Goodbye, World? 5 Languages That Might Not Be Long For This World

Nerval's Lobster writes As developers embrace new programming languages, older languages can go one of two ways: stay in use, despite fading popularity, or die out completely. So which programming languages are slated for history's dustbin of dead tech? Perl is an excellent candidate, especially considering how work on Perl6, framed as a complete revamp of the language, began work in 2000 and is still inching along in development. Ruby, Visual Basic.NET, and Object Pascal also top this list, despite their onetime popularity. Whether the result of development snafus or the industry simply veering in a direction that makes a particular language increasingly obsolete, time comes for all platforms at one point or another. Which programming languages do you think will do the way of the dinosaurs in coming years? With COBOL still around, it's hard to take too seriously the claim that Perl or Ruby is about to die. A prediction market for this kind of thing might yield a far different list.

Comments

If you wanted us to believe your Op-Ed...

[Bill_the_Engineer]

You shouldn't have made Perl and Ruby #1 and #2 respectively. Of course being on dice.com should have been enough.

On the plus side, I didn't waste much time reading the rest of the BS.

Re:If you wanted us to believe your Op-Ed...

[Anonymous Coward]

Of course! It's click bait from the parent company of Slashdot.

Re:If you wanted us to believe your Op-Ed...

[Bill_the_Engineer]

That can't be it... There are plenty of young and trendy languages that far outrank Perl and Ruby in the known-shithole category. At least they don't depend on white-spaces to define blocks.

Re:If you wanted us to believe your Op-Ed...

[i kan reed]

lol
"Syntax that every programmer uses to make their program readable is unreasonable as a semantically meaningful syntax"

Come on, python's got its problems, but forcing you to lay out your program in a naturally readable way to compile isn't one of them.

For example, duck-typing might be one of the worst ideas in the universe, because it's doing the exact opposite of the whitespace thing. It's decoupling easy-to-make mistakes with the output of compiling of your code.

But this whining about whitespace just comes off as having never actually tried it.

Re:If you wanted us to believe your Op-Ed...

[geminidomino]

If that were the case, PHP would have occupied all 5 slots.

Re:If you wanted us to believe your Op-Ed...

[Bill_the_Engineer]

I use Python. I have to also work with Python code written by other people.

Autopep8 is your friend and really it shouldn't have to be.

Re:If you wanted us to believe your Op-Ed...

[CastrTroy]

Whitespace is great, I use it all the time. I use the IDE to automatically fix the indenting of my code. That doesn't mean the compiler should attach any specific meaning to how things are indented. For one thing, once you let the level of indentation govern the execute, you lose the ability to auto-format your code. You can't autoindent the code based on where the blocks start and end, because the level of indentation defines where the blocks start and end. If you messed up the indentation at some spot, there is little to no indication that something went wrong. It's the same reason I don't like languages that don't require that you define that variables exist. Letting you define variables just by using them leaves you open to spelling mistakes and all the bugs it causes just so you can save a couple lines of code.

Re:If you wanted us to believe your Op-Ed...

[Bill_the_Engineer]

I don't like languages that don't require that you define that variables exist.

If only Python and Ruby offered something similar to Perl's "use strict".

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Duck-typing is fundamental to program polymorphism. It wouldn't be an object-oriented language without duck-typing.

I like Python. It's a strong, dynamically-typed language, and strong typing is critical for programmer success. Many a perl programmer have spent their frustrations trying to track down a problem caused by passing a string as an integer; and many have left forehead-shaped craters upon their desks and keyboards trying to make Perl interpret a string consisting entirely of numerical characters a string rather than an integer. That the interpreter believes types are fixed once set is a godsend: broken code should break immediately and with regularity, rather than hobble along until specific conditions are met.

Re:If you wanted us to believe your Op-Ed...

[Anonymous Coward]

It wouldn't be an object-oriented language without duck-typing.

lol no. An object-oriented language is language in which you can easily write object-oriented code. Object-oriented code does not imply dynamic typing.

broken code should break immediately and with regularity, rather than hobble along until specific conditions are met.

Broken code that can break in the compiler should break in the compiler.

Re:If you wanted us to believe your Op-Ed...

[Wycliffe]

I would like to see a list of languages that are not known-shitholes. Yes, perl and ruby have their problems
but so do python, c, c++, php, and java. Each has a place where they excel and each has places where they suck.
Name a programming language that doesn't start to have major problems somewhere when you start getting
into the details.

Re:If you wanted us to believe your Op-Ed...

[phantomfive]

Duck-typing is fundamental to program polymorphism. It wouldn't be an object-oriented language without duck-typing.

Duck typing becomes a problem when your program becomes big and lots of people are working on it, because it removes compile-time checks. Unless everyone on their project is extremely careful to write readable code, then random bugs will start popping up because it's not checked at compile time. Although I'm sure you personally write excellent, readable code so that's not a problem.

Also, come on, duck-typing is not necessary for polymorphism, inheritance and defined interfaces work just as well.

Re:If you wanted us to believe your Op-Ed...

[i kan reed]

No, strongly typed languages don't use duck typing. Even some weakly typed languages can still insist on interfaces being fully implemented. Python's duck typing falls apart specifically when the gate conditions don't update (possibly in distant parts of code) to match functional changes.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Programs with inheritance have duck-typing. You just define specifically which ducks are still ducks.

Re:If you wanted us to believe your Op-Ed...

[PRMan]

Oh PROGRAMMING languages! I was in here looking for German.

Re:If you wanted us to believe your Op-Ed...

[phantomfive]

ok, that's a different meaning than people normally give to duck-typing but whatever lol

Re:If you wanted us to believe your Op-Ed...

[3dr]

Python is my go-to language for quick code sketches, framework ideas, etc. That's the power of dynamically typed languages, it's very easy to throw code together to test ideas, and is what I value in "scripting" languages.

As much as I like Python, even with it's quirks like len() is a function on a sequence not a member, the one thing I despise is the whitespace-describes-structure. I have lots hours due to an auto-format of code run amok. Suddenly, all the code following an if-statement is now the body of that statement. It just doesn't make sense to not have block delimeters. With every other meaningful language under the sun using curly braces, why couldn't Python? I like the *idea* of clean code like Python code, and I enjoy reading Python code, but I prefer to have explicit block syntax.

As an aside about spelling mistakes, I agree, and Python doesn't help you there (unless you are reading a misspelled class field). One trick I use to fortify larger Python programs is to define slots on each class to explicitly define the members. If your code accesses a mistyped member name, that name will not be in the __slots__ list and the python runtime will raise an exception. Not only do __slots__ protect you from name typos, they are faster than regular fields for some reason. I've shared this tip with other pythonistas, and nobody else has heard of doing this; I can't believe others aren't doing this, too.

Re:If you wanted us to believe your Op-Ed...

[gstoddart]

Come on, python's got its problems, but forcing you to lay out your program in a naturally readable way to compile isn't one of them.

So, here's my problem with whitespace being syntactically significant ... everybody likes to see code with different levels of indent. There isn't one "naturally readable" way which everybody agrees on. And then suddenly you have a language which says "we're all stuck with whatever the whiniest coder wants".

Years ago I had a co-worker. He liked emacs, and I liked vi. OK, whatever, that's not a big deal, I can overlook that.

The problem was his electric mode in emacs was thinking itself oh-so-clever, and instead of storing the *actual* number of tab indents or whitespace, it just stripped them in favor of a single tab that emacs would then know how to render later.

Unfortunately, not all editors rendered the same way his beloved emacs did. So after he edited a document, every line showed as a diff, and every file he edited now rendered as gibberish in every other editor, because there was only a single whitespace character.

He was one of a team, but wouldn't listen when we said his editor was fscking up the code, and making it impossible for us to work on it.

So, where he wanted to see two character indents, and I wanted to see 8, and a co-worker wanted to see 4 ... once he'd edited it, in anything not using this electric mode didn't work.

Finally the solution was to lock him out of CVS, and tell him in no uncertain terms it was his problem to make his editor work with our coding standards, and that we didn't give a damn about his pretty little electric mode.

It took a lot of howling and gnashing of teeth for him to realize that he needed to fix it.

But this whining about whitespace just comes off as having never actually tried it.

No, it's a concept which had been ridiculed when COBOL used to insist you put things in specific columns. And as far as I'm aware, even COBOL stopped doing it, because we don't use punch cards any more.

For any of us who have taken compiler classes, a context free grammar specifically ignores whitespace. That's how compilers have worked for a very long time, if the grammar productions for your language involve counting whitespace ... well, my compilers prof would have failed me. Instead of having a visible thing to define a block, oh, well, just indent a few more chars.

You can't see what character whitespace actually is ... is that 8 spaces or a tab? Which means when your program suddenly won't compile, or is doing strange things, you have to spend extra effort to figure out WTF specific piece of invisible whitespace is the problem.

I've thought syntactically significant whitespace was dumb since I first saw it. Specifically because it creates many many places where it can go wrong, and no easy way to find them.

I've seen someone debug a python program, and even though things were in the same column in the editor, some were tabs and some were spaces, which had the very bizarre effect of making it semantically different than it looked. Because the block wasn't explicit, it was implicit based purely on indenting.

If you can break a program by changing the tab indent of your editor, the problem lies in the syntax of your language.

Making whitespace affect your syntax means "I'm too damned lazy to make my blocks explicit, so I'll rely on this cheap hack".

I've also worked with blind programmers who used code readers (yes, blind, and they were damned fine programmers). You know what whitespace being part of your syntax does to them? Renders them almost unusable.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Yeah, no, that's the cop-out answer. It's like saying you always have to trade something off: you can't make something better at X without making it worse at Y. Cheaper, faster, better, pick two.

Problem: I can make a language that is complete ass shit for anything, completely slow, expensive, terrible. It stands to reason a language may be, overall, better for a wider variety of tasks *and* faster *and* easier to implement, with different corner cases where another language might work better because it was designed for that or has some specific advantage that outweighs all of its disadvantages.

Some languages, for what they're good at, are not the best at that; and, in many cases, you can provide a different way of solving those problems without making a language that's particularly worse at anything than the original, inadequate language. That's called "better". Other languages are terrible for anything, or are just terrible as a model.

PHP for example has many overlapping functions, no standard implementation guidelines (core functions operate differently, take different styles of arguments, and are named differently; it's designed to solve similar problems in completely different ways; etc.), and so cannot be programmed in without either an at-hand reference or a huge volume of memory-by-rote. A language exactly identical to PHP except where every regex function took arguments in the same order and used a flag for lower-case searching would be, definitively, superior to PHP.

Python's most major problem is implementation: the reference implementation can't handle threads and, solving that, you have many extensions which are not thread-safe because they assume Python doesn't handle threads properly. Fixing that would be a beast: you'd have to minimize or eliminate the GIL, and then mandate that all broken extensions can fuck themselves.

C++'s most major problem is massive complexity in object orientation, which is not a problem in Objective-C--which itself is well-suited to solve all the same problems as C++. C++ mangling is also problematic; Objective-C less so. On the other hand, Objective-C is reference counted (superior to C++ manual memory management), while Python and Perl and PHP are implicitly reference counted (or, in some implementations, garbage collected), which is better, and Java and CIL are garbage collected, which can be slow and complex and harmful for performance. Both Python and Objective-C soundly thrash C++.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

C++ is strongly typed.

Re:If you wanted us to believe your Op-Ed...

[i kan reed]

Now you're talking about templating, which is a way of temporarily setting aside type restrictions for the sake of brevity in a specific and clearly denoted context. Allowing something isn't the same as it being your only option.

Re:If you wanted us to believe your Op-Ed...

[Anonymous Coward]

> because there was only a single whitespace character.

> So, where he wanted to see two character indents, and I wanted to see 8, and a co-worker wanted to see 4 ... once he'd edited it, in anything not using this electric mode didn't work.

No man, that's the ENTIRE point. You use 1 character, tab, that says indent here. Then everybody's editor renders tabs how they want. He sees the tab as the width of 2 spaces, you see the tab as the width of 8 spaces, and your co-worker sees the tab as the width of 4 spaces.

It works for EVERYBODY, and you only need 1 arrow key to move around, 1 backspace key to delete etc.

It's you fools that mandate spaces that mess it up for everybody else.

Re:If you wanted us to believe your Op-Ed...

[pigiron]

"Name a programming language that doesn't start to have major problems somewhere"

LISP 1.5

Re:If you wanted us to believe your Op-Ed...

[wol]

The two whitespaces in the beginning of this line are not like the other. Get your tabs out of my spaces.

Re:If you wanted us to believe your Op-Ed...

[Wootery]

Not really.

It's not a matter of just 'having inheritance'. Java and C++ have inheritance, but Carpenter#createTable is distinct from CSVReader#createTable (assuming no common ancestor class or interface), as the languages have static binding.

In, Ruby or Objective-C, which have late binding, you can 'pun' the createTable method name/signature: you can pass in a reference to any object which has a method by the name createTable (i.e. it 'looks like a duck and quacks like a duck'). Not so in Java or C++.

To simulate duck-typing in Java, you'd have to create an interface for each method, such as a HasCreateTableMethod interface, to be implemented by each class which has a createTable method (of the appropriate signature).

Re:If you wanted us to believe your Op-Ed...

[NonUniqueNickname]

You might think duck-typing got its name from "If it looks like a duck and it quacks like a duck". Common misconception. The name actually comes from "If it's declared like a Duck and compiled like a Duck" as pointed out by GP.

Re:If you wanted us to believe your Op-Ed...

[Wootery]

*looks like I should have used 'dynamic binding' rather than 'late binding'.

Which term I should have used instead of 'static binding', I'm not sure.

Re:If you wanted us to believe your Op-Ed...

[psmears]

It works for EVERYBODY,

... unless you want to look at diffs, in which case it breaks :-(

Re:If you wanted us to believe your Op-Ed...

