Swift Tops List of Most-Loved Languages and Tech

Nerval's Lobster writes Perhaps developers are increasingly overjoyed at the prospect of building iOS apps with a language other than Objective-C, which Apple has positioned Swift to replace; whatever the reason, Swift topped Stack Overflow's recent survey of the "Most Loved" languages and technologies (cited by 77.6 percent of the 26,086 respondents), followed by C++11 (75.6 percent), Rust (73.8 percent), Go (72.5 percent), and Clojure (71 percent). The "Most Dreaded" languages and technologies included Salesforce (73.2 percent), Visual Basic (72 percent), WordPress (68.2 percent), MATLAB (65.6 percent), and SharePoint (62.8 percent). Those results were mirrored somewhat in recent list from RedMonk, a tech-industry analyst firm, which ranked Swift 22nd in popularity among programming languages (based on data drawn from GitHub and Stack Overflow) but climbing noticeably quickly.

Comments

Matlab

[CurryCamel]

Matlab? Dreaded? Why?
I don't like it. But its good for its purpose, I find.

Re:Matlab

[Enry]

Take a look at the demographics of the people that responded. Overwhelmingly male, average age is 29, half have been coding for less than 5 years, and almost half don't have at least a BS degree in CS.

Re:Matlab

[bsDaemon]

Are you saying that if they were overwhelmingly female that they would appreciate MATLAB more? I don't see how that has anything to do with the other things, which merely point out how a bunch of hipsters who don't really understand computers would rather work with mostly easy languages to do "app" development than to work with anything in environments that indicate either some domain knowledge (MATLAB) or a "real job" (Salesforce, SharePoint).

Also, consider that it is Stack Overflow. It's mostly code-snipet "programmers" and people looking for help with homework.

Re:Matlab

[grimmjeeper]

That's what they've been saying about every new generation of languages since COBOL. They even said it about C++ and Objective-C when they came out and all the C programmers didn't want to learn it.

Re:Matlab

[gweihir]

I have no idea who "they" are. I am saying this about Swift. Incidentally, nobody halfway sane would say that C++ or Obj-C is for the incompetent in comparison to C. They are all three hard to master and C++ is a pretty bad design in addition. (Of course, C++, like the equally bad Java, was never intended to be a general-use language.)

Re:Matlab

[Gavagai80]

Are you saying that if they were overwhelmingly female that they would appreciate MATLAB more?

Possibly, since women are much more likely to be math majors than comp sci majors and nearly anyone who studies math learns MATLAB.

Re:Matlab

[Tom]

So, your average software developer. Which explains a lot about why software quality sucks so much. (and then someone writes six code analysis tools and ten testing tools to at least catch the shit before it hits the fan).

Re:Matlab

[EmeraldBot]

Are you saying that if they were overwhelmingly female that they would appreciate MATLAB more?

Possibly, since women are much more likely to be math majors than comp sci majors and nearly anyone who studies math learns MATLAB.

The gender doesn't matter, it's the experience level. These guys are hipsters, and quite literally, the tool is older than they are. Further more, I'm sure they're a lot of people who combine their dislike for math with MATLAB itself...

Re:Matlab

[grimmjeeper]

"They" are all of the programmers who have come before you. The computer engineers bemoaned COBOL because it let non engineers make programs to run on computers without knowing what was going on under the hood. Same with FORTRAN.

As languages have progressed, each generation provided more abstraction and let people get farther away from the hardware without having to understand what was going on at lower levels. And every time a language provided a new abstraction, the old guard pissed and moaned about how it was destroying programming by making it so the next generation wouldn't have to know all of the lower level details they knew. Object oriented languages were yet another in a long line of language advances that were decried by the previous generation who didn't want to learn the high level abstractions and were mad that the low level details they know well were no longer important. They claimed that the people who just wanted to program at a high level were "incompetent" and how it was leading to a "general cretinization of programming, and consequentially programmers."

Funny thing though. Most modern day programmers have no real clue what's really going on behind the scenes and they're pretty clueless about what happens once their source code goes through the compiler. And that's my definition of "incompetent". So when you say "Swift is for the incompetent", I say that incompetent people can write code in pretty much any language because I've seen them do it for decades over many generations of languages well before anyone knew what Object Orientation was.

Not only that, the old-timer "get off my lawn" types have been spouting the same nonsense since the 50's when it comes to newer technology. Sure, some languages are targeted at making tasks simple so they're easier for simpletons to use. And, of course, some languages are more poorly designed than others (*cough* *cough* C++ *cough* *cough*). And there are many languages that are more difficult to learn and master than others for no good reason. But people have been deriding next generation languages since the first generation. And that's why I laugh at your characterization of Swift.

Re:Matlab

[gweihir]

Yes, nice (and I know all that), but how does that relate at all to my comment?

I have no problems with new tech (I am currently doing quite a bit of Python 3.x), but there has to be a good reason for it, and making it easier for bad programmers to produce more bad code is not a valid one.

Re:Matlab

[TechyImmigrant]

Well swift has a bucket load of nice clean language features for a compiled language.
That seems like a reason to me. Cleanliness.

 

Re:Matlab

[aralin]

.... nearly anyone who studies math learns MATLAB.

Hence the dread! Back in '94 to '99 I've been sysadmin at the math department of my university. Already then all the students that were learning Matlab dreaded the subject. I was observing lesson after lesson, year after year, since they happened at the Comp. Lab. I was managing, always the same thing.

Re:Matlab

[Darinbob]

I'm not so sure, I was there at the start of C++ and a lot of programmers were very interested in trying it. One key feature of early C++ was that it was backwards compatible with C. Indeed it compiled into pure C which was then compiled into an executable.

Re:Matlab

[grimmjeeper]

Sure. A lot of the younger programmers (like myself at the time) were excited about it. But many of the old timers were very resistant to it. A lot of people complained that if it just compiled to C, why not just write it in C and save a step. And there was much bitching about how everything was so indirect and hard to figure out exactly what was going on. Of course, this was before there were good debuggers and direct compilers.

But it's like so many other things. Old folks get set in their ways. Young people want to change everything. I find myself starting to get the "get off my lawn" attitude now that I'm past my prime. But I do go out of my way to learn new things and try new ideas.

