in doubt, per the previously supplied references you're welcome to purchase SQL standards documents from ISO, IEC or ANSI.

Re:Verilog?

[angel'o'sphere]

Ofc there is no "single" standard. After all the language is evolving, like most "programming" languages. (The quotes have a particular sense).
However you seem not to notice that nearly every standard you care to mention .... is not turing complete
The conclusion is up to you ...

Most discussed = worst

[holophrastic]

So if a language is particularly awful, and troublesome to use, or used by students discussing homework exercises, then it'll score high here. Thanks for the useless popularity sketch.

I see that money is not the favorite language

[Anonymous Coward]

Should ask lawyers and managers.

Re:The Embedded list is the best

[ArcadeMan]

No I didn't. And there's no such thing as "Square-Enix", "Sonic games on Nintendo consoles" or "Star Trek reboot" either.

Now get off my lawn!

Best editor?

[pipatron]

What's next in flamebait-land? Rate the best editor?

Re:Best editor?

[Brett Buck]

That would be silly indeed, since its clearly VI.

        Brett

Re:Best editor?

[DMUTPeregrine]

Well, it's certainly better than timothy...

Re:Best editor?

[jd]

Already done. Check the Slashdot archives for the Vi vs Emacs paintball fight.

How did Java beat C

[Murdoch5]

It's a pretty good list but my only question is how on earth did Java beat out C. Java is a decent language for a lot of different areas but doesn't come to the table in any one area and own the hill. On the other hand C is the king of the embedded world, Operating System world ( such as kernels ) and can still rock it on the desktop with C++ and C#. If C and Java switched places then it would be prefect, until that happens I can't really agree with it.

Re:How did Java beat C

[phantomfive]

how on earth did Java beat out C....... the other hand C is the king of the embedded world, Operating System world ( such as kernels ) and can still rock it on the desktop with C++ and C#.

Because most programming isn't exciting new hip startups, it's not embedded.
Most programming in the world is boring business software. And that is where Java shines, for various reasons. As someone else pointed out, it's like the COBOL of the 21st century.

Re:How did Java beat C

[Murdoch5]

Java is what a lazy developer uses to get free security and free memory protection, a child could write a business application in Java and have it secure. To me that doesn't sounds like a good language as much as it does a language for lazy programmers, which should loose it a few points.

Re:How did Java beat C

[phantomfive]

Java is what a lazy developer uses to get free security and free memory protection, a child could write a business application in Java and have it secure

No, if you think Java will give you security and free memory protection, then your program has both security holes and memory leaks.

Re:How did Java beat C

[gnupun]

but my only question is how on earth did Java beat out C

That's like asking why isn't Assembly language on the top of the list? It runs circles around C in the performance area, both in speed and size (important for embedded apps).

Java beats C because you can accomplish more in Java than C with fewer lines of code and less mental effort. Things like exceptions, OO, garbage collection, a massive library, etc. save a lot of time compared to C. Debugging is also relatively painless because you get a stack trace in Java, but not in C.

I too have a few questions:
How on earth did VHDL beat Verilog? I thought everyone in companies used verilog because VHDL was too complex, like Ada.

Why is Julia not on the list? Its syntax is similar to Python but performance approaches C/Java.

Re:How did Java beat C

[Murdoch5]

Well I'm an embedded system developer and I can tell that compared to C, Java gives you everything and the kitchen sink.

Re:How did Java beat C

[phantomfive]

I'm an embedded programmer too, and I can tell you that if you're having trouble with memory or security, the problem isn't the language, it's you.

Re:How did Java beat C

[mvdwege]

fewer lines of code

I'm sorry? Have you ever seen the huge amounts of boilerplate Java requires?

Re:How did Java beat C

[Murdoch5]

I never said or implied that I have any problem with memory management or security. The problems come when languages like Java, C++ or etc... try to insert protections for you without knowing exactly how you want the protection to work. Java also extracts the programmer way to far away from the hardware, which is essential for any good developer, who doesn't work in web land.

Re:How did Java beat C

[gnupun]

That's often a symptom of bloated corporate design. You can write clean Java code that resembles C in succinctness. Granted the language is verbose using words where other languages use symbols. Why isn't there a sensible default for method access -- 'public or private?' The default 'package' access is rarely used.