[phantomfive]

"If it's declared like a Duck and compiled like a Duck"

Now that you mention it, I seem to remember that phrase from a John Donne poem....

Re:If you wanted us to believe your Op-Ed...

[Garfong]

Really? This has to be something specific to how he had emacs configured, because as a long time emacs user this certainly isn't the default.

On the other hand, emacs has the equally insane default behaviour of assuming tabstops are 8 characters, and if you have more than 8 characters of initial whitespace it will fold it down into the minimum number of characters by replacing spaces with tabs. Fortunately there's a setting which will turn this behaviour off and force emacs to always use spaces.

Re:If you wanted us to believe your Op-Ed...

[JoeMerchant]

Romansh will fall first. (I came here with a similar thought...)

Re:If you wanted us to believe your Op-Ed...

[TechyImmigrant]

>For example, duck-typing might be one of the worst ideas in the universe,

You don't have to duck type, but it's a convenience when you do. I tend not to.

But you do have to type "this." in front of everything and when you forget to, it might work but break in subtle ways. That's what I call a language flaw.

Re:If you wanted us to believe your Op-Ed...

[UnknownSoldier]

> Letting you define variables just by using them leaves you open to spelling mistakes and all the bugs it causes just so you can save a couple lines of code.

Indeed. Javascript is a piece of shit without this hack:

"use strict"; // Must declare all variables before using

Didn't we learn anything from BASIC in the 80's??

Re:If you wanted us to believe your Op-Ed...

[gstoddart]

Really? This has to be something specific to how he had emacs configured, because as a long time emacs user this certainly isn't the default.

He eventually found something to fix it.

But he simply wouldn't listen until I locked him out of CVS, sat him down, and showed him what the hell it was doing to the code in a hex dump.

Because all he gave a damn about was how he he saw it and how his editor worked, and didn't seem to care about the other developers, or that it was a codebase which was around since before he'd worked there and it wasn't up to him to determine our coding standards.

It was pretty stupid, but it was also 15+ years ago.

Re:If you wanted us to believe your Op-Ed...

[Stele]

You can easily do automatic memory management in C++ using reference counted smart pointers. This allows you to control when "memory management" occurs when necessary, as is common in the case of high-performance applications (games, imaging software, etc) where C++ excels. The ability to overload operators allows you to write vastly more readable (and efficient) code than with Objective C. And in Objective C all method dispatches are effectively virtual, where in C++ you can control when you pay the cost.

Objective C is definitely NOT well-suited to solve all the same problems as C++. It's fine that you don't need to write high-performance (or portable) applications, but sweeping generalizations like this just show your ignorance.

Disclaimer: I use C++, Objective C, and Python on a daily basis.

Re:If you wanted us to believe your Op-Ed...

[pak9rabid]

As the saying goes, there's two types of languages: the one everyone bitches about, and the one nobody uses.

Re:If you wanted us to believe your Op-Ed...

[tibit]

Template functionality doesn't set aside any type restrictions. After doing template type deductions and substitutions, the code must typecheck.

Re:If you wanted us to believe your Op-Ed...

[tibit]

Python can actually be parallelized for certain tasks. And it can be done without even touching the GIL, and without affecting the performance of "regular" code. And it can really perform very, very well.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

The ability to overload operators allows you to make your code readable to people who took a study in your code base, and not to people who can program in the language you write in. This is why most languages don't allow operator overloading. Operator overloading is commonly cited as one of the worst mistakes made while implementing the C++ programming language.

Re:If you wanted us to believe your Op-Ed...

[lgw]

Duck-typing in C: we do it with macros. And people mostly get that macros are dangerous to over-use, and limit the abuse of the concept (mostly).

Duck-typing: in C++ we do it with templates. The new-ish "concepts" approach that will eventually make it into the C++ standard is duck typing, but it's pretty clearly meant for well-reviewed library code to solve problems with no other good solution.

While duck-tying should be seen as an anti-pattern in OOP, there are still dark corners where you need it.

Re:If you wanted us to believe your Op-Ed...

[ceoyoyo]

So, here's my problem with whitespace being syntactically significant ... everybody likes to see code with different levels of indent. There isn't one "naturally readable" way which everybody agrees on. And then suddenly you have a language which says "we're all stuck with whatever the whiniest coder wants".

Use tabs. Set your tab stop to however many spaces you like. There you go. As a bonus, I can read your code with my preferred level of indent as well.

One of the few things that irritates me about Python is the PEP that suggests you should use spaces for whitespace.

Re:If you wanted us to believe your Op-Ed...

[lgw]

He has a good point though. Python's fundamental flaw is that it allows both tabs and spaces as indentation. If it had picked one, and made the other a syntax error, it could have been a great concept.

Curly braces are meaningless clutter. Indention level is what matters to the human eye, and all that the compiler needs. But you have to freaking normalize indention if you're going to take that path, or it ruins the whole deal.

Re:If you wanted us to believe your Op-Ed...

[lgw]

C++ is a bad language to do complex OOP shit in - use a managed language for that.

C++ is an awesome language to do low-level coding in thanks to automatic resource cleanup. All the speed of C, none of the error-prone "allocate at the top, free at the bottom, never ever return from the middle". When I write code that, back in the days before out-of-order execution would have been hand-tuned assembly, I choose C++ and let the optimizer work it's magic. And std::algorithms is quite handy for the sort of shit that make you dust off your algorithms book from college, all carefully tuned and well-tested.

Re:If you wanted us to believe your Op-Ed...

[Stele]

Which do you think is more readable?

Vec4 a = 1.0;
Vec4 b(2.0, 3.0, 4.0, 5.0);
Vec4 result = matrix * a * b / 10.0 + 0.5;

or

Vec4 *a = [[Vec4 alloc] initWithScalar:1.0];
Vec4 *b = [[Vec4 alloc] initWithX:2.0 Y:3.0 Z:4.0 W:5.0];
Vec4 *result = [matrix multiplyVec:[[a multiplyVec:[b multiplyScalar:1.0 / 10.0]] addScalar:0.5]]; ...
[result release];
[b release];
[a release];

Personally, I prefer the first one.

Re:If you wanted us to believe your Op-Ed...

[nmr_andrew]

A diff should be just fine. Parent talked about individual preferences for rendering (i.e. displaying) a tab. So as long as the editor maintains the tab character when saving the file, all should be well.

If, OTOH, some screwed up editor "renders" a tab by converting it to some number of spaces and then saves that...

Re:If you wanted us to believe your Op-Ed...

[ShanghaiBill]

If only Python and Ruby offered something similar to Perl's "use strict".

If only programmers always used code they wrote themselves.

I put "use strict" at the top of every perl file, javascript file, etc. I also compile C/C++ with "-Wall -Wextra". But that doesn't protect me from libraries and headers written by incompetent slobs.

Re:If you wanted us to believe your Op-Ed...

[psmears]

A diff should be just fine.

Unfortunately not. The thing is, tabs don't mean "<x> number of spaces" (with <x> configurable by the user) - they mean "Move rightwards to the next column that is a multiple of <x>". So, a tab in column 0 moves to column 8, but so does a tab in column one, two or seven.

This doesn't play well with diffs, because diffs tend to add some number of characters to the start of each line (typically one for "unified" diffs, or two for "context" diffs). So if a changed line in the file being context-diffed is indented with a single tab, and the user has set their terminal/editor/whatever they're viewing the diff in so that tabs are (say) three characters, then since the tab is s now preceded by two extra characters, the next multiple-of-three column is column 3, so the tab will result in the line being indented by only a single character (relative to the rest of the code in the diff).

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Option B if option A is C or C++, otherwise Option A if Option A is some mathematical language.

Re:If you wanted us to believe your Op-Ed...

[lorinc]

Exactly. Now let's speak about all that non strongly typed trendy crap too.

Re:If you wanted us to believe your Op-Ed...

[Jane Q. Public]

TFA is just plain wrong.

Not just about Ruby (which is still among the top 10 popular languages, and shows absolutely no sign of fading), but also about Twitter.

Twitter is a hodgepodge of languages, depending on just what "part" of Twitter you're talking about. For example, search is now done in Java. Some of the back-end is done in Scala. Etc.

Further, when one of the Twitter founders decided to switch part of the system from Rails (which is not the same as "Ruby") to Scala, he did so because he didn't understand how to properly do it in Rails.

Perf tests using the same kind of system in Scala, and Rails done properly, showed virtually identical results. But Twitter had already switched over.

It's no surprise at all that Java works faster in most ways than interpreted languages in general. Big deal. (That's why, by the way, jRuby exists... so you can pre-compile your apps if you want to and get Java speed.)

Ruby is an active, vibrant, and still growing community. It is still one of the most popular language.Not only did TFA get history wrong, it conflates Rails with Ruby, which is like saying Django and PHP are the same things, or that JavaScript and Node.js are the same things. It's all just BS.

Re:If you wanted us to believe your Op-Ed...

[Chris Mattern]

Nope. Sections/Paragraphs still start in column 8, columns 1-6 are ignored, 7 is used for comment characters and on some systems for condtionally compiled code, and 72 is the end of the line.

Not in any modern COBOL. A free form option has been required in the standard since COBOL 2002.

Re:If you wanted us to believe your Op-Ed...

[Kojiro Ganryu Sasaki]

Could be worse. Could be javascript.

Re:If you wanted us to believe your Op-Ed...

[RabidReindeer]

lol
"Syntax that every programmer uses to make their program readable is unreasonable as a semantically meaningful syntax"

Come on, python's got its problems, but forcing you to lay out your program in a naturally readable way to compile isn't one of them.

For example, duck-typing might be one of the worst ideas in the universe, because it's doing the exact opposite of the whitespace thing. It's decoupling easy-to-make mistakes with the output of compiling of your code.

But this whining about whitespace just comes off as having never actually tried it.

No, whining about whitespace is what you do when people use tab characters.

Because python treats tabs and blanks as identical characters, but they don't indent the same. Especially when fed to different formatters.

It's quite easy to create a python program that when printed or displayed looks perfectly fine but is a total train wreck when executed because of these literally invisible differences.

The only thing that makes it even remotely tolerable is that most editors have a Python mode that's sensitive to such misdemeanors.

Re:If you wanted us to believe your Op-Ed...

[mrwolf007]

Whitespace is great, I use it all the time. I use the IDE to automatically fix the indenting of my code. That doesn't mean the compiler should attach any specific meaning to how things are indented

This.

Forgot a closing bracket? Usually quick to find after auto-indenting.

And seriously. It doesnt even save space. Instead of a single bracket you will end up with a bunch \ts.

Re:If you wanted us to believe your Op-Ed...

[Matheus]

Long Live Lisp!

Parenthesis Parenthesis Everywhere... No white space required ;-)

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

Not requiring delcarations is one reason Python is a scripting language. It is indeed reasonable that a scripting language be allowed some shortcuts otherwise we may as well write those quick and dirty utilities in C++. Yes, this can cause bugs. But so what?

For indentation, I still get the impression that people think Python invented this or is the only language to do so. It's not like just one language designer got a stroke of insanity one day and decided upon this, there were quite a lot of people who thought such things were a reasonable idea.

Re:If you wanted us to believe your Op-Ed...

[gringer]

Come on, python's got its problems, but forcing you to lay out your program in a naturally readable way to compile isn't one of them.

I wouldn't mind python's indentation quirk so much if it didn't give me bugs when rewriting and transferring code, or if I could use braces (or something else) to indicate program flow. In short, changes in the control structure of the code can require indentation changes that cannot necessarily be determined by the editor:

A copy-paste of code from one place to another will only be easy to do in the case where the surrounding indentation is the same in both places.

When a particular bit of code needs to be surrounded in an IF statement (or a loop), it is necessary to manually re-do the indentation to indicate where the statement starts and stops.

If my editor decides that the default indentation is 2, while the code has a default indentation of 4, there's a chance I could inadvertently alter the program flow by re-indenting a line.

With other languages where structure is specified by non-whitespace, I can tell the editor to re-indent refactored code to accomodate new changes, taking a second or so instead of a minute or so (plus any time taken fixing bugs).

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

We did learn from BASIC in the 80s to never touch Visual Basic in the 90s.

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

Actually I never even heard of 'duck typing' until a few years ago, and I've been programming for decades and have been in academia. If there's one thing to be said about computer science is that we invent new words for old ideas on a constant basis.

Re:If you wanted us to believe your Op-Ed...

[John Bokma]

In short: the problem was not Emacs but an asshat developer; and it's not like those are rare.

Re:If you wanted us to believe your Op-Ed...

[serviscope_minor]

C++ with the right overloading is "some mathematical language".

And it's a whole lot more readable than doing the same in C or Objective C.

Re:If you wanted us to believe your Op-Ed...

[Stele]

Option A is provably easier to type, easier to understand, and vastly more efficient to execute, usually optimizing down to a handful of vector instructions.

You don't have a single good reason why Option B is "better" except that you don't like C++, which really isn't a good reason. Let me guess - you also hate Python because of the indentation.

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

Coding style standards in the workplace. Use them. This is not just a problem with Python. I've seen this problem in C. Yes, in C you don't have to worry that indentation screws you up, but if every member of the team is allowed to use their own indentation styles then you end up with an unreadable mess. I've been there and seen it. If the coding style says "4 characters and no tabs" then you follow that style even if you hate it. It's more important that the team work together well than to satisfy personal egos.

As for whitespace, Python didn't invent this and is one a unique oddball, it just happens to be the most popular language that does this.

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

The vast majority of coding styles I've seen in the wild will dicate that tabs never be used, ever, even in languages with delimiters, and even if some personal favorite IDE refuses to use spaces.

The major problem with using tabs and defining the indentation as however the user wants is that it does not translate to other programs and tools or users. For example, trying to print out something that looks nice in an IDE may end up with half the lines falling off the side of the paper.