Re:Matlab

[Enry]

Are you saying that if they were overwhelmingly female that they would appreciate MATLAB more?

I don't know. In my only actual programming job in the early 90s I worked with some great MUMPS developers who happened to be female and they were at a far higher percentage of the developers I worked with than listed in the survey. I'm just pointing out that the demographics of the SO survey may not really represent what the programming community is, so what they say as what they like may not be reality.

Also, consider that it is Stack Overflow. It's mostly code-snipet "programmers" and people looking for help with homework.

That's true. It's been a real help in (re)learning Python.

Re:Matlab

[Darinbob]

C was so new anyway at the time, about ten years old, that the real old timers were probably bitching about wanting to use PL/1.

Re:Matlab

[arth1]

And there was much bitching about how everything was so indirect and hard to figure out exactly what was going on. Of course, this was before there were good debuggers and direct compilers.

I think the point was that one should know exactly what was going on before it was run, and being able to follow the low-level flow by looking at the source without "and then, magic occurs" moments.

Simula and Ada were arguably much better languages for understanding exactly what was going on, being as unforgiving as leather clad mistresses, but "C with classes" won out because it was so similar to C. C++ is great, but it does give the programmers enough rope to hang themselves with, while obscuring what's really going on behind the scenes.

Re:Matlab

[Jeremi]

there has to be a good reason for it, and making it easier for bad programmers to produce more bad code is not a valid one.

If all you've got is bad programmers, and their bad code is nevertheless good enough to accomplish the tasks you need to get done, then a tool that allows bad programmers to produce more bad code may be just the thing you need. (of course some would argue that that niche is already filled by Java, but time will tell)

Re:Matlab

[gweihir]

While at the first glance, you argument is compelling, that is not what is going to happen. There is a feedback-loop here: "Management" will hire even worse coders, as the language clearly does most of the work now. As a result, not readily obvious properties like security and reliability will get even more abysmally bad and code quality will drop through the floor.

On the plus side, we are missing a few really impressive IT catastrophes that cannot anymore be swept under the rug to drive home the point that creating software is hard. Swift will surely contribute to reaching that goal.

Re:Matlab

[lucm]

Well swift has a bucket load of nice clean language features for a compiled language.

Can you name 3? (insert cricket chirps here)

Re:Matlab

[TechyImmigrant]

Well swift has a bucket load of nice clean language features for a compiled language.

Can you name 3? (insert cricket chirps here)

Inferred types, for less typing and typos.
Tuples, very handy for lots of uses.
Simple iteration syntax.

None of these are new, but the ensemble is nice and clean in a compiled language. I like python for its simplicity and power, but it's generally interpreted. Swift is a but pythonic, but compiled.

Re:Matlab

[serviscope_minor]

It's more (or less) than that though: it isn't ObjectiveC. Seriously if I had to program in ObjectiveC otherwise, I wouldn't just love Swift, I would worship it.

Re:Matlab

[iris-n]

Matlab is the worst piece of shit I ever had the displeasure of working with. I only use this crap because the SDP interpreters I like (YALMIP and CVX) don't support anything else.

Once I had a bug in a code I had been writing for more than a month, and after lots of work I tracked it down to a vector declaration of the kind

v = [a, b +c];

You know what Matlab thought this meant?

v = [a, b, c];

Yep. It interpreted a whitespace between 'b' and '+' as creating a new vector element. And it's not even consistent, if I type [a, b + c] if gives again a two-element vector instead of a three-element one.

The fact that the programmer who introduced this "feature" was not instantly smitten by a thunder is proof that Zeus does not exist.

Btw, if you know any sensible SDP interpreter that does not require Matlab, I'd really like to know.

Re:Matlab

[slashdice]

Linus Torvalds said that about C++. For some people, there is a very strong C/C++ divide. I think from their perspective (I use C++ but didn't drink the koolaid) C++ encourages you to be lazy (yet overcomplicated!) and they have a point. Every time you type a character into the chrome address bar, there are 25,000 std::strings created. Shit like that wouldn't happen in C.

Maybe in C it's harder to do things, so you need to think about it more and do it correctly.

Re:Matlab

[macs4all]

Indeed. Swift is for the incompetent. There is nothing wrong with Objective-C, you just have to understand what you are doing. This is just one more step on the way to general cretinization of programming, and consequentially programmers.

I also agree that there is nothing wrong with Obj-C; but OTOH, is there anything wrong with attempting to update and/or clarify syntax?

For example, I doubt you would find that many developers that would argue that C++ can be arcane, and most that have developed in both would say that Obj-C is much "cleaner" than C++ overall; so why can't Swift be seen as simply a further refinement on Obj-C, especially for less "ambitious" projects?

Bottom line, I don't think Swift is in danger of displacing Obj-C anytime soon; so why all the hate?
br. Oh, I know: Because Apple.

Re:Matlab

[macs4all]

I have no idea who "they" are. I am saying this about Swift. Incidentally, nobody halfway sane would say that C++ or Obj-C is for the incompetent in comparison to C. They are all three hard to master and C++ is a pretty bad design in addition. (Of course, C++, like the equally bad Java, was never intended to be a general-use language.)

Yeah, but that never stopped Pascal!

Re:Matlab

[david_thornley]

Stroustrup made C with Classes as a version of C that had Simula classes. He liked Simula, but had to rewrite stuff into C for performance. I don't know Simula, so I can't comment on how transparent it was compared to C with Classes, but for Stroustrup it was a performance issue.

Re:Matlab

[david_thornley]

I haven't looked at Swift yet, but are the Swift features better than the C++ equivalents? "for (auto i: foo)" is about as simple as iteration syntax gets, tuples were added in the library, and C++11 introduced inferred types.

Re:Matlab

[david_thornley]

Why do you think that would never happen in C? It looks like a lot of bad design went into that, and C++ is easier to do strings well in.

Re:Matlab

[TechyImmigrant]

It certainly comes with fewer opportunities to shoot yourself in the foot in subtle ways than C++.

The Edsel had comfy seats, air conditioning, electric windows and a big engine, but that didn't make it a good car.

What has Rust been used for?

[Anonymous Coward]

Aside from the Rust implementation and Servo, is anything of note (and not written by the creators of Rust!) actually built using Rust?