Re:How did Java beat C

[kauaidiver]

Lazy? Java is overly verbose, a lot of typing.

protected static transient synchronized const unsigned int m;

It loses points in my book for being too verbose.

Re:How did Java beat C

[Murdoch5]

SecureSocket sslSocket = new SSLSocket(blah);
List linklist = new List();
Tree tree = new Tree();

So how is it not for the lazy programmer?

Re:How did Java beat C

[kauaidiver]

Work in Java for a few years or on a large Java project then go back to C or switch to Python.

It boils down to what you want to do: use best tool for the job. But I wouldn't be caught dead working on a new project/team who decided to use Java as the primary language.

Re:How did Java beat C

[ultranova]

Java is what a lazy developer uses to get free security and free memory protection, a child could write a business application in Java and have it secure. To me that doesn't sounds like a good language as much as it does a language for lazy programmers, which should loose it a few points.

Java should lose points because you think it's not challenging enough? Seriously?

Re:How did Java beat C

[ComputerGeek01]

Java is a decent language for a lot of different areas but doesn't come to the table in any one area and own the hill.

Except, you know, that whole Android API thing...

Re:How did Java beat C

[Murdoch5]

Fair enough

Re:How did Java beat C

[dkf]

The default 'package' access is rarely used.

Huh. I use it quite a bit when implementing an API. (You hardly need to use public at all inside interfaces.)

Comment removed

[account_deleted]

Comment removed based on user account deletion

IEEE, It's Perl, not PERL

[nick_urbanik]

IEEE shows "PERL" at number 11. IEEE, It's Perl, not PERL.

Verilog?

[TechyImmigrant]

Why do they only list VHDL and Verilog HDLs?

SystemVerilog and VHDL are now the primary HDL languages we use in chip design.

Popularity != Quality

[jd]

There should have been modifiers for typical bugs per kloc and security holes per kloc.

Also, there are many more layers to the industry. Scientific computing? Avionics? Publishing?

The subdivisions between languages are also a bit... strange. Java/Oak isn't truly uniform, whatever anyone claims. C and C++ have standards that aren't always backwards-compatible - if you ignore such changes, why bother listing C# or D as distinct? Lump the lot, together with B and BCPL under a single header.

My guess is that accurate representation of languages isn't possible (when does a dialect become a distinct language?) but that if it was, none of the so-called "big three" languages would be in the top 10. Computer languages are as bad as natural languages when it comes to classifiers.

Last, but by no means least, people rarely directly code any more. They code within engines, usually using some weird fringe language nobody has ever heard of that turns out to be Lua or Visual Basic with the keywords words renamed for the theme. Real programmers (as opposed to integer or complex programmers) tend to be in the minority, have become rarer after Qualcomm outlawed them, and are mostly in mourning for Freshmeat. But as a lot are Goths anyway, it's hard to tell.

Re:Popularity != Quality

[jd]

I am the One True God of..... ummm..... damn, what's that word for when you can't remember something?

Re:Java?

[jd]

Go sit in a corner! Thou shalt never bring those two words in close proximity, for should they ever come in contact, the reaction could destroy reality as we know it!

Re:Java?

[linearz69]

Sorry. How about "Dalvik"?

You forgot the part about

[Latent Heat]

. . . we loaded programs by flipping toggle switches for the binary op-codes until our finger bled. And we liked it!

Re:You forgot the part about

[JoeCommodore]

Luxury.

My mates and I had to wire the core logic for each operation using barbed wire and hand tooled relays that we had to build in metal shop. I'll tell you, laddie, we smashed many a finger just to resolve 2 + 2 = 5 - but back then it was close enough.

Re:You forgot the part about

[ultranova]

My mates and I had to wire the core logic for each operation using barbed wire and hand tooled relays that we had to build in metal shop.

From metal you crushed with your bare hands from primordial hydrogen, I presume? No? Thought so.

Re:Let the flames begin!

[narcc]

Somehow, I don't think the parent was actually wrong...

So what are good languages to get into?

[blindseer]

I look through the comments here and it seems a lot of people are unhappy, displeased, confused, or otherwise negative about the rankings of languages in the list. As someone that is seriously considering going to graduate school to update my programming skills I'd like to know where I could get the best return on my investment.