Most of the people who argue against coding styles for projects or teams often fall back to the same idea that their personal editor is the best ever and can't everyone else just adapt to me instead of me adapting to you?

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

8 characters for tab is the worldwide computing standard that was set in place before IDEs even existed. All the early terminals used this, all the early editors that did not allow adjusting them, and all the printers that would take straight up ascii. It wasn't even until people starting using other widths that rules of no-tabs starting being added.

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

First language I used with whitespace for blocks (before Python was invented) had that problem also, and with the vi editor I used at the time I didn't know how to prevent it from automatically converting whitespace to tabs (yes, I never even typed in the tabs the editor just decided to be unhelpful). After some baffling debugging sessions I had to seek out a vi expert to figure out what settings to use.

Re:If you wanted us to believe your Op-Ed...

[Wycliffe]

So you state it's a cop-out answer and then go on to trash all the remaining languages. I was specifically thinking
of php when I wrote that answer. Php is a mix of randomness. So, yes, in theory you can write a better language.
The bar's actually rather low. You could even branch off an existing language if you were willing to break backwards
compatibility (add threads to python, standardize all the conventions in php, etc...) but that still doesn't disprove
my point that all the languages kindof suck especially in the web development and interpreted domains.
Java and C are ill-suited for just in time compiling and web development. Ruby, python, php, and perl are the best
we have. They all 4 suck but pick your poison or write a decent language and try to get everyone else to join you.
Oh, and I might go ahread and throw this in: on the frontend, javascript/html/css are even worse.

Re:If you wanted us to believe your Op-Ed...

[DamnOregonian]

modern COBOL

You just blew my mind, sir.

Re:If you wanted us to believe your Op-Ed...

[HornWumpus]

I've been meaning to write a jargon translator for decades. The problem is jargon is deliberately obtuse, harder then translating normal languages.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

No, I like Python. I just don't like operator overloading, since I can never predict wtf is going to happen when looking at code that uses it without doing an inordinate amount of research.

Operators are language-defined. Functions, even polymorphic functions, are programmer-defined. multiplyMatrix(a,b) or a->multiply(b) make more sense. The only thing wrong with Option B is the amount of verbosity, whereas it could be expressed exactly the same by breaking it out into multiple procedural lines or just inserting newlines for formatting.

It would also read better if you initialized a results matrix, loaded it, and used it as an accumulator. You don't optimize programs by compacting down the source code.

Re:If you wanted us to believe your Op-Ed...

[lgw]

Back in the day EMACS and VI was the holy war to have. These days I'm baffled there are people who still don't use an IDE for real projects.

Re:If you wanted us to believe your Op-Ed...

[DamnOregonian]

The JVM is a pretty badass VM, really.
Even Java itself as a language isn't too vile, it's really some of the things people do with it (JVM abuse allowed by the language) that can really turn the code into a festering pile of ebola-ridden excrement.
Normally, I'm pretty adamant about not blaming the language for the authors (I'm a 9th Circle Perl Acolyte), but finding a piece of Java code greater than the length of a single class that isn't an obvious gateway to the Dark Side is fucking impossible.

Back on topic, I'm super open-minded about languages, and I regularly use many of them in my line of work... But Python.. No. Fuck that garbage. I learned it, I became proficient in it, it still churns my stomach. The sheer level of skill it took to take all the good ideas in Python and ruin them with a few unforgivable paradigms impresses me every time.

Re:If you wanted us to believe your Op-Ed...

[fahrbot-bot]

You can't see what character whitespace actually is ... is that 8 spaces or a tab?

In addition, the X Windows server converts tab to 8 spaces on cut/copy and paste - which was really unhelpful when editing Makefiles back in the day. (yes, I'm old)

Re:If you wanted us to believe your Op-Ed...

[Stele]

We'll have to agree to disagree then.

Nobody writes math like Option B, unless they are forced to.

If someone wants to do this:

Vec4 operator+(const Vec4 &lhs, float scalar) const
{
        reformat_hard_drive();
        return 0.0;
}

Well, they're an asshole. That doesn't prevent the same asshole from doing this:

- (Vec4 *)multiplyScalar:(float)scalar
{
        reformat_hard_drive();
        return nil;
}

Programmers can be bad (or assholes) regardless of the language. I personally am thrilled that I've been allowed to shoot myself in the foot for 30+ years. It's made me a better programmer. I can't imagine being shacked to the incomprehensible mess that is Option B out of fear of what some bad egg might do to the api.

Re:If you wanted us to believe your Op-Ed...

[Aaden42]

That’s completely not what duck typing is.

interface Duck {
    String quack();
}

class Mallard implements Duck {
    String quack() {
        return “no”;
    }
}

class Mythbuster {
    String quack() {
        return “Damn you”;
    }
}

In a duck-typed language, any method which worked when passed an instance of Mallard would also work when passed an instance of Mythbuster. I’m not aware of any language which has an interface structure like the above where that would compile successfully without an explicit cast. And at that point, you’re punching the duck, so all bets are off.

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

I don't use one. The one's I have to use for a particular purpose are just plain awful. They don't get along well unless the entire project (build, regression, etc) are along for the ride as well. When used in a group context they often screw up badly when used with a source code control system (ie, someone wants to check in a new file to the project and they end up checking in their personal preferences as well because the IDE didn't separate things into separate files).

Yes I'm sure there are better IDEs but I've yet to see them. (don't mention visual studio, I'm not on windows and it was indeed the one that screwed up the source control and I had to regularly fix up its project files using Emacs)

I also want to see an IDE that is not a single window MDI interface. I want multiple source code windows at once sitting side by side, not merely multiple tabs, and interspersed with windows for other tools. If the tool does some great refactoring then that's good but it all gets undone and wastes my time when it's a lousy editor.

Re:If you wanted us to believe your Op-Ed...

[wrook]

OMG! A rant about Emacs vs VI with a sub-rant about indenting with spaces vs tabs!!! And you even throw in a story about locking someone out of CVS (although in my day we used RCS and we were damn happy to have it).

You know how to make an old man feel young again.

Re:If you wanted us to believe your Op-Ed...

[Aaden42]

Honestly, a lot of what you complained about actually turns out to be strengths of whitespace dependent syntax. You don’t have “every line’s a diff” problems since there’s no reformatting. Tools don’t touch the whitespace, so everyone needs to get along. Whitespace within a single file is consistent because if it’s not, it doesn’t work.

But that’s about it for the advantages... I’ll expand on the main disadvantage you mentioned because it deserves repeating:

When mixing tabs and spaces, most traditional editors will render two semantically different pieces of Python code with a pixel-identical display that programmers can stare at all day and not realize what’s going on. Short of enabling “control codes” mode or similar, there’s no visual indication of what’s wrong. I hurts my head to think about how many person-years of developer hair pulling and teeth gnashing must have been burned by this.

Granted most Python-specific editors tend to draw in left-gutters or color highlight the levels of hierarchy to represent the blocks which helps with this. But if I wanted to use an IDE-bound language, I’d pick up PowerBuilder or something... The greatest strength of plain-old-text based source code gets lost when you NEED to use a language-aware editor to be able to see functional differences in a piece of code.

I know one thing: If I had the misfortune to have to work in Python, our revision control server would have a pre-commit hook to reject any *.py which contained any ASCII 0x09 characters. No tabs at this establishment, buddy!

Re:If you wanted us to believe your Op-Ed...

[Dynedain]

And even worse, once you start nesting languages within each other (like HTML inside a JSP, JS, PHP, etc...) then any attempts at consistency for indentation is completely unachievable.

Re:If you wanted us to believe your Op-Ed...

[pmontra]

I feel your pain. I don't use any language that uses spaces to define the structure of code. They made me lose too much time.

Re:If you wanted us to believe your Op-Ed...

[Garfong]

My objection to this feature is that (as far as I know) emacs is the only editor that has it. So if you're in a mixed vi/emacs/whatever shop you quickly get a tab-space soup. Worse, unless you're careful you can easily start introducing whitespace-only changes to portions of a file you're working on, making merging changes between branches unnecessarily difficult.

Re:If you wanted us to believe your Op-Ed...

[CastrTroy]

Actually, in Basic/VisualBasic, the keword(s) for required you to define variables was Option Explicit. "Option Strict" required that you define the type of variables and that you cast between things properly when converting between different types. As far as I'm aware, "use strict" in Javascript does not stop you from reusing a variable to store different types, which is a whole other class of error I like to avoid by using a strongly typed language.

Re:If you wanted us to believe your Op-Ed...

[shutdown -p now]

It's no surprise at all that Java works faster in most ways than interpreted languages in general. Big deal. (That's why, by the way, jRuby exists... so you can pre-compile your apps if you want to and get Java speed.)

Just because it runs on JVM doesn't mean that it'll automatically get "Java speed" - it's not just about the JIT, but also about static typing; a dynamically typed language will always be slower, requiring very complicated runtime code analysis and type inference to even remotely approach the same perf as a statically typed one.

Re:If you wanted us to believe your Op-Ed...

[shutdown -p now]

Are you confusing strong typing and static typing? Python itself is an obvious example of a strongly (but dynamically) typed language with duck typing.

Statically typed languages can also have duck typing, by the way. Duck typing is basically just inferred structural typing, and that can totally be static. Objective Caml has just that, and its type system is far stricter than, say, C++ or Java.

Re:If you wanted us to believe your Op-Ed...

[shutdown -p now]

Operator overloading is not different from functions at all - operators are just functions with fancy names. And a programmer can just as well define a function that has "multiply" as a name, but does something else entirely, as he can define operator * that does something else entirely. These two are identical in all aspects. The only reason why operators were historically not redefinable is precedent, and it doesn't have any reasoning behind it.

Re:If you wanted us to believe your Op-Ed...

[Forever Wondering]

According to one of perl's developers, slashdot is powered by perl. See http://developers.slashdot.org... and look for this link: https://www.youtube.com/watch?...

Re:If you wanted us to believe your Op-Ed...

[Lotana]

Do you want it blown again?

It is object-oriented now!

COBOL

Re:If you wanted us to believe your Op-Ed...

[gewalker]

Some guy named Guido said

          __slots__ is a terrible hack with nasty, hard-to-fathom side
          effects that should only be used by programmers at grandmaster and
          wizard levels. Unfortunately it has gained an enormous undeserved
          popularity amongst the novices and apprentices, who should know
          better than to use this magic incantation casually.

Using __slots__ to enforce your programming style is very much not Pythonish. It break pickle an other things too.

Re:If you wanted us to believe your Op-Ed...

[Wootery]

Again, not really. You can't have proper duck-typing in C or C++. Macros and templates can provide a compile-time semblance of duck-typing, but it's not the really same. In Ruby, an array can hold instances of anything. That will never be the case with C++.

A C++ Vector looks somewhat like a List: they both have a push_back member-function, say, but you can't create a 'pointer to either a Vector or a List' type. It has to resolve at compile-time, even with concepts (though I admit I don't know a lot about them).

You could get proper duck typing using hash-maps mapping strings to function-pointers, but afaik that's about it.

Re:If you wanted us to believe your Op-Ed...

[UnknownSoldier]

> Javascript does not stop you from reusing a variable to store different types, which is a whole other class of error I like to avoid by using a strongly typed language.

Yes, I hate typeless languages, such as Lisp, Javascript, etc. for exactly this reason.

I would rather catch things at compile time instead of run-time.

Re:If you wanted us to believe your Op-Ed...

[Sique]

In the grey ages of the late 1980ies and early 1990ies, there was LPC, which had ducktyping to the extreme. If you called a method in an object which didn't implement that method, you just got 0 as result.

Re:If you wanted us to believe your Op-Ed...

[AmericaRunsOnDunkin]

Fortunately there's a setting which will turn this behaviour off and force emacs to always use spaces.

That's like saying "Fortunately, this handgun has a feature to always disable the safety and aim the barrel at your foot." Repeat after me: spaces are not indentation.

Re:If you wanted us to believe your Op-Ed...

[guacamole]

Agreeing on white space conventions are important regardless of what programming language you use. In Python world, it's always better not to use tabs and only use whitespace.

Re:If you wanted us to believe your Op-Ed...

[Bucky24]

You can't see what character whitespace actually is ... is that 8 spaces or a tab? Which means when your program suddenly won't compile, or is doing strange things, you have to spend extra effort to figure out WTF specific piece of invisible whitespace is the problem.

The very first time I used Python this exact thing happened to me. Transferred a largeish script from school computer (which used tabs) to my own IDE (which used spaces). It took me forever to even realize what was wrong. I still use Python from time to time, but I wouldn't use it for anything large that requires multiple people.

Re:If you wanted us to believe your Op-Ed...

[haruchai]

Hah! German will never die!!

The European Commission has just announced an agreement whereby English will be the official language of the EU rather than German which was the other possibility.
As part of the negotiations, Her Majesty's Government conceded that English spelling had some room for improvement and has accepted a five year phase-in plan that would be known as "Euro-English".
In the first year, "s" will replace the soft "c". Sertainly, this will make the sivil servants jump with joy. The hard "c" will be dropped in favour of the "k". This should klear up konfusion and keyboards kan have 1 less letter.
There will be growing publik enthusiasm in the sekond year, when the troublesome "ph" will be replaced with "f". This will make words like "fotograf" 20% shorter.
In the 3rd year, publik akseptanse of the new spelling kan be ekspekted to reach the stage where more komplikated changes are possible. Governments will enkorage the removal of double letters, which have always ben a deterent to akurate speling. Also, al wil agre that the horible mes of the silent "e"s in the language is disgraseful, and they should go away.
By the fourth year, peopl wil be reseptiv to steps such as replasing "th" with "z" and "w" with "v". During ze fifz year, ze unesesary "o" kan be dropd from vords kontaining "ou" and similar changes vud of kors be aplid to ozer kombinations of leters.
After zis fifz yer, ve vil hav a reli sensibl riten styl. Zer vil be no mor trubl or difikultis and evrivun vil find it ezi to understand ech ozer. Ze drem vil finali kum tru! And zen world!