Rust gets a lot of hype about how great it supposedly is. But I don't see anybody doing anything particularly interesting with it. As a systems programmer who uses C, C++ and Ada, I haven't been impressed by Rust. Its claims are lofty, its implementation is buggy, its syntax is mediocre, its standard library is convoluted, and it's taking forever for them to get a 1.0 release out.

So I think that anyone who claims to like it is either:
A) One of its creators.
or
B) A hipster who is repeating hype.

Re:What has Rust been used for?

[gweihir]

Rust is neither very good, nor finished. It promotes the myth that secure software is a question of the tool used. (It is not. It is 100% a question of coder, designer and architect skill and experience.) It just has a rabid follower community with little actual understanding of what they are cheering for and how it actually compares. That alone already makes it a problem, not a solution.

Re:What has Rust been used for?

[caspy7]

Rust hasn't even hit 1.0 yet.
Due to language changes many have chosen to wait until the 1.0 release. So no, it's not surprising not to see anything "of note."

Despite this they have a very active subreddit with many people coding things - especially now that they've hit beta and the language is fairly stable.

There have been a few notable interested parties (that I've seen), including indications that Google is playing around with it and another large project investigating core usage.

Re: What has Rust been used for?

[gweihir]

It is a rather strong hint that this Rust-hype is just hype. That said, Ruby is a pretty nice language, but in the hand of incompetents that does not mean anything. I does point to the actual problem though.

Re: What has Rust been used for?

[TechyImmigrant]

So you are confirming that the hype is without any basis in reality. People are claiming it's a good language without actually having tried to use it to build anything substantial. That is not encouraging at all.

Software security people are saying it's a good language because it has features that very much matter to software security and there's nothing else out there that is even attempting to solve the same problems.

That's not something without basis in reality. It's hard learned experience creating requirements that rust is out to meet.

Re:What has Rust been used for?

[lucm]

So I think that anyone who claims to like it is either:
A) One of its creators.
or
B) A hipster who is repeating hype.

It's called the Go syndrome.

Re: What has Rust been used for?

[jma05]

> there's nothing else out there that is even attempting to solve the same problems.

You mean move semantics? That would be the main innovation of Rust. C++ also has them now. Perhaps Rust has them better, but it would be inaccurate to say that no one is even attempting to solve these problems.

Re: What has Rust been used for?

[drewm19801927]

Don't worry, nobody is going to make you program in Rust unless it really proves valuable in practice. One good litmus test for this will be if Servo gets productized and ends up being the only browser ~not getting hacked every year at the pwn2own. Rust has a steep enough learning curve that you pretty much already have to be a good programmer to even get started. You have to not only grok what pointers/references are, there is the additional overhead of mutability and ownership that you have to understand to write even trivial code. I think once Rust matures a bit it will still have a shallower learning curve that C++, though; I have a hunch new users will learn faster from Rust's compiler errors than from C++'s segfaults (and data races once you get to parallel programming).

Re:What has Rust been used for?

[drewm19801927]

If you really have a significant amount of Ada experience, and spend enough time with Rust once it reaches 1.0 to get over the initial pain of a new language, please write a blog post about it, and be very constructive with your technical criticisms. I have yet to see an in-depth comparison between Rust and Ada; just a handful of short forum posts by the Rust devs saying that Rust has stronger safety guarantees for parallel code. I attended an Ada talk at a high profile open source conference recently and was a bit underwhelmed; I came away with the impression that aside from ranged integers, the compile time safety features of Ada (beyond memory safety) are a bit kludgy, i.e. they don't fit will into the syntax, are vendor specific, or only work in restricted cases... Of course, that could have been a problem with the talk, rather than the Ada ecosystem.

Re:What has Rust been used for?

[serviscope_minor]

It promotes the myth that secure software is a question of the tool used. (It is not. It is 100% a question of coder, designer and architect skill and experience.)

That's complete and utter rubbish, and like the best persistent rubbish has a kernel of truth. Yes the programmer matters (duhhhhhhhhhhhhhh).

But the tool matters too. Certain tools eliminate entire classes of mistake. Any mistake a tool prevents is one that cannot be made, even the the programmer is ill, sleep deprived and hung over.

Quiting errors are possible in shell, but not, say C++.

Buffer overflows are completely impossible in haskell.

etc.

Yeah perfect programmers can write perfect code. Trouble is perfect programmers don't exist.

Re:What has Rust been used for?

[gweihir]

But the tool matters too. Certain tools eliminate entire classes of mistake. Any mistake a tool prevents is one that cannot be made, even the the programmer is ill, sleep deprived and hung over.

You model is fundamentally flawed: Preventing certain classes of bugs does not make software more secure. That is a myth. People that do not have good skills will just make other mistakes instead, and, worse, a lot of "managers" will hire even worse coders if the language used is believed to be "more secure". That is why in actual reality production Java code does not have fewer vulnerabilities than production C code.

There is one area where language matters: If it is not well matched to the task, it drives up cost.

Fuck off Dice

[b1ng0]

Fuck off Dice we don't care about your shitty fluff pieces passing as news! You will not get any ad dollars from me and I encourage everyone else not to click or respond to this garbage.

Test of Time

[organgtool]

It's easy to love Swift now since it's relatively new. Enough time hasn't gone by yet for projects to grow big enough to discover all of its shortcomings. I did like many of the core concepts behind Swift when I first heard about it, but I'm not a fan of its low type safety as well as the fact that it only works on one platform.

Re:Test of Time

[bluefoxlucid]

Yeah, it's a fad language. We'll see if it sticks like Python, C#, or Java.

Re:Test of Time

[samkass]

This was my reaction to that comment, too. Swift is a strongly typed language, it just infers the type at compile time so the programmer doesn't have to manually enter it when declaring the variable. Considering it also differentiates between variables which can contain "nil" and ones that can't in its type system, I'd say it's one of the most strongly typed languages in common use, so I don't understand the statement.

The one platform thing is a bummer, but it's changing pretty fast right now so I don't blame Apple for not wanting to lock any decisions in with third parties yet. I hope they add it to their pile of open source projects before too long, though. Considering the reference implementation is LLVM-based, it shouldn't be hard for it to become very portable very fast.

(Besides, who doesn't like a language which has the entire unicode character set available for variable names, including the symbols? Can make for some colorful code.)