I did VHDL and Verilog primarily for a few years. As is the nature of the beast there was some mix of programming in a lot of other languages that went with that to make tools work, add in existing legacy code, or accommodate the preferences of co-workers and managers. Some that come to mind are C/C++, shell scripting, PHP, Python, Perl, XML/HTML, and JavaScript. I came across some MatLab, tk/TCL, and Fortran but I didn't have to know the language but know it was there, how to run the code, and who to talk to when it broke.

I did some development on MAMP, LAMP, and WIMP systems off and on over the years. This meant using SQL, PHP, HTML, CSS, JavaScript, and since the development platform I was using was a Mac I had to know some shell scripting, AppleScript, and probably one or two more I've since forgotten.

Wanting to get back into web development I'm considering classes on Python, Perl, Ruby, JavaScript, and HTML5. I'd probably also brush up my knowledge of C/C++ and Java since they see unavoidable as both a computer science student and someone in the real world of programming just about anything. Old code for web pages will very likely have C++ and Java in them that need to be reused or replaced with something newer, faster, and more compatible.

A masters in computer science program means taking about 10 three credit courses to get the degree. That means learning potentially 10 different languages. Which 10 would you choose? Which of those 10 are a must to learn, which would be merely advantageous to know?

I thought I had my mind made up on which classes to take but reading the fine article and some of the comments here leads me to think I may need to reconsider my priorities. What shold my priorities be?

Re:So what are good languages to get into?

[john.c.earls]

I'm working on my PhD with a focus in Computational Biology, so I may be biased toward the academic side of things, but I love to sling code so I'll give answering this a shot. You say that you are mainly interested in web development, so I will try to focus on that world. For client side dev, Javascript is really the only choice these days. If you have any experience with pure Javascript, you probably know that it is a ridiculous mess of a language. It has been made (reasonably) useful by several frameworks in recent years. You will want to look into angular, backbone or react in order to make anything useful. The big thing to get clear in your head with modern Javascript is how to handle asynchronicity. On the app server side, the new hotness appears to be Go. It is popular because it makes concurrency easier. Most of the push these days has been taking advantage of concurrency and the ascendant languages (Go, Node, Julia) generally in some way try to make this easier. For interpreted languages, Julia looks really interesting and IMHO has a good shot at being very popular in the near future. It has a smart, friendly community and you might get in on the ground floor of a language that will be hot when you are done. You should learn to use some NoSQL datastore, and more importantly when and where you want to use a one. Of course, C/C++ is essential to understanding (almost) any modern programming language, not to mention Data Structures. If you believe you need brushing up on these, then that should be your first stop. Take a systems class your first semester and you should get a nice refresher on what is actually going on under the hood with most of the rest of the languages. One thing to keep in mind is that all programming languages go through periods of growth and decay. You should focus on concepts, with the languages being merely tools that have certain niches which sometimes map better to certain concepts than others. The key concepts of the future are concurrency, distributed and heterogeneous architectures(cloud/devices/IoT), data management and security. That said, you asked for a top ten, so if I was looking for Masters with a focus on web dev at this point in time, my top 10 would be- C/C++, Javascript(Client side), Go, Julia, Python(probably less important in the future, but I feel there is nothing better for rapid prototyping), Node, Java( I don't like Java, but it is very versatile), Haskell(token functional lang), SQL, HTML/CSS. Again, the languages themselves are less important than the concepts. The future is asynchronous, concurrent, distributed, heterogeneous and data rich. Prepare yourself for that world and you should be in good shape

Re:So what are good languages to get into?

[dkf]

A masters in computer science program means taking about 10 three credit courses to get the degree. That means learning potentially 10 different languages. Which 10 would you choose? Which of those 10 are a must to learn, which would be merely advantageous to know?

Take at least one OO language (Java's fussy and bureaucratic, but its a pretty good example of the breed and is likely to be useful after you get your masters), at least one functional language (probably Haskell these days), at least one declarative language (Prolog or SQL), and don't just learn programming languages. You also need to learn about data, about data structures, about algorithms and their analysis, about parsing and compilation, and about concurrency; these are all independent of any programming language.

But computing is well served by not just learning about computing. If you have time, learn about math, stats and logic too, and learn how to communicate your ideas effectively; you'll never get far if you can't communicate with other people well.