Re:If you wanted us to believe your Op-Ed...

[arvindsg]

well in C/C++ you can always have a array of void * pointers, pointing to very different kind of objects.... Never say never especially for something so low level as c

Re:If you wanted us to believe your Op-Ed...

[gstoddart]

Thanks, I knew some of the kids wouldn't remember the old fashioned steam powered revision control systems we used to have, but I knew some of us old farts were still around. ;-)

Re:If you wanted us to believe your Op-Ed...

[jeremyp]

Curly braces are meaningless clutter.

No they are not, they denote the beginning and end of a compound statement in a nice explicit way that the the human eye can see and the compiler can parse.

Indention level is what matters to the human eye, and all that the compiler needs. But you have to freaking normalize indention if you're going to take that path, or it ruins the whole deal.

But normalisation of indentation is not a solved problem whereas pretty printing of blocks delimited by braces is.

In both of the editors I currently use frequently, typing an opening brace also puts in the closing brace nicely aligned and places the caret between them indented the correct amount ready for some code. How does that work in Python?

If I copy a piece of C code (or move it) from one place to another, my editor correctly indents it for the new context. How does that work in Python?

I do like Python, but its indent rule coupled with the rule that stops you from breaking long lines anywhere you like is a bit of a pain. However, it's pain I'm prepared to live with because of the other nice features.

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

I use Python. I have to also work with Python code written by other people.

Autopep8 is your friend and really it shouldn't have to be.

This. How freaking hard is it to set tabs to spaces and make every number referencing tabs or indents in your editor 4?

Re:If you wanted us to believe your Op-Ed...

[jeremyp]

They missed a trick with Objective-C. In Smalltalk the binary operators are messages just like anything else, but in Objective-C you can't use the binary operators as messages.

It actually wouldn't have been that hard for the language to define the binary operators as syntactic sugar for messages e.g. you could have * being syntactic sugar for multiplyBy: when it appears in the context of a message selector, so your above example would then be:

Vec4 *result = [matrix * [[a * [b / @(10.0)]] + @(0.5)]];

@(...) is the modern Objective-C syntax for boxing a literal as an NSNumber.

It's still not as readable as your first example, but it's an improvement on the normal syntax.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Point is you can't say one language isn't particularly bad. A language can fill a niche and be terrible--and even be good at that niche. A language can also be impossible to implement for a niche without implementing the language poorly, thus a language will be general-purpose and broad-use; whereas a niche language may be restricted to a niche by necessity, or may be able to sacrifice a small subset of broad uses to be general-purpose but suited for a specific niche that other languages fail at.

The case where a language is best suited broad-purpose, whether it's targeting a specific niche or not, is the best case: a well-designed broad-purpose language is in general easier to code quickly and correctly, and thus any special purpose it's designed around will benefit from its general usefulness. A language exposing functions and constructs which make a particular use possible can still provide a terrible, complex interface for the programmer, giving you easier control over that purpose than some other language but still being a pain in the ass to use at all.

And of course the bar's low. I can stomp all over every other language except Python, and that's only because the things wrong with Python sound silly and are evasive--even if they're important. That and because I don't understand Python well enough to find some critical flaw besides a lack of trigraphs; but that's a holy war topic. I personally like the "a = ((x > 2) ? b : c)" construct, over "if ( x > 2 ) { a = b; } else { a = c; }" construct, but most people see it as unreadable noise.

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

So, here's my problem with whitespace being syntactically significant ... everybody likes to see code with different levels of indent...

You're in a work environment. It doesn't matter what I like to look at, it matters what the standard is. If you throw fits over 4 vs 8 vs tabs vs 2, that's just primadonna crap. Its not as if you can't grok separation if its not at least 4 spaces.

Not only that, there is PEP 8. 4 spaces, no tabs is the preferred way. Code to the standard. Period. How complex is this?

When I got a job in a C++ house, I was shown their internal style guide - 2 spaces, no tabs. There was a reason they made it this way at some point in the distant past, and I never bothered asking why. Now I could have thrown some autistic OCD fit about my absolute need to have 6 spaces and tabs and pretty whatevers, but instead I made sure I set a "work style" template in all of the editors I use and drove on.

So, tl/dr: That dude was a whiny ass, style guides are there for a reason, follow the style guide, follow the style guide, follow the style guide.

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

But still, this still boils down to operator error, not an inherent problem with the language itself.

This is why we have style sheets, and why we follow them. If people want to march to their own drum they can do it on their own code bases at home.

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

They did, with PEP 8. If its really this much of a problem and you're in a python shop, one of your first interview question sets should be this:

"Do you know what PEP 8 is?"

"Do you know what a style sheet is?"

"What editor do you use?"

"Do you understand the process to make your editor comply with the style sheet?"

After this, the first code review for the new employee should be checked immediately for non-compliance and action taken accordingly.

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

If you want a nice python IDE, pycharm is pretty amazing.

As for my realm, I've been enjoying QT Creator quite a bit. It just uses a simple text file of directories and locations to index in, making it pretty easy to script a grabber to generate that file with just the things you want to see. The gdb integration is great, and the IDE itself is pretty clean and uncluttered. Might be worth checking out (if you do C/C++).

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

The vast majority of coding styles I've seen in the wild will dicate that tabs never be used, ever, even in languages with delimiters, and even if some personal favorite IDE refuses to use spaces.

The major problem with using tabs and defining the indentation as however the user wants is that it does not translate to other programs and tools or users. For example, trying to print out something that looks nice in an IDE may end up with half the lines falling off the side of the paper.

Most of the people who argue against coding styles for projects or teams often fall back to the same idea that their personal editor is the best ever and can't everyone else just adapt to me instead of me adapting to you?

THIS. I work in a C++ house and the rule is 2 spaces, period. Why does it even matter? Evidently it did at some point, so the rule was made.

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

Good, then you should be able to overload 'if' and '{' in C, right?

Re:If you wanted us to believe your Op-Ed...

[TheBilgeRat]

C++ is a bad language to do complex OOP shit in - use a managed language for that.

C++ is an awesome language to do low-level coding in thanks to automatic resource cleanup. All the speed of C, none of the error-prone "allocate at the top, free at the bottom, never ever return from the middle". When I write code that, back in the days before out-of-order execution would have been hand-tuned assembly, I choose C++ and let the optimizer work it's magic. And std::algorithms is quite handy for the sort of shit that make you dust off your algorithms book from college, all carefully tuned and well-tested.

This argument is made a lot, but ignores the reality of the state of C++ today. I think the biggest hurdle for C++ is the fact that fewer and fewer colleges are taking the time to teach languages like C++. Most are Java houses, because its easier to spit out code drones if they don't ever have to worry about memory management (which isn't quite true, but whatever). So the pool of qualified C++ coders gets smaller and less projects in the app/widget space get started in C++.

Having been exactly one of those java drones and getting a job in a C++ house and having been thrown in the deep end to drown, I can say that complex OOP in C++ is just as easy as it is in java, with the added bonus of the ability to go "low-level" easily if need be.

Re:If you wanted us to believe your Op-Ed...

[CastrTroy]

Yeah, I think that VB.Net offers a good compromise though. In the default configuration you can do things like assign an integer to a string variable without explicitly casting it, because the conversion of obvious and does not lose any data, but you can't do crazy things like assign a List to a string, because it doesn't know how to do a conversion. It's nice because print functions and other things that would usually take a string can be called without casting if you just want to print out an integer, but you still get most of the advantages of strong typing.

Re:If you wanted us to believe your Op-Ed...

[lsatenstein]

lol
"Syntax that every programmer uses to make their program readable is unreasonable as a semantically meaningful syntax"

Come on, python's got its problems, but forcing you to lay out your program in a naturally readable way to compile isn't one of them.

For example, duck-typing might be one of the worst ideas in the universe, because it's doing the exact opposite of the whitespace thing. It's decoupling easy-to-make mistakes with the output of compiling of your code.

But this whining about whitespace just comes off as having never actually tried it.

My favorite language is Iverson's APL. It will never die. It is used by financial statistical and string oriented language used by actuaries moreso than other professions.

Re:If you wanted us to believe your Op-Ed...

[beastofburdon]

That really hurt my brain reading that.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

More to the point, didn't we learn anything from FORTRAN in the 60s?

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

Doesn't Haskell duck-type when it can, in the sense of deducing types? It's strongly and statically typed.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

Template functionality depends on type deduction, and what operations are available for the particular type in question. That's at least awful close to duck typing.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

emacs is an IDE. vim can be made to function pretty much as one. Personally, I like multiple xterms instead of an IDE.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

The old Unix make language had tabs and spaces mean different things. I have no idea who decided that, what he was smoking, or where I can get some.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

Actually, operator overloading is a really neat idea that can be fantastically misused. If you overload properly, you've made the program more readable, If you overload improperly, you've made it incomprehensible. C++ has other features like that. It's a great language but requires a good deal of knowledge and skill to use well.

Re:If you wanted us to believe your Op-Ed...

[cwsumner]

Do you want it blown again?

It is object-oriented now!

COBOL

Wow! Way to put the newbies in their place.

And yes, Cobol is still used. Really...

Note: The health of a programming language is based not on the number of apps sold, that used it, but on the number of programmers who bought compilers / development systems. Quite a different thing, sometimes.

Re:If you wanted us to believe your Op-Ed...

[cwsumner]

I would like to see a list of languages that are not known-shitholes. ...

Clarion. Used all over the world by professionals who need to get things done reliably as well as quickly.

But it's Windows only, at least for now.

And it's not cheap...

Re:If you wanted us to believe your Op-Ed...

[Darinbob]

I actually interviewed with him, and he seemed sane. I agree that really it could have been simplified; all rules are left justified with no leading whitespace and all command are intended with one or more whitespace. Works identically, is backwards compatible, and remains readable.

Re:If you wanted us to believe your Op-Ed...

[ConceptJunkie]

When the codebase you are working on has been garbled by 12 years of random tab indent settings and other similar butchery from who-know-how-many programmers (and in some cases totally incompetent consultants) and other nonsense. The indentation in the code (not Python, but C++) is horribly mangled in many places and sometimes all but unreadable, but no one will allow it to be fixed. The other developers I've talked to insist either that it's not a problem, or that we should keep tabs, and if everyone who ever touches the code would simply do everything perfectly and never make a mistake, it wouldn't be a problem.

This is further complicated by a plethora of styles that includes such nonsense as right justifying code, and people who almost literally seem to think the only whitespace you are allowed to use is tabs. Yes, in other words, there are large swaths of code where there are tabs between many of the tokens rather than spaces. Apparently, tabs are not just for indenting for some people.

You would think this would be simple, but in practice, it's not. You would think coding standards and style guides would help, but not if they're ignored, or they were mandated after millions of lines of code was already written in every conceivable style and a few that would leave you gibbering like a Lovecraft character. Then there are the files that had tabs randomly converted to spaces, with indent levels ranging anywhere from 2 to 8, and often at a different indent level than the code with spaces only uses.

As someone who takes code formatting seriously, it drives me crazy, but no one else seems to have a problem with it. I've pleaded to allow the code to be formatted with uncrustify or something similar, but no one likes the idea because apparently I'm also the only person who does any kind of automation and for whom adding an additional step wouldn't affect my work routine. Note: the complaint is not that uncrustify, etc., might break the code, which is a real concern, but that it would be too much trouble to use. This is very incongruous with the fact that these are very good developers who know their stuff and write good code (as opposed to a lot of the people who apparently worked on it in the past.) I'm just surprised how set in their ways people get.

Regarding Python, I thought the whitespace thing was a lousy idea until I tried it. I love it now, but I'm someone who is very OCD about code formatting, so I would format my code in any language the same way that Python requires. It took a little getting used to, but I think it helps make Python code more streamlined to read than C++.

Re:If you wanted us to believe your Op-Ed...

[ConceptJunkie]

It's not like just one language designer got a stroke of insanity one day and decided upon this

Yes, that would be whoever defined the format for make...

Like I said above, I thought it was a pretty bad idea until I started using Python. Now I really dig it. On the other hand, I'm very strict about how I format code, so I write anything as if the whitespace matters.

Re:If you wanted us to believe your Op-Ed...

[Pseudonym]

Nope, that's not duck typing. That's type inference.

Haskell does support a principled form of duck typing, in the sense that you can make any type a Duck:

class Duck d where
    quack :: d -> IO ()

instance Duck BelittlingBird where
    quack _ = putStrLn "I don't think so"

Unlike Simula-style OO languages, Haskell has a strict distinction between a type and a class (the same as the distinction between a set and a class in Goedel-von Neumann-Bernays set theory). The equivalent in Java would be if you could declare a class or even a built-in type to "inherit" from a superclass without having access to the source code for that class.

Re:If you wanted us to believe your Op-Ed...

[Pseudonym]

For any of us who have taken compiler classes, a context free grammar specifically ignores whitespace.

People have written whitespace-aware grammars because that is what the language specification requires. One of the more famous historical examples was Jon Levine's Fortran 77 parser written entirely in Yacc with no lexical analyser. Parser combinators have made the technique popular again in some circles.

Most compiler writers aren't happy with the current state of parsing tools, by the way. Almost all of the big production compilers use hand-written parsers because programming languages are complicated when you take into account language variants and extensions, and because error reporting from automatic parser generators is terrible.

From a theoretical point of view, once upon a time, people wrote recursive descent parsers to parse the block structure of a language and then switched to an operator precedence parser to do expressions. The distinction between lexical analysis and syntax analysis seems similarly artificial.