Re:Test of Time

[Dutch Gun]

It will stick because it was purposefully designed and deployed for practical use among a very broad audience. Objective-C would likely still be a niche language were it not the native language for developing products on Apple platforms. This mass deployment for Apple devices essentially assures it's success - along with the fact that it looks like a pretty decent language on its own merits.

Re:Test of Time

[RyuuzakiTetsuya]

I don't think Swift is ready for prime time just yet.

Given that Swift doesn't really let out any of Apple's Magic, I think we might see a source release for Swift someday. It's not like opening CocoaTouch or CoreVideo or CoreStorage.

I just want a platform independent release so I can build web apps on it rather than Python or PHP. It works through libcgi so.

Re:Test of Time

[gnasher719]

Besides, who doesn't like a language which has the entire unicode character set available for variable names, including the symbols? Can make for some colorful code.

Actually, you can create completely invisible variable names. Shame that superscript 2 and superscript 3 are not valid for custom operators.

Re:Test of Time

[Alomex]

Yes, the GP post is confusing variable declaration and strong type safety. Experience has shown that we want both.

E.g. "variable x is a string", so if by accident the one time I used it, I happen to write "x=3" I want the compiler to complain and not silently infer the type "x is an integer".

Second, we want variable declarations so that if you accidentally write "usrename" instead of "username" the compiler complains.

Swift doesn't require you to give the type during declaration, which is a minor saving at a cost of many headaches. This is the wrong design decision, though I don't think is particularly critical.

Variables do need to be declared, which is a nice improvement over python and one of a few big things holding python back from total dominance (the others are white space, interpreted not compiled and somewhat weak pointer/data structure support).

Re:Test of Time

[Xest]

Whilst I don't really know anything about Swift, what the GP describes makes it sound like the features he's describing are basically exact copies of the way C# does things.

If so, then what you describe isn't really correct. In recent versions of C# you can simply declare a variable with "var". The situation you describe where it determines something to be an integer when you wanted a string shouldn't be an issue because in C# what would happen is the initial assignment of an int would have the compiler treat it as an int, and then any subsequent attempt to use it as a string would throw a compiler error because the compiler already figured out that it's an int by that point.

So you still declare your variables, you just let the compiler automatically determine the type.

This said however, I'm still not a fan of it, I don't like C#'s var keyword. It makes it easy for sloppy developers to write code quickly, but it kills readability and maintainability for other developers, because they have to expend effort figuring out what type the variable is actually meant to be rather than being able to just look at the definition to see what it's actually declared as.

It's a feature that's been added in to try and woo sloppy developers, but frankly all it does is reduce the average level of code quality by allowing people to write such sloppy code in the first place.

Re:Test of Time

[Alomex]

It's a feature that's been added in to try and woo sloppy developers, but frankly all it does is reduce the average level of code quality by allowing people to write such sloppy code in the first place.

Exactly. It is a false convenience.

Re:Test of Time

[gmiller123456]

I can certainly see the desire for things like "var", but I actually think it's something that could be better served by an IDE rather than the language itself. E.g. type "var i=3", IDE changes that to "int i=3". Many IDE's today do a good job of suggesting variables based on types of parameters, but so far I haven't noticed any a variable's type from a return type of a function.

Re: Test of Time

[MemeRot]

var is useful when used with factory functions. Makes changing the type returned by your factory function much easier if all the calls use "var x = MyFactory.GetConnection("foo");" Otherwise you may find yourself replacing 3000 SqlConnection declarations with DbConnection as we had to do.

Re:Test of Time

[david_thornley]

In C++, you can write "auto x =", and if you initialize it with a string then it's a string variable. You can't use "auto" unless the compiler can deduce the type.

It's useful in several ways. It allows templates to be much more type-flexible. It makes loop syntax much cleaner. There are things in C++ where it's difficult or impossible to write the type (such as lambdas), and "auto" allows assignment in those cases.

Re:Test of Time

[garote]

A very _wealthy_ "niche group" of developers, too.

C is "cross platform" the way that a tricycle is an all-terrain-vehicle: Easy to roll out the driveway, useless for anything ambitious. C++ is better but better alternatives are everywhere now. Javascript is half a solution at best: You need back-end support, and nodejs is just a less useful alternative to Python. Plus there are things you really just can't do in Javascript. Your examples of "cross-platform" all have serious flaws, which should not be surprising.

I noticed you completely failed to mention Java, which was hailed as the total cross-platform solution for a while until web browsers started crapping on it, and now it's synonymous with Android programming and would be decisively in the rear-view mirror of the tech industry by now, if not for that. What's funny is, Google almost had to choose it by default. What were they going to use instead? C# from Microsoft? ObjC from Apple? What else are you going to implement and entire OS in that isn't 20 years old?

At this point it's clear that politics is just as much a factor as intrinsic quality when determining the fate, and staying power, of a language. Javascript would be a lot cleaner now if it didn't have to support a dozen years of really crappy legacy on the internet. Visual Basic would have been a brief, unpleasant fart if it hadn't been crowbarred into every Office product for a quarter century.

If you're placing bets that Swift will dry up and blow away because Apple is due - any day now - to do the same, you're probably a little TOO old-school. You know what will die before Swift dies? In terms of popularity and profitability? C#, because its fate is tied almost entirely to Microsoft. And that's not going to die for quite a while.

Why the hate for VB

[JerryLove]

Modern (.NET) VB is nearly C# with more english-like syntax. I don't understand all the hate for the language.

Gonna agree with SharePoint though. :)

Re:Why the hate for VB

[LWATCDR]

Sharepoint and Wordpress are languages?

Re:Why the hate for VB

[bsDaemon]

They said "languages or technologies" in the summary. Sharepoint, WordPress and SalesForce are platforms that apparently people who responded to the survey don't want to work with.

Re:Why the hate for VB

[PRMan]

Other than working with interfaces, where C# will build a framework for you and VB will not, I really see very little difference to account for the hatred. In fact, the string functions in VB are actually preferable.

Re:Why the hate for VB

[Tom]

Same reason that fascism and communism are unlikely to win any elections anytime soon - the name has been tainted by a horrible first version, even if you came up with a perfect current version, nobody would believe it.

Re:Why the hate for VB

[Marginal Coward]

The time was on this site where you would have been laughed off praising a Micro$oft proprietary language. How times have changed. Pathetic.