Re:If you wanted us to believe your Op-Ed...

[Pseudonym]

These days I'm baffled there are people who still don't use an IDE for real projects.

IDEs work if your language and toolchain are well-integrated with the IDE. If they aren't, you spend a significant proportion of your time fighting the IDE. Also, if you're the sort of software engineer who spends some of their time doing sysadmin-related stuff, moving between two types of text editor incurs a significant context switch.

Of course, what I really want is vim and/or Emacs embedded in the IDE. There are some projects for doing this with varying levels of success.

Re:If you wanted us to believe your Op-Ed...

[Pseudonym]

It was the 70s. We didn't know as much about syntax as we do now.

Re:If you wanted us to believe your Op-Ed...

[strikethree]

Heheheh. Cute. Where did you find that at Bloodguard?

Re:If you wanted us to believe your Op-Ed...

[bluefoxlucid]

The most important benchmark is this: if a thing gives advantage X with a great deal of knowledge, skill, care, or some combination; and advantage X is not significantly better than some other implementation which solves the same problem but isn't as easily screwed up and, thus, produces better results with less knowledge, skill, and care; then the other implementation is better, and will produce as-good or better results more often.

Basically, advantage of consistent suitability versus advantage of a powerful tool requiring an exquisitely trained and rigorously practiced specialist. Of course, we take this to an unfortunate extreme in our society, making things wholly unsuitable for all but the most narrow and inadequate cases, but usable by idiots with zero knowledge or understanding. The goal is for your simpler tool to produce the same value with almost exactly the same efficiency (if not better) when similarly skilled as to be able to use the more complex tool, not to cut away features and call it "better" because there are fewer things to do and, thus, fewer ways to fuck up.

Re:If you wanted us to believe your Op-Ed...

[haruchai]

It's an oldie that was sent around in the late '90s. I think I first saw it in '98

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

Those are good design goals for a language, but they aren't the only ones. In particular, they weren't Stroustrup's goals for C++.

I didn't say operator overloading required a great deal of knowledge and skill, since it's fairly easy to know when not to use it. Also, I don't think other languages solve the same problem as elegantly. Consider complex numbers. Either they're built into the language, they can be made with operator overloading, or they're clumsy to use. Consider memory allocation. There are advantages to being able to use default memory allocation or some sort of user-designated memory allocation without having to specify explicitly every time memory is allocated or deallocated, particularly when using more abstract concept (C++ make_unique and make_shared, for example). C++ does that with overloads for new and delete.

Re:If you wanted us to believe your Op-Ed...

[david_thornley]

Interesting. However, you're using the misleading word "whitespace" here, which isn't really applicable when tabs and spaces mean different things.

Re:If you wanted us to believe your Op-Ed...

[Wootery]

No, that's still nothing like duck-typing. There's no dynamic dispatch.

Re:If you wanted us to believe your Op-Ed...

[Dutch Gun]

I've heard about this "overloaded operator" problem a number of times. I've been using C++ for almost 20 years now. Do you know how many times I've been bitten by crazy nonsense happening under the cover of an operator overload? Zero. Nada. It's never ever happened to me in practice, and I've worked with a variety of codebase quality. I'm a sample of one, of course, and therefore can't extrapolate out to all use cases, but still a sample of one who's been using the language a really long time now. If it was really such a problem, you'd think I'd have actually experienced said problem at least once at some point.

Maybe it's a problem if people do really horrible things in operator overloads. But come on - just don't do that! Why the hell would someone write a database transaction as an operator overload, for instance? That would be insanely bad practice. It's not something you can do accidentally, like forgetting to free some memory. It's a deliberate, very bad design decision, and you're blaming the language? It's like blaming a kitchen knife for slicing your fingers if you don't treat it with the proper care and respect.

C++ has plenty of things wrong with it as a language, and it's most definitely not suited to every programmer or every programming task. Operator overloading has never even made the list, as far as I'm concerned.

Who cares?

[PhilHibbs]

Some people like to make a big deal over languages dying, particularly if the language is one that they never really liked. I say, why make a fuss? Sure, some languages will decrease in popularity, but they're still there to use if you want, and there will always be a die-hard community of fans that keep it alive. Why hold a big whoop-de-doo circus to celebrate the ebb and flow of language popularity?

Re:Who cares?

[oodaloop]

Besides which, is there any language that is completely dead? Honest question.

Re:Who cares?

[Lovepump]

Modula-2

Re:Who cares?

[Anonymous Coward]

How about B, the precursor to C?

Re:Who cares?

[__aaclcg7560]

I've always wondered about Modula-2. Jerry Pournelle mentioned Modula-2 from time to time in his monthly Byte Magazine column in the early 1980's. Back then Pascal was a popular programming and I was stuck with BASIC on the Commodore 64. Whenever I looked at Modula-2 in recent years, I think "meh" and never dived in.

Re:Who cares?

[jandrese]

Plankalkül. An 80 year old language for a machine that was never successful.

Re:Who cares?

[micahraleigh]

Some of us monitor adoption as if staying current affected our employability.

Re:Who cares?

[lgw]

Besides which, is there any language that is completely dead? Honest question.

PL/S is, as far as I know and dearly hope, quite dead. If not someone needs to drive a stake through its heart, saw off its head, and bury it at a crossroads. If you think the 5 languages listed are bad, you haven't tried PL/S.

Re:Who cares?

[rokstar]

If there is a just and loving god, MUMPS should be dead by now. http://en.wikipedia.org/wiki/M...

Re:Who cares?

[T.E.D.]

Besides which, is there any language that is completely dead? Honest question.

CMS-2. It was the US Navy's first attempt at a "standardized" language, before they joined with the other armed forces and sponsored the creation of Ada. I know back around the turn of the millennium there were still some ships around using the "AN/AYK" (affectionately pronounced "Anne Yuck") 16-bit CPUs that ran it, and occasionally they'd want a tweak. But I highly suspect that's all been upgraded by now. The last person I know of who actually knew the language retired in 1998.

Re:Who cares?

[budgenator]

IBM RPG II and of course RPG I. It reminded me a lot of PERL format programing for some reason. I don't think you can get any deader than RPG II.

Re:Who cares?

[david_thornley]

I did a quick google for free Modula-2 compilers. It would appear to be alive. I've never had any desire to learn it, but if I wanted to start using it I could.

Re:Who cares?

[cwsumner]

Besides which, is there any language that is completely dead? Honest question.

Fourth?

Arguments passed explicitly, on a macro-stack, in reverse-Polish notation like an HP calculator.

Actually, I think the astronomers still use it for telescope drives.

Adoption by large organizations limits extinction

[Terry Pearson]

Once a language is adopted by a large organization, it is almost impossible for it to go extinct. Just the way that larger companies tend to work, means that the language will exist in some form for decades. If I were to predict a language to go extinct, I would say that it has to be one that has not been widely adopted already, has not made its way to mainstream organizations, and basically reproduces what is already done by another, more popular, language.

Ruby?

[Chris Mattern]

His main complaints about Ruby seem to be that C programmers find it hard to use (because C is at the forefront of innovative computer languages, you know), and that Twitter has stopped using it (oh noes!).

Re:Ruby?

[Electrawn]

And completely forgetting that both Puppet and Chef are both ruby based and not going anywhere. Rails may die.

Re:Ruby?

[Chris Mattern]

Exactly. In fact, Puppet is why I'm learning Ruby, as we are doing Puppet-based deployment these days.

Re:Ruby?

[ron_ivi]

Rails may die.

Thank you for making that distinction.

I liked using Ruby before ever seeing Rails, and have a moderate dislike for Rails.

I wish people would stop judging Ruby for one IMHO awkward and bloated framework.

Re:Ruby?

[wiredlogic]

I don't get that either. I was under the impression that a lot of Perl refugees switched over to Ruby.

Re:Ruby?

[Luthair]

Like C or not the syntax has been the standard for most languages, and its the syntax taught to a generation of programmers. Language designers take note, deviating from the syntax does not increase clarity nor does it save time so don't unless using a different paradigm.

Having the poster child abandon the language and need to re-write its entire system was a pretty strong statement to the rest of the users. In reality though I think a lot of ruby users just moved onto node.

Re:Ruby?

[Wdomburg]

JRuby and Rubinius have been using JIT for years.

Re:Ruby?

[jp10558]

Did you see the puppetconf presentation? They're pretty much going away from Ruby and going to APIs. Yea, you can still use Ruby, but c++ native client is what's going to be shipping next year... The server is going Java.

Re:Ruby?

[thetoadwarrior]

Ruby never fully took off like PHP or Perl. It gained popularity but then was quickly replaced by node / JS. Now about the only people I see advertising Ruby jobs are Yammer in London or the odd company that was horribly stuck in the past and decided to move forward to 2007.

Re:Ruby?

[mr_mischief]

I haven't watched the presentation but I'd heard about some of this. I have a few questions to clear up.

Will the Java server handle all our Ruby custom facts?
Will it handle our ERB templates?
Will it actually be Java or like PuppetDB will it actually be Clojure?
Will it have Jetty in front of it like PuppetDB and take 30 seconds or more to restart and listen on a port vs. restarting Apache in 2 seconds with the puppetmaster Rack application behind it?

Re:Ruby?

[T.E.D.]

I share your scoffing at the supposed importance of "looking like C" in a language. Not looking like C is generally a selling point in my book. However, the prevalence of brace blocks and 0-indexed arrays in other languages makes his point for him.

Re:Ruby?

[Aighearach]

Which is hilarious, my favorite thing about Ruby is that I can write my Ruby in plain C. There is a simple C API for the whole language, including interacting with the pure Ruby parts.

Unlike Perl, which requires a whole extra glue language.

Re:Ruby?

[Aighearach]

Rails may die.

Rails may die, but ActiveRecord will live on. It used to suck, but since 3.x it is the best ORM in any language. I'd love a C port.

Re:Ruby?

[spiralx]

Yes, this is the case, and it's very obvious if you look at popular NPM modules - there are lots of ports or variants of Ruby libraries. Plus there's CoffeeScript, which I'm told has a Ruby-esque feel, but can be trivially used interchangeably with JS.

Perl and VBA will live for a long while yet

[Beetle B.]

I work in an engineering firm. There's so much legacy Perl out there that there'll be a need for it for at least another decade.

As for VB, it'll remain as long as Microsoft Office is used in companies. It's way too handy and there's no alternative.

Re:Perl and VBA will live for a long while yet

[SJHillman]

"It's way too handy and there's no alternative."

My understanding is that, with very few (and increasingly fewer with each new version) exceptions, anything you can do in VB.NET, you can do in C#. Does Office only understand VB.NET?

Re:Perl and VBA will live for a long while yet

[geminidomino]

VBA != VB.Net

Re:Perl and VBA will live for a long while yet

[neoritter]

No, the primary Office applications that like to use VB.net can also use C# instead.

Re:Perl and VBA will live for a long while yet

[Anonymous Coward]

Office doesn't even understand VB.NET. Just VBA. Which has the same syntax as VB6, but lacks some features.

Re:Perl and VBA will live for a long while yet

[jeffmeden]

"It's way too handy and there's no alternative."

My understanding is that, with very few (and increasingly fewer with each new version) exceptions, anything you can do in VB.NET, you can do in C#. Does Office only understand VB.NET?

The Office functions might be known to C#, but Office admins (those creating/maintaining the scripts) definitely do not know C#.

Re:Perl and VBA will live for a long while yet

[jmac_the_man]

The VB that Office does is VBA, which is essentially VB6. The language that they're predicting will die out is VB.Net, which has a very different syntax.

Re:Perl and VBA will live for a long while yet

[psmears]

The Office functions might be known to C#, but Office admins (those creating/maintaining the scripts) definitely do not know C#

This - and the fact that C# is not built into Word/Excel/Powerpoint/Outlook, whereas VBA is.

Re:Perl and VBA will live for a long while yet

[TemporalBeing]

As for VB, it'll remain as long as Microsoft Office is used in companies. It's way too handy and there's no alternative.

VB != VBA != VB.net

They only said VB.net will go away (at least in the summary). Most VB programmers stuck with VB instead of moving to VB.net because it was such a substantial change to go to VB.net.

VBA will never go anyway so long as MSO doesn't add support for other stuff, but even then there's so much written in VBA that it will only die when MSO dies.

Re:Perl and VBA will live for a long while yet

[dwpro]

you mean, VBA VB.net.

Re:Perl and VBA will live for a long while yet

[dwpro]

damn, slashcode ate my chars, VBA <> VB.net

Re:Perl and VBA will live for a long while yet

[thetoadwarrior]

Yeah, most popularity contests seem to be based on what start-ups use like the only programming done is making novelty shit for middle class westerners and their 1st world problems. But once you get into something a bit more intelligent and long lasting then you find these people stick with what works because their company can't afford to spend all its time converting to the new shiny thing nor are they some short lived dumb start-up that can ignore the long-term implications of their decisions.

Re:Perl and VBA will live for a long while yet

[squiggleslash]

Despite the superiority of C#, my experience is that VB.net is by far the dominant language in (enterprise) .net shops, and (enterprise) .net is about as popular as Enterprise Java in practice.

Two major reasons:

- People are very familiar with VB and can program in VB.net easily.
- Lots of critical custom business apps were written in VB in the 1990s. As the software has been migrated to the web, VB.net allowed large amounts of source code containing critical business logic to be literally cut and pasted over.

The summary is completely absurd. VB.net will be around for another 25 years, at least.

Re:Perl and VBA will live for a long while yet

[TemporalBeing]

Despite the superiority of C#, my experience is that VB.net is by far the dominant language in (enterprise) .net shops, and (enterprise) .net is about as popular as Enterprise Java in practice.

Two major reasons:

- People are very familiar with VB and can program in VB.net easily. - Lots of critical custom business apps were written in VB in the 1990s. As the software has been migrated to the web, VB.net allowed large amounts of source code containing critical business logic to be literally cut and pasted over.

The summary is completely absurd. VB.net will be around for another 25 years, at least.

I've done VB (VB5, VB6) programming and VB.net programming. There were so many subtle differences between VB6 and VB.net that a vast majority of the VB programmers continued with VB6. Code didn't get migrated in the sense that it did for VB5 to VB6 where all you had to do was reload and recompile - it got completely re-written going from VB5/6 to VB.net. It was a long standing complaint for 5 or so years after VB.net was released.

Now so much has been re-written, that it's probably not as a big a deal and VB.net may have taken over as a result. But I can certainly guarantee you that VB6 stuff is still out there and won't be rewritten for VB.net; probably will still be when VB.net goes away.

Ummm, no.

[Anonymous Coward]

These languages may not be the "cool" languages at the moment, but to say they are "dead tech" (or even on their way) is classic hyperbole, and /.'s owner dice should be ashamed for soliciting ad views with this nonsense.

Re:Ummm, no.

[Austerity Empowers]

It's just the usual groupthink that cliques like to propagate. There are tens of thousands of lines of new perl and ruby (not to mention python, tcl and javascript) being created in the engineering world every day. There's been no shift away from them, and in many cases they are integral parts of CAD tools. I have yet to work in a place where Perl wasn't an institution. I keep hearing that C is dying, obsolete, ancient, yet I would advise any child who wants to pursue CS to learn it first, and learn it well, that if you knew only one language that it should be C. It isn't going to "die" in my lifetime, and probably not my children's. It's an excellent language where pragmatism is advantageous (read everywhere, but primarily OS/Systems/embedded) .

The key thing here is that these languages are tools we use, but we do not define ourselves as "C Programmers" or "Perl Programmers". My title has something like "design verification" or "design engineer" or something in it, my resume lists C, perl, python and (gasp!) various assembly languages. I'm not hired because of those skills, but those skills definitely increase my chances of being looked at. I think if you say looking for a "Perl Programming" job, you'll strike out. But there's tons of jobs out there for which perl programming is a hugely valuable skill.

But if you want to chase trends, and you want to chase job req's... well I have no advice for you except that you're doing something fundamentally wrong anyway.

Re:Ummm, no.

[NoImNotNineVolt]

I still list Tcl/Tk on my resume. At a recent interview, this was pointed out, and both I and the interviewer had a good laugh about it.

Not MY language!

[Bovius]

I suppose this is where I'm supposed to be indignant because the language I use got listed. But, I suppose it's fair. Ruby has always been one of the trendier languages, regardless of its utility.

Really struggling to avoid defending it, though.

Re:Not MY language!

[mooingyak]

I suppose this is where I'm supposed to be indignant because the language I use got listed. But, I suppose it's fair. Ruby has always been one of the trendier languages, regardless of its utility.

Really struggling to avoid defending it, though.

Feel free. I mean, perl is on that list, and as much as I'd love to see every line of perl vanish from the face of the earth, there's no way in hell I'd say the language is on the verge of extinction.

If you're thinking that ruby isn't headed for the trash heap, you'd be right too. And you don't have to like (or dislike) it to see that.

Perl?

[just_another_sean]

Perl 6 might be languishing in academia but in the meantime Perl 5 is chugging along nicely with bug fixes released regularly and CPAN content growing week over week. Not to mention Debian and BSD's heavy use of Perl in the base system.

They can have my Pathologically Eclectic Rubbish Lister when they pry it from my cold, dead hands!

Re:Perl?

[Trepidity]

It's used in a ton of places in the bowels of big companies as well.

Re: Perl?

[Anonymous Coward]

What a coincidence, the exact same place that shit accumulates.

Re: Perl?

[Anonymous Coward]

Remember, something has to sort through all that shit and make sense of it!

Re:Perl?

[jandrese]

Perl 6 is one of the most blatant examples of the Second System Effect in modern times. Perl 5 wasn't exactly small and svelte, but it worked. Perl 6's design had the ambition of fixing all of Perl's problems in one go with a gigantic rewrite and complete rebuilding of the underlying architecture. I'm not saying Perl 5 is the last version of Perl that will ever be widely used, but I am saying that Perl 6 in anything resembling its current incarnation is not likely to be the replacement, especially as more and more of the really useful fixes get backported into Perl 5.

Re:Perl?

[LQ]

When your legacy shell scripts start to get unwieldy, it's time to rewrite them in perl. It fills a similar niche but is more amenable to structured programming.

Re: Perl?

[Kiaser Zohsay]

Shit is what's left after all the sorting is done and the useful parts have been sent where they're needed.

Re:Perl?

[Kiaser Zohsay]

Unmaintained windows packages? Are you using ActiveState? Try Strawberry Perl instead. It comes bundled with a Mingw32 build of GCC, which might seem like overkill, but it has been able to build from source any package I have thrown at it. Just use cpan.bat instead of ppm.bat.

Re:Perl?

[spiralx]

Oddly enough quite a bit of the design PowerShell was inspired by Perl.

Re: Perl?

[gilby]

The problem with "a traditional interpreter in C" is that Perl6 code can change how it's parsed.
Which means that the easiest language to date with which to write a parser for Perl6, is Perl6.

Which is why the Rakudo implementation is written mostly in Perl6 and NQP (Not Quite Perl).

Speaking of C level stuff there is a new VM designed specifically for the Rakudo implementation of Perl6, MoarVM.
You can also run Rakudo on the JVM.

Logo?

[The Grim Reefer]

I don't see Logo getting much use these days. ;-)

Re:Logo?

[Qzukk]

But APL lives on!

Re:Logo?

[The Grim Reefer]

INTERCAL was great. In a funny, ironic way of course. I absolutely love the idea of a COMEFROM command.

Re:Logo?

[DanielOom]

APL may be forgotten before too long, along with SNOBOL, MUMPS, ALGOL 58, and a host of other impopular languages.

Flash

[rujasu]

Also on the list: Adobe Flash. Please let them be correct about this. No more Flash. Anywhere. Ever.

Re:Flash

[dave420]

Flash is not a language. The language it uses - ActionScript - is an ECMA compliant scripting language, meaning it's essentially syntactically JavaScript.

Re:Flash

[Anonymous Coward]

flash isn't a language, its a platform. the language that it uses is actionscript 3, which happens to be a decent language - object oriented, robust, mature, reasonably decent to work with. it has a few quirks, but i haven't met a language that didn't. the flash player is the issue - it's buggy and full of security holes. then, take the fact that Flash Pro allowed amateurs to create content that used machine generated code (which is rarely good), combine it with how ridiculously widespread the flash platform used to be, and you have a recipe for disaster. i happily agree that flash (the platform) needs to go away, but im more than happy to keep using AS3 - as are most developers who have used it. its quite nice on the AIR platform (although you have to know what you're doing to get good performance out of the platform. on the plus side, it weeds out the amateurs pretty quick), and there are some projects in the works for converting AS3 code to C++. there's also HAXE, which is very syntactically similar to AS3 and has excellent performance.

Re:Flash

[tepples]

So what's better than the Flash platform for creating and exhibiting things like Lesko's Revenge? Just rendering to video takes 10 times the bandwidth, which means viewers hit their caps 10 times faster.

Re:Flash

[rujasu]

Yes, I'm aware that Flash itself is not a language. It still needs to go away. ActionScript can go with it.

Re:Flash

[GoodNewsJimDotCom]

Except for the fact that it is the best language to develop for mobile.

Its a very easy language to write in compared to C/C++, and it ports to Android/iOS/Web/Desktop(Linux/Mac/Windows).

The people in the know are writing cell phone/tablet aps in Adobe AIR now

Its actually quicker, more efficient, and you get more done in it than if you tried to code natively in Android or iOS.

Flash developers really have an unfair advantage to developing aps :) But you guys just keep saying its dying.

Re:Flash

[rivercityrandom]

But if they kill Flash, how will we get our fill of Newgrounds and Homestar Runner?

Scripting language du jour

[Pro923]

This is why I don't waste my time with Python. There will always be a latest and greatest scripting language to come along and replace the previous one. We all know that real code is written in C/C++, but it seems that in the corporate world this has been deemed too difficult to understand. The recent trend that I've noticed is to create your system from piles of scripted modules. Also, part of the complexity in C++ is self inflicted. Years ago, C++ code was like a more flexible C - but with cool objects that you could use to create flexible, inheritable objects. More recently, people have taken the whole template aspect to an extreme and it (in my opinion) has really screwed the whole thing up.

Re:Scripting language du jour

[Pro923]

I would agree... But in practice, every UI that I've used that isn't built from native code works like total crap. I don't understand why, but it just always seems slow, sluggish, and the components just don't work exactly like their native counterparts (like you'd expect them to function). Also, in reality every implementation is a bit different. Finally, we add an entire additional layer of security holes (all of the problems that I fix these days on my friends computers seem to result from unpatched java or adobe products).

Re:Scripting language du jour

[bangular]

Ruby was supposed to be the Python replacement and it never happened. Love it or hate it, Python has some staying power. Not because it's better than other languages in any meaningful way either.

1) It seems to be popular in the scientific community, especially with non-programmers. In my experience, scientists will learn a language they like and use it their whole careers. It's the same reason R is still going strong.

2) There's an active community of volunteers. A sense of community can take something average and make it very popular.

3) Large companies use it. Google being the most notable.


Most of the scripting languages mentioned (and I got on the bandwagon for many of them) never really #1 and #3. They have community for awhile, but it fades away as the next big thing comes around. Python's community is as active as ever.

Re:Scripting language du jour

[Anonymous Coward]

R has a huge set of statistical operations built into it. There's no replacement in the languages pipe-line.

Re:Scripting language du jour

[phantomfive]

More importantly, univerities teach Python.

Re:Scripting language du jour

[steveha]

...I don't waste my time with Python. There will always be a latest and greatest scripting language to come along and replace the previous one.

Maybe so, but Python is getting more popular and widely-used rather than dying out.

IMHO Python hits the sweet spot: it's powerful and expressive, yet the code is readable and maintainable. The worst thing about Python is that it's pretty slow, but it has a vast library of extensions (written in C or even FORTRAN) and the extensions aren't slow. (Like, if you wrote your own FFT in Python it would be glacially slow, but you don't need to write your own FFT because fast ones are available... and if your program is mostly doing FFTs it will be nearly as fast as a C program, because the slow Python glue code isn't where the program spends most of its time.)

In the world of science, everyone is converging on Python because of SciPy (which rocks). As people get fed up with legacy systems, they adopt Python as the replacement. I attended a keynote lecture at the SciPy conference a few years back, and a senior guy from the Hubble Space Telescope project talked about how they were leaving a language called IDL and switching to Python, and how much happier they were with Python.

I have heard that the Ruby guys had a project to make a "SciRuby" but (a) progress was slow and (b) the science guys are already using SciPy and won't switch unless some really compelling advantage appears.

Python is a clean, well-designed language that can have anything you need put in as an extension. So you can replace Matlab with Python and it's mostly a win. You can replace R with Python (and I think it's probably mostly a win, but I'm biased toward Python and have never seriously used R so feel free to ignore my opinion).

Python can be used by sysadmins, web site developers, cloud app developers, scientific researchers... really almost everyone can do their work in Python, and they can talk to each other about it if they are all using the same language. That's not a trivial benefit.

So, IMHO you would not be "wasting your time" to try actually using Python.

Re:Scripting language du jour

[biodata]

The more I see of python the less I like it. Every time I try to install some python package there is some incompatibility with the system python, or some wierdness in compiling the underlying libraries, and I have to maintain several versions of the language. For some reason it doesn't seem to be backward-compatible, which leads me to question the sanity of whoever decided to make it that way. It is way worse than perl that at least 'just works' a lot more of the time. Maybe it's just me?

Re:Scripting language du jour

[micahraleigh]

Having been a C/C++ programmer for 7 years and a C# developer for 4 years, I can tell you for sure the C#(/Java?) guys can crank it out faster, and for most of business the bottleneck is development.

The anecdotal indicator is that C#/Java devs stand to make a lot more money and see a lot more listings.

But I can't comment on how C++14 will or will not affect that except to say people are starting to look more seriously at C++ now.

Re:Scripting language du jour

[ceoyoyo]

R is going strong because there isn't really an alternative. It's a pretty horrible language, but that part is copied from something else (SPlus).

Python is much the same. R evolved to replace things like SAS, which are expensive, proprietary and clunky for research. Python got adopted to replace MatLab, which is expensive, proprietary and can be clunky for many kinds of research. In both cases, there's a lot of ancillary stuff that needs to work well - plotting, signal and image processing, interfaces to peripherals, databases, etc. In science, whatever most people choose is going to become dominant and stay that way until there's a good reason to switch.

Re:Scripting language du jour

[ceoyoyo]

The best thing about Python is Cython. You can use all the nice bits of Python and then when you want something to run fast you can write some C code and just use it.

I often use Python to help write C where standard debuggers don't cut it. Need to graph something or display an image to debug your algorithm? No problem. And when you're done, you have Python wrappers as a bonus.

Re:Scripting language du jour

[cerberusti]

Really? C is quite possibly the most portable language which currently exists if you care to write it that way.

I nearly always use C for programs which must be portable due to the ease with which it can be made to run on a variety of platforms. I have quite a few which run on more than a few entirely different platforms with a simple recompile.

You can write C code which is tied to a specific operating system of course, but for any task where you have language options other than C you can generally write a portable C program which does that. The tasks which absolutely cannot be portable also tend to be the type of program where you have little to no choice of language anyway.