Maybe at some point somebody directed those folks to this.

Re:Why the hate for VB

[angel'o'sphere]

Why are variables declared with DIM in VB, or DIM ... AS.
You really think that is english?

Re:Why the hate for VB

[careysub]

...Almost always, a goto statement indicates sloppy design on the part of the coder. I think I have only come across one instance in my professional life where a goto was actually not a bad option (maybe even the best, or least worst option). And I've been coding for around 30 years now. Also, there is a reason why coders almost instantly fell in love with the object-oriented paradigm. Almost overnight, it cleaned up a lot of code. Granted, it is not a perfect paradigm, but it does seem to work well in a surprising number of cases. Just sayin'.

Structured programming constructs already did the heavy lifting on the Curse of the Go-To, a decade before OO languages became generally available.

But yes, with exceptions available (handling error breakouts in otherwise clean logic) the last reason to use a go to died.

Re:Why the hate for VB

[narcc]

Wait, that's your argument? Fewer keystrokes?

How incredibly stupid.

Re:Why the hate for VB

[Keruo]

It's an abbreviation: Dim

Re:Why the hate for VB

[Xest]

Because VB brings us such fucking abominations as:

If myVar IsNot Nothing AndAlso myVar = "something" Then
' do something
End If

The problem with VB is in it's attempts to be English like it's just ended up requiring you to spout nonsense. No one says "Is not nothing", they say "Is something".

It's too verbose and ends up forcing you to write stuff that's inherently less readable than if it didn't try and fudge English into it's syntax.

Re:Why the hate for VB

[angel'o'sphere]

I know what DIM means ;D
As many others here I started 30 years ago with Basic.

Hence I know it makes no sense and it is certainly no english to write in VB something like this:

DIM myObject AS SomeType

Because that object has no DIMension! Keywords like VAR or VAL or even DEF would have been more appropriated.

Re:Why the hate for VB

[david_thornley]

All other things being equal, terseness is good in a programming language. Verbosity means things are harder to type, and display less actual code in a given screen area. I'm a very good typist, but I'd get frustrated at typing longer keywords unless there was an advantage.

Re:Why the hate for VB

[narcc]

The advantage is usually in readability. Programming languages are, for the most part, terse enough. Most could benefit from a little verbosity. Brevity is not always beneficial.

Java is interesting as it's not the syntax and keywords that lead to it's absurd verbosity. That's a function of the language itself and the JCL.

Re:Why the hate for VB

[JerryLove]

Because VB brings us such fucking abominations as:

If myVar IsNot Nothing AndAlso myVar = "something" Then
' do something
End If

The problem with VB is in it's attempts to be English like it's just ended up requiring you to spout nonsense. No one says "Is not nothing", they say "Is something".

It's too verbose and ends up forcing you to write stuff that's inherently less readable than if it didn't try and fudge English into it's syntax.

Right: because if (myVar != null && myVar == "something") {
'do something
}

Yea. That's much closer to English. I routinely say things with more special characters than letters in them.

And that's a terrible line.. since it must be not nothing to be something. You could have cut the entire first part.

I'm a big fan of self-documenting code. If you want to give up readability to gain brevity; then I'd hate to try to follow your variable and function names.

Re:Why the hate for VB

[david_thornley]

I find readability to be a non-issue, personally. I dislike having to type unnecessarily long keywords, and that would go into whether I like a language or not.

My identifier names tend to be fairly long, but that's my decision, and I can see the benefit directly.

Re:Why the hate for VB

[Xest]

You seem obsessed with the idea that programming languages are better if they're close to English, but this is a long debunked theory so you're living in the 70s/80s on that one. If closeness to English was an overriding priority then we'd all be using COBOL.

We don't however, because regardless of closeness to English we still have to learn the syntax, and if we have to learn the syntax either way then we read code as if it is English. Anyone that knows C style syntax can read your C style example just as easily as they read plain English, and yet the VB style syntax requires you to parse a double negative which is bad English.

As such, all you're left with with VB is bug inducing double negative mindfucks, and increase unnecessary verbosity resulting in lower productivity.

It boils down to this:

"Is not equal to null"

vs.

"Is not nothing"

The former is perfect English, the latter is terrible, broken English.

The only way you can have a programming language that works with plain English is by allowing it to have a large number of keyword combinations, so that you can express "Is not nothing" as "Is something". Until you do that attempts to create a programming language in plain English are a long verified dead end.

Don't try and pretend my example was intended to be anything other than an example of VB's terrible syntax and verbosity. If I was creating a demonstration of great coding style I'd stay out of any VB discussion in the first place because VB is the antithesis of that. There is nothing readable about VB's syntax because it's broken English end to end and that gets in the way of clean syntax that can be read logically.

Have fun writing low readability code with your reduced levels of productivity if you enjoy spitting out such unnecessary verbosity if that's your thing, but don't try and pretend it's superior. There's a reason VB is hated and unpopular, and that's because it's shit for the reasons I've described here, if you think otherwise it's not because you're some super coder who just sees something no other coder gets, it's just that you're an inept VB fanboy.

Re:Why the hate for VB

[drakaan]

Actually, all variables have a dimension. Think of your "myObject" as having dimensions of [1].

Re:Why the hate for VB

[angel'o'sphere]

That is a pretty dumb way of looking at it.

It might work in plain old C like that, but in most other languages you get a compiler error.

Re:Why the hate for VB

[drakaan]

You get a compiler error from doing what? I wasn't giving a code example, I was noting that scalar variables occupy space in memory and have to be able to be referenced, hence they have a dimension.

What's your unmentioned better way of looking at it?

Re:Why the hate for VB

[angel'o'sphere]

Sorry, the discussion is pointless.

Sane languages use: "def, var, val, let" or the type to declare a variable. Only VB uses DIM.

So what exactly do you want a new programmer to learn? "DIM" is the keyword you use to declare/define a variable?

Wow, then you tell him: "DIM" means dimension. "Hu?" What exactly did he learn now?

He learned he has to remember an arbitrary bad chosen keyword to do something that has nothing to do with the meaning of the word.

And now you come and try to tell me that every variable has an implied dimension of 1?

And that makes it "logical" ... or how ever you want to call it, that you declare variables with the keyword DIM in VB, instead of using a keyword like "var"?