I also tend to pair it with another language for the UI unless I need something specific (although that really has more to do with my desire to hand projects off to someone else for minor UI tweaks, color changes, etc.) If you want portable UI code there are C libraries which can do that, although you end up with the same problems other languages with a portable UI library have (mainly lack of consistency with the rest of the UI, as you are not using the native API.)

If the project is the kind of thing where a web based UI is desirable, you can even make the UI language HTML. It is easy to have a C program bind to port 80 and respond to HTTP requests, or use CGI and a web server.

Re:Scripting language du jour

[__aaclcg7560]

That's the downside to Python being a teaching language. Like Java before it, many students will have no choice but to take every course in Python without ever being exposed to a different programming language. The last thing we need is a glut of Python programmers who adequately know Python but nothing else.

Re:Scripting language du jour

[shutdown -p now]

Why would you need that kind of stuff built into the language, though? With sufficiently flexible syntax, the libraries should do the trick.

And, FWIW, I hear a lot about people switching from R to Python (with the corresponding libraries, naturally - including an R bridge for gradual migration) lately.

Re:Scripting language du jour

[shutdown -p now]

Unlike Java, no-one is trying to sell Python as "the language to rule them all", though, and the performance makes it clear that you want other stuff for some things. Heck, even the official docs have that huge section on writing C extensions.

Re:Scripting language du jour

[shutdown -p now]

It's mostly just you.

With respect to different versions not being backwards compatible, it's more of a branding issue. There are two languages. One is Python 2, the other is Python 3. Within each of those, there's backwards compatibility, but between them, they're not compatible in general (though with some care it's possible to write code that works in both, and with some more care it'll even do the same thing - but you'd best leave that to library writers...). For your own project, pick one and stick to it. For other people's code, you might need both 2 and 3, but you should only need one version of each.

Re:Scripting language du jour

[stub667]

Well done. I've not heard Python descripted as 'scriping language du jour' for a decade. I tend to date the time when Python started getting taken seriously at version 1.5.2, April 1999.

Of course, if real code is written in C/C++, then no true scotsman would ever consider scripting languages the right tool for the job :)

Re:Scripting language du jour

[Hognoxious]

So instead of the usual backwards/forwards conundrum (if that wasn't fun enough) we have to consider another dimension as well? Is it left/right? Strange/charm perhaps?

Re:Scripting language du jour

[shutdown -p now]

There is no other dimension. They are two related but different languages, so treat them like e.g. C# and Java.

Re:Scripting language du jour

[rdnetto]

There are two big problems with Python (and no, the whitespace thing is not either of them).

The first is that the Python community has effectively been fragment by Python2 vs Python3. Many libraries support one but not the other, which can lead to dependency hell.

The other big problem is duck typing. No language which uses duck typing by default (with no way to strongly type variables) should be used for programs more than a few hundred lines long. Strongly typed variables significantly improve the static verification of programs, and throwing something like that away just seems foolish. The most common justification I've seen for it is that it's more concise and saves time, but there are plenty languages which are strongly typed and use type inference instead of requiring explicit declarations (e.g. Haskell).

Re:Scripting language du jour

[Hognoxious]

C# and Java, which also differ only by a numeric suffix. Just like VB and COBOL or Pascal and Lisp.

Seriously, who thought naming them like that was a good idea?

Re:Scripting language du jour

[shutdown -p now]

Perl (5/6)?

Re:Scripting language du jour

[Dutch Gun]

I've been programming C++ for almost 20 years and C# (part time) for about 7, and I'd agree with your general assessment about productivity. If you don't meet the requirements for using C++, you should absolutely be using C# or a higher-level language. It's not only the language that's easier to use - the .NET framework has a lot of really useful functionality and is far less obtuse than older APIs like the Windows native API (although that's a pretty low bar to hurdle).

If you need maximum performance, seamless interoperability with C and other C++ APIs, and excellent platform portability, then you use C++. C++ 11/14 don't really affect the language's use case a significant amount, in my opinion. It just makes things a whole lot nicer for those who would still be using C++ even if it was still stuck at C++ 98 because of the core language characteristics and ecosystem available. Some developers are coming back to C++ because many of the predictions about native-equivalent performance of managed languages really never panned out, at least in non-contrived real-world examples. A lot of us, though, like in the videogame industry, simply never left.

My guess is that C++ is rapidly evolving because of the renewed interest in it due to it's performance characteristics, not the other way around. Who knows for sure, though...?

Maybe Perl is just "complete," not dying.

[Arakageeta]

"Perl is an excellent candidate, especially considering how work on Perl6, framed as a complete revamp of the language, began work in 2000 and is still inching along in development."

This does not imply that Perl is on its way out. I don't use the language myself (I despise it, personally), but I know many who use it on a daily basis. It is still a go-to language for many programmers (albeit, who may no longer be in their 20s) who need to quickly hack together a test harness for a larger system. It could merely be that Perl is "complete" for applications where it is useful. Further revision is no longer necessary.

Also, I'd hardly say that C++ is on it's way out, even though C++11 took so long to be ratified.

Re:Maybe Perl is just "complete," not dying.

[marcosdumay]

Perl was once my go-to language for small scripts.

It isn't anymore. Perl is so full of idiosyncrasies that it's easier to learn another couple of languages than to code a couple of scripts on it. Even after it's your go-to language.

Re:Maybe Perl is just "complete," not dying.

[Arakageeta]

I agree. I like to use python instead.

Re:Maybe Perl is just "complete," not dying.

[fahrbot-bot]

It is still a go-to language for many programmers (albeit, who may no longer be in their 20s) who need to quickly hack together a test harness for a larger system. It could merely be that Perl is "complete" for applications where it is useful.

Careful grasshopper. Ya, I'm 51 and have been using Perl since it was invented - along with Emacs. But Perl can be a go-to language for anyone, *even* those still in their 20s. I currently develop software that runs on Solaris/Unix, Linux and Windows using about 10 different programming languages and among all of them, Perl is the most useful (with Java second) for cross-platform things. We also use Python for some things, and it could probably replace Perl for others, but, seriously, why bother.

Newer doesn't always mean better and old doesn't always mean obsolete.

ADA: Programming Languages Never Die

[Anonymous Coward]

A friend of mine is currently re-writing some old ADA code.

What, never heard of ADA? It's based on PASCAL. It was a contractual DoD standard for a while.

https://en.wikipedia.org/wiki/...

Re:ADA: Programming Languages Never Die

[neoritter]

I'm unsure if you're suggesting Ada is dead, should die, or don't want it to.

Re:ADA: Programming Languages Never Die

[mr_mischief]

Not the AC, but I'd say Ada should live on in descendants. It should be split into the eight or so complete programming languages that they bundled together when the committee made it. One or two of them might be a nice tool.

Thinking back to my undergraduate days (late 70's)

[14erCleaner]

Fortran: will live forever
Cobol: ditto
PL/1: probably a goner
Pascal: is that still around?
LISP: was already for hipsters only by the 80's

COBOL - (Perl,Ruby)?

[MiniMike]

With COBOL still around, it's hard to take too seriously the claim that Perl or Ruby is about to die.

Why would you make that assumption? Have Perl or Ruby been suggested as replacements for COBOL? Is the future usefulness of a language based inversely on age? I'm not seeing the direct connection between the lifespans of COBOL, Perl, and Ruby.

Also, how can they not mention FORTRAN in the article? No self respecting article on the topic of "soon to be dead programming languages" in the last 30 years has failed to mention FORTRAN. I see it as a staple of these articles for years to come.

Re:COBOL - (Perl,Ruby)?

[plopez]

Nah, not Fortran. It is still widely used. Often linked into R or Mathematica applications. And Fortran 2008 is OO and natively supports parallel processing. When time and huge data sets need to be crunched it is the best tool for the job.Can Java or C# do all that? Data analytics anyone? Geological modeling anyone? Fluid modeling, perhaps used for wind turbines? Or maybe weather modeling?

All good problem domains for a fast lean language.

Re:COBOL - (Perl,Ruby)?

[MiniMike]

I didn't say that FORTRAN was going anywhere. I said that it's almost always listed in articles like this, and always will be. I still write and maintain code in FORTRAN too.

We got the kids a Kano ...

[MyNicknameSucks]

We got the kids a Kano. Daughter 1 asks, "Do you two know how to program?" Parents, "Does the Kano have Turbo Pascal?" Daughter 2, "Is that like some kind of 80s cartoon character?"

Re:We got the kids a Kano ...

[Chris Mattern]

"The Kano"?

Well, it'd be appropriate to have a language named for an 80s cartoon character when the system is named for a 90s video game character...

Not going to happen

[Carewolf]

That is not dead which can eternal lie, in unmaintained hardware burried deep in your organization.

Re:Not going to happen

[WillAffleckUW]

That which is unmaintained is not broken.

Press cares only about stuff that breaks. Stuff that runs the power systems and never breaks never makes the news.

Re:Brainfuck?

[Anonymous Coward]

Nope, it was never the rage. Someone thought, "If I make the name controversial enough, at least people will talk about it."

Obscure ones

[Hydrated Wombat]

It will not be anything I know how to name. It will be the obscure ones that most people on /. haven't heard of.

That's not how languages die

[gurps_npc]

Languages die when people stop using them, not when they cease to be 'hot'.

Look at human languages. They die when the last person speaking them dies. What makes anyone think computer languages are different?

Re:That's not how languages die

[T.E.D.]

Well, there are actually multiple levels of human language death.

1. Extinct - No living speakers.
2. Dead - Perhaps still known, but no longer used in general conversation
3. Pseudoextinct - No living speakers, but there are speakers of a child language
4. Moribund - Speakers are shifting to other languages. Nobody new is learning it.

IMHO, most supposed "dead" computer languages are in actuality barely even moribund by the above definitions. There are some pseudoextinct languages though. K&R C would probably be a good example.

Re:That's not how languages die

[Livius]

I'm starting with the assumption that any five languages popular enough that the writer has even heard of them, might be on the decline but there is no way they will be the next five to die completely.

Re:That's not how languages die

[Hognoxious]

Esperanto: Latin with all the grammar taked out.

Re:That's not how languages die

[candlebar]

They die when the last person speaking them dies.

I think human languages die when the second to last person speaking them dies. The last speaking person would no longer have any use for the language--written or oral.

First, define your terms

[John Jorsett]

Before making a prediction, you have to define "death". If it's the complete absence of users of the language, then the answer is easy: no language ever dies. There'll always be somebody in some obscure corner of the planet who'll know and use it from time to time.

"Microsoft's long love of BASIC...."

[markhb]

From TFA:

Microsoft’s long love of the BASIC programming language extends all the way back to 1991, when the company purchased a pretty awesome (for its time) visual programming designer from Alan Cooper.

I'd say that MS's love of BASIC goes back at least a decade before that; they wrote the ROM BASIC for the TRS-80 (as I found when doing a PEEK scan through it).

Re:"Microsoft's long love of BASIC...."

[operagost]

It is a pretty ridiculous statement, considering that the BASICs for most early PCs and home computers came from Microsoft.

Re:"Microsoft's long love of BASIC...."

[tadas]

I'd say that MS's love of BASIC goes back at least a decade before that; they wrote the ROM BASIC for the TRS-80 (as I found when doing a PEEK scan through it).

Umm.. try 1975, when Harvard dropout Bill Gates and his friend Paul Allen wrote a BASIC interpreter for the Altair, the first microcomputer. http://en.wikipedia.org/wiki/Altair_BASIC/

Re:"Microsoft's long love of BASIC...."

[markhb]

If the WP article is accurate, Commodore's PET BASIC was a licensed Microsoft product, and as I mentioned so was the version in the TRS-80. So other than Apple's, I would say that the parent's statement that "the BASICs for most early PCs and home computers came from Microsoft", regardless of the fact that any other OS layer may or may not have, is accurate. And as others have pointed out, the first MS BASIC was in 1975 with the Altair, but I never used one of those so I went with "at least" as old as the TRS-80.

Re:"Microsoft's long love of BASIC...."

[psmears]

When the Apple II, PET 2001 and TRS-80 were all released in 1977, all three had BASIC as their primary programming language and operating environment. Upon boot, a BASIC interpreter in immediate mode was presented, not the command line interface used later.

This is where BASIC took off.

And who wrote the (level-II) Basic for the TRS-80?

Re:"Microsoft's long love of BASIC...."

[MooseMiester]

Yes, because all these people learned to program on PDP-11 RSTS/E... BASIC, and it's son BASIC PLUS was the first programming language learned by millions of college students, because smaller universities couldn't afford IBM 370's.... You had COBOL (Batch, JCL, definitely not exciting) and you had PDP-11 BASIC - that was done on a Terminal, you interacted with it, it talked back to you in real time. So you could sign up for the COBOL or FORTRAN course and learn to make punch cards... or you could sign up for BASIC and use the terminal, maybe even play a game.

The other hot language was FIG-Forth, because people wanted to write their own TIL (Threaded Interpretive Language) -- All the scripting languages ever written have been TIL's, which goes to show you just how little has really changed since the 70's when it comes to scripting...

As to any of these languages "dying" dream on. There's still jobs maintaining COBOL-74... We just got a request to convert a VB6 application into VB.net. Seriously... There's probably millions and millions of lines of VB3 still out there. VBA is the most horrible language ever, and it still ships with office, it is positively ghastly as languages go but if you need to do something fancy in Excel, it's what you use.

Re:"Microsoft's long love of BASIC...."

[david_thornley]

As far as Apple goes, you might want to look at Applesoft BASIC. It pretty well replaced the original Integer BASIC. FWIW, the original TRS-80s could come with Level I BASIC, which was a 4K BASIC. When I got mine, it took some time for enough Level II BASIC ROMs to be shipped, so I spent a couple months with Level I.

Re:"Microsoft's long love of BASIC...."