How exactly do you even declare an array in VB?

DIM myVar as myType of Array ...
or: DIM myVar as myType[dimension] ????

How retarded does that sound if you speak it out loudly?

Pascal:
myVar : Array [1 .. 99] of integer;

Or Ada:
myVar is Array from 1 to 99 of integer;

The discussion was about how easy to read/pronounce/spell a programming language.

Some idiot proclaimed VB is the easiest one. No it is not. It is the worst.

Re:Why the hate for VB

[drakaan]

I'm not arguing that VB is good, but that's a pretty trivial example of what makes it bad. Here's a discussion of where DIM came from: http://stackoverflow.com/quest... You used to be able to just say "variablename = value" without a dim statement. That's improved in the current syntax for VB.Net. If you want an example or two of what makes VB bad look at some of the legacy VB6 and classic ASP code hanging around out there. PHP has some (most) of the same problems.

Re:Why the hate for VB

[angel'o'sphere]

And strictly speaking both top voted (and the accepted) answer on stackoverflow are wrong.

The important thing about DIMension is: it does not really refer to the size of the array, but to its dimensions.
That means: is it a one dimensional array of a certain size, or a multidimensional one, e.g. a matrix or some cubic structure.

Hence from a standpoint of a mathematician or physician the keyword made sense. Especially as it only was needed for arrays over a certain size and with more than one dimension.

A short array, with less then lets say 30 elements (don't remember the limit) you could use without declaring the dimensions/sizes.

Perhaps we once will have a main stream language that makes sense from a linguistic point of view as well as from a programmers point of view ...

So is it REALLY good?

[txoutback]

Is it good, or does it just suck a lot less than Obj-C?

Re:So is it REALLY good?

[bluefoxlucid]

Having used both Objective-C and C++, I can point out Objective-C as being a vast improvement over C without bringing the major drawbacks of C++. For one thing, Objective-C specifies symbol resolution against what is actually available, rather than what is theorized: in C++, if you recompile a library with a new class member--even a new private member--the data structure of the C++ class is now incorrect, and all applications using that library must be recompiled; in Objective-C, you can extend a new interface onto a class.

Largely, Objective-C is C with classes. Along with that, it's C, however you want to take that, with a better OOP implementation than C++. It's not Python or C#.

Re:So is it REALLY good?

[serviscope_minor]

All depends what you mean by "better".

I primarily use C++ classes for encapsulation. I have a bit of inheritance, but often use it as an aggregation mechanism where you import the parent's functions (see e.g. curiously recurring template pattern). Compile time resolution is 100% necessary for high performance code, and thats largely what I write.

Also the C++ model allows classes to look very much like builtins, so the code uses the language syntax, so containers etc don't feel worse than arrays.

IOW it depends on what you use it for. Neither is definitively better.

Re:So is it REALLY good?

[bluefoxlucid]

Compile time resolution is 100% necessary for high performance code

When you first run a C++ program, it runs through a list of symbols, resolves them against library addresses, and puts their addresses into a procedural linkage table (PLT). Each time you CALL, you get an indirect jump to an address in the PLT. In lazy resolution, the PLT is populated with a resolver function address, which does the symbol look-up and populates the PLT.

When you run an Objective-C program, it does exactly the same thing; however, the PLT lookup is done through a function that provides symbolic lookup, symbol to symbol, instead of address to symbol. It's effectively the same look-up (it looks in the symbol table), but handled differently (the structure of the class has to be resolved at run time, and the PLT can be later altered by using a different object and changing how the class looks up its members).

It's not much different. It resolves on first, unknown call; then it continues to use the cached address.

Re:So is it REALLY good?

[serviscope_minor]

When you first run a C++ program, it runs through a list of symbols, resolves them against library addresses, and puts their addresses into a procedural linkage table (PLT). Each time you CALL, you get an indirect jump to an address in the PLT. In lazy resolution, the PLT is populated with a resolver function address, which does the symbol look-up and populates the PLT.

That only happens for some functions. Very many, almost certainly the vast majority in terms of numbers called get inlined out.

Also, the way ObjC works, I'm assuming that classes don't support stack allocation, because that requires compile time knowledge of the size, which implies you can't make arbitrary changes without recompiling. Stack allocation is of course much faster (it's free!).

The result is that the C++ class model works exceptionally well for large numbers of tiny classes (as abstraction---not a virtual function in sight), things like std::complex and so on.

As with all things it loses flexibility, and from what I gather, ObjC has substantially more in terms of run-time tweaking of stuff.

Re:So is it REALLY good?

[bluefoxlucid]

I'm talking about things that are external, not inline. Do you know how many libraries Chrome, Firefox, LibreOffice, Gnome, or ls load?

[c_jmoser@vmohssibg002 ~]$ ldd /bin/ls linux-vdso.so.1 => (0x00007fff79487000)
librt.so.1 => /lib64/librt.so.1 (0x00007fe67c310000)
libacl.so.1 => /lib64/libacl.so.1 (0x00007fe67c109000)
libselinux.so.1 => /lib64/libselinux.so.1 (0x00007fe67bef1000)
libc.so.6 => /lib64/libc.so.6 (0x00007fe67bb98000)
libpthread.so.0 => /lib64/libpthread.so.0 (0x00007fe67b97b000)
/lib64/ld-linux-x86-64.so.2 (0x00007fe67c52a000)
libattr.so.1 => /lib64/libattr.so.1 (0x00007fe67b777000)
libdl.so.2 => /lib64/libdl.so.2 (0x00007fe67b573000)
libsepol.so.1 => /lib64/libsepol.so.1 (0x00007fe67b32c000)

Re:So is it REALLY good?

[serviscope_minor]

I'm talking about things that are external, not inline. Do you know how many libraries Chrome, Firefox, LibreOffice, Gnome, or ls load?

Yes, but I was specifically referring to those inline. Compile time resolution happens a LOT in C++, and makes things go very fast. Combining that and stack allocation you get custom classes which look and feel like types built in to the language.

That only works with compile time resolution with classes.

You can (and one often does) use C++ classes for abstraction not traditional OO programming.

Re:So is it REALLY good?

[bluefoxlucid]

That only happens on modules compiled into the same source file, not on anything external. That always happens: Compilers totally optimize out any interfaces for things that aren't interfaced with.

Re:So is it REALLY good?

[serviscope_minor]

That only happens on modules compiled into the same source file, not on anything external.

Not any more: GCC supports whole program optimization, as do other compilers. But regardless, C++ has a class encapulation system which allows early binding, which ObjectiveC does not, whereas ObjectiveC has a class system which allows very late binding in all classes. Anything related to a template class will have the methods defined in the translation unit, so those will be subject to optimization out. I don't think you can override functions within a program either. You can at shared object boundaries of course. Assuming you haven't compiled a static executable.

I gather in ObjC you can pretty much override everything to do with any class at any stage. The downside of flexibility is speed.

Anyway, my point is that C++ and ObjC have two quite different class systems. Neither is objectively better since neither has the capabilities of the other.

Re:So is it REALLY good?

[bluefoxlucid]

So you're telling me gcc inlines functions from libqt when you compile a program against qt's headers?

But regardless, C++ has a class encapulation system which allows early binding, which ObjectiveC does not

Binding is still done at run-time, and is still a matter of setting a pointer somewhere. It's functionally similar to a page fault: if it happens and it's satisfied, the computer just calls as normal; if it happens and isn't satisfied, the computer does some work first. All you're changing is the point in time the work is done, not the number of times.

Re:So is it REALLY good?

[serviscope_minor]

So you're telling me gcc inlines functions from libqt when you compile a program against qt's headers?

I believe anything linked in statically will do yes, if the archive contains object files with the correct metadata.

But anyway what's that got to do with it? C++s object model allows for this, where as ObjC explicitly doesn't.

Binding is still done at run-time, and is still a matter of setting a pointer somewhere.

For a lot of stuff no. Write some code with std::vector and compile with -O3. Then look for the calls to the operator[]() function. They'll all have disappeared completely.

Re:So is it REALLY good?

[bluefoxlucid]

I don't think the linker does that sort of inlining. It would require complex compiler optimization. It just resolves the PLT to the local binary.

Re:So is it REALLY good?

[serviscope_minor]

I don't think the linker does that sort of inlining. It would require complex compiler optimization.

It does, and it does. Well, actually the linker is a fake. GCC writes out ELF object files with the normal object code as usual and also some intermediate representation which can more or less be loaded by the compiler. It then loads up all the intermediate code and runs the optimizer on it, before spitting outthe final object code.

It's a pretty complex process, especially when multiple compiler versions exist which is why it took from 4.5 when it deuted to 4.9 before it was any good.

I believe the GOLD linker has plugin support and so when linking .a archives, the plugin can call back to gcc to start up and do the linking.

http://hubicka.blogspot.co.uk/...

A the new and shiny

[prefec2]

Yes, the new and shiny is most popular. How could that happen?

By the way, it might be interesting to see this "most loved" thing filtered for "full time employee" and "full time freelancer". I assume that then other languages might appear in the list on top. But anyway, the stack overflow statistics are interesting.

Most-loved or Most-infatuated?

[QuietLagoon]

It is way too early in Swift lifetime to come to any sort of real opinion about using Swift and maintaining Swift projects over the course of many years.

.
At best, this is little more than puppy-love.

Re:Most-loved or Most-infatuated?

[PRMan]

Jesus said, "Those who have been saved from much, love much." IMO, this is more an indictment of the years of suffering in Objective C than anything else.

Re:Most-loved or Most-infatuated?

[ameoba]

"There are only two kinds of languages: the ones people complain about and the ones nobody uses"

Re:Most-loved or Most-infatuated?

[euroq]

Of course, good code is 99% due to the person producing it and only 1% (if that) due to the language used

I agree that it's at least a majority of the person producing it, but I wouldn't say 99%. Good language design, in the right and appropriate circumstances, do indeed help. For example, Java (which you mention as bad) actually makes it MUCH harder for software to crash. For example, I lead a mobile team with Objective C (iOS) and Java (Android). The number of users experiencing crashes on the iOS app is around 2.5%, while the Android app is a perfect 0%. I don't think the Android devs are any better than the iOS devs - I am certain it's the language.

For the record, I have nothing good to say about C++ :)

Re:Wordpress is not a language

[prefec2]

The title states languages and tech. This includes platforms, frameworks, libraries etc. And I do not see where the problem is with wordpress if it is used as a blog.

My most loved language

[Anonymous Coward]

Is whatever pays the bills. C++? I love it. Java? Can't do any better. C#? It's Java but better! Ruby? I think I just came, pissed and shat myself at the same time. Groovy / Dart / Swift / Python? I think I spurted so I hard that some went in my mouth.

Most languages are trivial to learn. Most use familiar concepts and conventions so it takes maybe a month to learn the syntax and the frameworks to do stuff. Anyone who makes the effort to keep pace will never find themselves out of employment. Pick a stalwart language and learn some trendy ones on the side. Do some practical projects in those languages to make yourself proficient and then throw it on the resume. You can't go wrong.

Re:My most loved language

[serviscope_minor]

I think I just came, pissed and shat myself at the same time.

I think a urologist may be of more use to you than slashdot at the present time.

Re:Another example of how closed source rules

[epyT-R]

but closed sores leave scars...

Re:Swiftly first post

[Keruo]

Didn't you get the memo?
We're replacing systemd with systemVBd, it'll be vastly superior with its familiar basic syntax.

Re:forced by instructor...

[TechyImmigrant]

Our senior project/software engineering class/writing intensive capstone class... was a JOKE... the school formed a second section when they saw how many wanted into the class. The prof was paid 1/3rd of a normal class.. thought he was only getting a few students... and ended up with 25.

Anyway... the class turned "learn swift at the 101 level" for 2 months and work on a meaningless project you would never, ever show a prospective employer for the other 2 months.

I hate swift. It's like writing javascript with a keyboard made of razor blades. Nothing ever works. Nothing ever feels right.

I suspect you hate xcode. Swift is fine. xcode is a project obfuscation system.

VB-Classic, VBA & VB-Script [Re:Why hate for V

[Tablizer]

The survey respondents may have meant VB-Classic, VBA, and/or VB-Script.

Incidentally, they are generally fine for smaller projects in my opinion. It's when you try to build something complex with them that you get into knots. Languages best for big projects are rarely best for small projects and vice verse. Use the right tool for the job.

Re:Don't see it

[gTsiros]

holy shit

a three-digiter

quick, where's my bug-net, before it flies away

Can't say I love it *yet*.

[jcr]

Coming from many years of Obj-C development, I can acknowledge several ways in which Swift is superior, but the learning curve is somewhat steeper than the transition from C to Objective-C was.

Aside from the language itself, Swift playgrounds are wonderful. We're getting closer all the time to a Smalltalk way of writing code.

-jcr

Swift loved 'cause ObjC (and frameworks) supported

[SuperKendall]

I've been using Swift for production work since shortly after it was released (much of it on an internal enterprise application which is how we were able to start using it right away).

What makes Swift really nice to use in its own right is that it has a lot of useful language features (like closures, generics, tuples, etc) with a syntax that can be kind of boiled away to the degree that you choose, to keep code clear and understandable. I think the best way I could describe it, is that it's like a functional language buy is very practical and doesn't get preachy about it.

So already the language is very pleasant to use. The real benefit Swift enjoys that give it such a high rating though, is that it comes with very advanced tooling and a super-integrated mirror-counterpart language (Objective-C) right out of the box.

Think about it, how many new languages like Rust suffer because you have to build up syntax highlighting support in the editors you like, figure out a new build process tailored to that language, how to run the applications and so on. With Swift if you knew XCode you could easily just start writing Swift and all of the annoying overhead was gone. Even if you DIDN'T know XCode, at least it's a pretty advanced tool dedicated to helping produce running code in very short order (VERY short order with Playgrounds).

Then along with that, you have a new language which invariably has some missing features or capabilities, that make some particular thing you are trying to do hard in the new language. Well in those cases, Objective-C is very close at hand - you can mix code from both languages easily in the same class even. For example Swift itself is strongly typed and has very few reflection or dynamic method lookup features yet. Objective-C is kind of the opposite way, full of dynamism and runtime reflective use, so you can jump over to those abilities as needed.

I don't think people outside the iOS community realize just how fast everyone doing iOS development is switching to Swift. Swift (for me) has actually worked really well since day1, the tooling was rough for a while (with the syntax highlighter/code completion crapping out regularily on Swift code) but I THINK it may finally be OK.

It's definitely not a case of people hating Objective-C, because a lot of the people that like Swift also liked Objective-C. It's a case of having some good tools already, and being given another tool that seems to work really well for some tasks and thus appreciating having an expanded toolbox...

Re:Don't see it

[MassacrE]

Damn, a sub-700-er!

Re:Don't see it

[Enry]

n00b

Have to wonder which will come first

[SuperKendall]

Rust adoption seems to have been slow. Given that Swift took a lot of ideas from Rust, and is evolving more rapidly to completeness, I have to wonder if Swift will not take over positions Rust holds before Rust gains much of a foothold...

That's all predicated of course on Swift being released as open source, which will probably happen in a year or so.

But... I love them!

[lucm]

Fuck off Dice we don't care about your shitty fluff pieces passing as news! You will not get any ad dollars from me and I encourage everyone else not to click or respond to this garbage.

Why are you so harsh with them? It's a great company that creates interesting, unbiased and entertaining content!

Oh wait, I thought you meant Vice, not Dice. Sorry about that.

Re:Don't see it

[jpatters]

Damn, a sub-800-er!

Gotta love goto

[lucm]

There is a reason why people despise goto statements. May you never have to learn from personal experience what those reasons are.

I once had the "opportunity" to work with someone who was using goto to implement a fascinating code reuse strategy. Basically he was copy-pasting old code from previous projects and using goto to jump over lines of codes that were irrelevant in the new project. The best part was that he was *commenting out* old gotos, in case they could be repurposed.

That guy really loved goto, although he frequently complained about the fact that you can't pass variables to goto.

Re:Relative

[lucm]

Yeah. That's like this bleeding-edge feature now available in Oracle 12c: "multitenant architecture with pluggable databases".

Thanks to this fantastic innovation (which requires an additional license), it's now possible to have multiple databases hosted on the same Oracle instance! Crazy.

Of course this has been available (with no extra fee) for decades in SQL Server, MySql, Postgres and others.

Fluff

[lucm]

Those "clean" features have been in C# for years. Even java has now somehow the equivalent.

None of this gives Swift any kind of "cleanliness" edge. This is pure fluff, nothing else.

The reason for Swift is that Apple just doesn't want to compete head-to-head with the Android framework. There is simply no other reason.

Re:Fluff

[TechyImmigrant]

Yes. But why the hell would I program for windows?

Languages come and go. Some are great. Some are not. So far swift is ok.
 

Re:Fluff

[macs4all]

The reason for Swift is that Apple just doesn't want to compete head-to-head with the Android framework. There is simply no other reason.

Oh puh-lease. Do you even believe that?

Yeah, Java is the pride and joy of the Android Development community. Riiiiight.

HOW much bitching have I heard here on this site alone about Java? Hmmmm???

You're just Butthurt that Android doesn't have anything as clean and modern as Swift.

Re:Fluff

[lucm]

The reason for Swift is that Apple just doesn't want to compete head-to-head with the Android framework. There is simply no other reason.

You're just Butthurt that Android doesn't have anything as clean and modern as Swift.

See, you don't bring anything new in the conversation, you only repeat meaningless buzzwords already said by others. A lot like what Apple has done for the last 10 years.

As for Swift vs Java, nobody in this thread has provided actual points proving that Swift is indeed a superior language. Which pretty much demonstrates the fact that it's more Apple fenceware, and nothing else.

Re:Relative

[DrXym]

That example says more for the suckiness of Obj-C than it's replacement. Maybe that's explains why Swift devs are so joyous to escape this BS.

Re:Relative

[trumpetplayer]

Interesting!, and you wrote the same information on the relevant Wikipedia page too!

Re:Loved ... and ... C++???

[cwingrav]

Yeah, but it's familiar and behaves how you would expect. There's something to be said for familiarity.

Re:Loved ... and ... C++???

[david_thornley]

C++ has got a fairly nasty learning curve if you want to do things right, but it's extremely powerful and extendable. The syntax is frequently ugly, but there's a lot you can do with it. Perhaps more importantly, there's a lot of really powerful stuff that takes one guru to write and is reasonably easy to use thereafter.