[cwsumner]

From TFA:

Microsoft’s long love of the BASIC programming language extends all the way back to 1991, when the company purchased a pretty awesome (for its time) visual programming designer from Alan Cooper.

I'd say that MS's love of BASIC goes back at least a decade before that; they wrote the ROM BASIC for the TRS-80 (as I found when doing a PEEK scan through it).

Microsoft originally made their money on MBASIC, which ran on the 8-bit CPM operating system. It was at that time the most popular OS in the world, if you go by number of installations. And MBASIC was on almost all of them. It was also converted to other computers, or the computers were upgraded to run a daughter board just so they could run MBASIC.

But it required more RAM than many computers had, and the available boards were smaller static RAM boards.

So Microsoft came out with a 4Kb S100 Dynamic RAM board. (Yes Kb. And yes they once made hardware.)

But the boards didn't work. 8-(
In fact, I never heard of a single one working. We put instruments on one, and even tried changing the connections. The chips just were not compatible with the CPUs used. Any of them.

So Microsoft required that -every- copy of MBASIC be bundled with the RAM board. People screamed, but they wanted the MBASIC so much that they paid it.

Microsoft never made any more hardware, AFAIK. But they still do business that way, sometimes...

Re:"Microsoft's long love of BASIC...."

[Hognoxious]

Cicero? No, too recent. God, maybe?

Netcraft confirms it

[Blackknight]

I think it's pretty safe to say that we can dismiss claims that perl is dying as pure rubbish. Sure, the language isn't as trendy as ruby or whatever the new hotness is but it's still a language that is used by thousands of companies and applications every day and will continue to be used for quite some time.

Re:Netcraft confirms it

[iggymanz]

No, Perl is being thrown aside for new projects, the built in utils and things in distros that would have been done in Perl now done more and more in Python. And it is not even on the list when most corps build or choose web frameworked apps

Perl is dying, Larry is killing it.

Re:Netcraft confirms it

[Blackknight]

Maybe where you work. Our company is still 100% perl for our core products.

Re:Netcraft confirms it

[iggymanz]

Having worked for ten years at place with dozens of large client corporations, your employer is very unique and peculiar

Single-language platforms

[tepples]

Sure, some languages will decrease in popularity, but they're still there to use if you want, and there will always be a die-hard community of fans that keep it alive.

But that doesn't necessarily mean that skills in the language will help you put food on the table. For example, if one popular client platform requires that all programs be written in C#,* and another requires either Objective-C++ or Swift, and another requires Java,** and another requires JavaScript,*** then companies hiring programmers will disproportionately demand those languages.

* The platforms I'm thinking of require verifiably type-safe CIL compatible with the .NET Compact Framework, which in practice means C#.
** Or another language that compiles to 100% pure JVM bytecode.
*** Or another language that compiles to JavaScript, even on JSVMs that don't specifically accelerate asm.js.

Re:Single-language platforms

[PhilHibbs]

Knowledge of a variety of programming languages, even if some of them are dead, will always help. Even if it's just in "I remember when Perl was still a thing" bragging competitions.

Re:Single-language platforms

[__aaclcg7560]

then companies hiring programmers will disproportionately demand those languages

Let's not forget the HR requirement of having five years of experience in a technology that just came out six months ago.

Re:Single-language platforms

[Austerity Empowers]

Or you just keep writing in C and rely on tools to create code in those bullshit languages for you, because who wants to define their career by the platform-du-jour? Sure, someone will have to know C#,Java,ObjC,Swift,JS, but not really that many people.

This stuff would go away if we could just agree to boycott these things. Corporations know how to make products, they are absolutely terrible at creating anything that lasts.

Re:Single-language platforms

[tepples]

So what's the preferred tool to compile C to verifiably type-safe CIL or JVM bytecode? The subset of C++/CLI that corresponds to standard C is not verifiably type-safe.

Re:Thinking back to my undergraduate days (late 70

[MobyDisk]

Pascal: Delphi still lives.

Let me save you some time

[Somebody Is Using My]

Here are the dead and dying languages

1) Perl - because it's a "piecemeal" language with features pile atop one another
2) Ruby - because its difficult to learn if you know C
3) Visual Basic.Net - because C#
4) Adobe Flash & AIR - because iPhone
5) Delphi Object Pascal - because it isn't well-supported

Now you don't need to read the article

Re:Let me save you some time

[TechyImmigrant]

That article was written by a youngster.

Pilot, PIC, mathchat, hypercard, IPL, SIMULA, etc.

These are the dying or dead languages.

Re:Let me save you some time

[charlieo88]

Dang, wish I saw this great summary before reading this stupid click-bait of a... this fine article.

Re:Let me save you some time

[Dogtanian]

Visual Basic.Net - because C#

It's arguably misleading that Visual Basic .Net was chosen as a dying language, because I don't think it was ever that successful in the first place.

It's more VB6 (and its predecessors) that are the ones that were once very successful and now dying if not almost dead by now. Despite sharing a similar syntax, VB.Net was never the same, as it was .Net-based, and not directly compatible with its predecessor. As the article notes, since MS basically forced everyone onto .Net, everyone chose C# over the retro-flavoured VB.Net anyway.

Re:Let me save you some time

[Anonymous Coward]

That article was written by a youngster.

Worse. A youngster has access to wikipedia, and can easily and quickly research programming languages. This article was written by an idiot, or someone trying to troll for clicks. (More likely alternative: both.)

Re:Let me save you some time

[MooseMiester]

You'd be surprised how much VB.net is out there in corporate America.

Re:Objective C, Ada, Tcl, Lisp

[neoritter]

Ada is still around an quietly chugging along. It's still used in the fields it was designed for, missile systems etc. And they just had a new stable release just shy of two years ago.

What about those dead before being born?

[ramirodt]

I mean Dart and Go particularly.

Perl isn't going away anytime soon...

[labmixz]

Lets be honest here, Perl is sticking around for awhile, it's more widely used than people realize, especially in the realm of bitcoin applications/interfaces with the hardware miners, so needless to say Perl has been picking up more traction due to that alone in the last couple years, not to mention RaspberryPi's and all the other DIY boards that have become popular. However for the other languages on that list, (at least my personal opinion)... Ruby seems like it died awhile back, it was a very short lived language in the eyes of many true developers. I personally saw the biggest gain/support of Ruby from "script kiddez", which also made me shake my head. VB.NET has seen to be on it's why out for some time, the last time I can even remember coding in VB was VB 3.0, granted it was useful back then, for certain aspects. Adobe Flash & AIR, Flash's wide adoption, basically due to very few options at the time when Flash reigned supreme made it so popular, but it's absolutely horrid these days, the fact it's held on this long is a complete mind-fuck, one can only attribute that due to the wide adoption, of which companies/people aren't willing to pay/learn enough to move away from it. One can only PRAISE the day Flash is complete no more. Objective Pascal, yup, that's dead, somewhere in the far corners of the world we have a few developers trying to maintain some code base, begging the higher ups to re-write it in another language.

Re:Thinking back to my undergraduate days (late 70

[LWATCDR]

Pascal was/is a much better language than Fortran or Cobol.
I would be shocked if it completely died out.
The one I wonder about is Java. It has sort of replaced Cobol as the language that you use to write programs that no one ever sees but will it keep that place now that Oracle bought it. I know that it is the language of choice for Android and that IBM and other people have their own JVMs but will Oracles lawyers kill it.

Latin?

[NotQuiteReal]

I've heard that Latin is a "dead" language, but people refer to it all the time, and even act smug if they know it!

Re:Latin?

[biodata]

Homo sapiens will never stop using Latin.

Re:Latin?

[cwsumner]

Homo sapiens will never stop using Latin.

Almost all of the European and American spoken languages -are- Latin, with local accents.

APL

[Squidlips]

Too bad. It was the most terse language ever created.

Re:APL

[LynnwoodRooster]

Brainfuck would like to have a word with you. Well, a few symbols perhaps - 8 to be precise.

what we'd want to see...

[buddyglass]

What we'd want to see is a ranking of languages by "new project starts" utilizing that language. There's still COBOL around but how many new projects are started that use COBOL? Etc. I suspect few people starting a project today that requires a Perl-like language would actually choose Perl unless they were already a Perl expert and it was definitely going to be a one-man job. They'd choose Python/Ruby/PHP instead. So, in that sense, Perl is dead.

Re:what we'd want to see...

[david_thornley]

Define "project". I'm not sure how many people would do a large new system in Perl, but there's lots of places where Perl (or COBOL or whatever) is entrenched, and thus that's the language for new projects.

Re:what we'd want to see...

[buddyglass]

My experience isn't comprehensive, but I haven't personally heard about any shops that still make extensive use of perl for new work. I'm not calling you a liar; I guess they're out there.

Re:what we'd want to see...

[david_thornley]

It depends heavily on what you mean by new work. Very few people would start a new Perl shop, but if there is one (and there is at least one) it's likely to continue to be used for new projects. It's hard to move away from a shop standard language, be it COBOL or Perl or whatever. (I know of two attempts to move away from COBOL, one of which was a disaster and the other a success.)

When does a language die.

[ZombieBraintrust]

A language dies when there are no longer programmers who are able to code in it. The langauge literally dies as people retire and die off. Well documented languages that are simple to learn will never die. COBAL is dying as a useful language as new programmers stop learning it and people retire. Economics still provide enough incentive for people to learn the language so it is dying slowly. They things listed at the top are still on peoples resumes. So they are not dying

Perl will never die

[Anonymous Coward]

We use it in power systems, we use it in scientific labs around the world.

Nice try.

Re:Thinking back to my undergraduate days (late 70

[elmohound]

The programming expert who wrote that article provided Pascal code that probably won't compile because of a missing "END" statement.

Re:Thinking back to my undergraduate days (late 70

[rgbscan]

I still hack around in Pascal from time to time. Kyan Pascal produces good results on my Apple IIe when I'm in the mood, and I dink around on System 7 from time to time on my old Quadra and all the libraries and toolkits for the Mac from back then are better supported in Pascal.

I don't think it has any practical use nowadays, even with Lazarus out there, but it's still fun if you're into the vintage scene.

Yeah, seriously...

[sirwired]

Microsoft's Very First product, written before Bill Gates even dropped out of college, was BASIC for the Altair 8k. This dates it back to '75.

What about those dead before being born?

[slashdice]

They're not dead, they're just resting. The real problem is languages that were born retarded, like javascript. And were then shaken vigorously and given a bag of paint thinner to huff on, just in case that helps perk them up. Yeah, node.js, drooling over in the corner, I'm calling back to you.

Perl in Latin

[Anonymous Coward]

Lingua::Romana::Perligata -- Perl for the XXI-imum Century

http://www.csse.monash.edu.au/~damian/papers/HTML/Perligata.html
Abstract
This paper describes a Perl module -- Lingua::Romana::Perligata -- that makes it possible to write Perl programs in Latin. A plausible rationale for wanting to do such a thing is provided, along with a comprehensive overview of the syntax and semantics of Latinized Perl. The paper also explains the special source filtering and parsing techniques required to efficiently interpret a programming language in which the syntax is (largely) non-positional.

Re:Perl in Latin

[NoImNotNineVolt]

Mod parent up --- Awesome!!

Reminds me of the Coy module and proves once again that Perl is the most bestest language ever.

Re:Perl in Latin

[cwsumner]

Lingua::Romana::Perligata -- Perl for the XXI-imum Century

http://www.csse.monash.edu.au/~damian/papers/HTML/Perligata.html ...

Ok, we surrender! 8-)

Completely off the mark...

[Mysticalfruit]

Whose writing programs in PL/1 these days?
What about Delphi? Whose writing Delphi apps these days?
How about Cold Fusion?
What about ActionScript. After standing up the Air platform they've basically abandoned it as a language.

Re:Completely off the mark...

[Oligonicella]

Then let's find this Whose guy and make him stop. He's the one who's keeping them alive.

Re:Completely off the mark...

[Mysticalfruit]

Nice :-) I stated it as a question, so if someone knows who this "Whose" person is, we should ask if they're still writing programs in PL/1...

blah blah blah

[Anonymous Coward]

I have become so tired of these people who try to make engineering into a fashion show. "You wrote you code in what language? OMG, that's sooo 5 minutes ago"...

Languages are tools. Use the one that is appropriate for the task. Stop inventing "new languages" (which are usually more about the implicit libraries that come with them than the language itself) and think that you are advancing anything. We don't need any more languages. Stop worrying about how many keystrokes it takes to write a piece of code (that is irrelevant- how long it takes to get a stable,maintainable, shippable product is what matters).

Years ago, I saw a "time line" for programming languages - every 5 years there was a tag that read "death of FORTRAN predicted" - it repeated every 5 years for the entire time line... Some guy jumping on a soapbox and proclaiming himself to be the prophet of programming and delivering "verdicts" on programming languages is just silly.

Stop worrying about whether a language is "trendy". Using a particular tool doesn't make one a good programmer. Writing solid, robust, un-obfuscated code that is well thought out and architected makes one a good programmer. That can be done in any language..

Re:Adoption by large organizations limits extincti

[cant_get_a_good_nick]

I agree to this.

We have millions of dollars riding on perl scripts. Yeah, we want to move to python, but while we're on perl we're on perl. There's a lot you can do with maintenance and upgrading to better perl with better constructs.

A language is not like a cellphone. We don't toss perl because the new iPhone is out next week. Perl doesn't fade. There's not a battery that will slowly begin not charging as deeply as time goes on. Perl remains perl. The problem domain doesn't radically shift month by month where we need a new language every month. What we have works.

Re:Thinking back to my undergraduate days (late 70

[David_Hart]

Pascal was/is a much better language than Fortran or Cobol.
I would be shocked if it completely died out.
The one I wonder about is Java. It has sort of replaced Cobol as the language that you use to write programs that no one ever sees but will it keep that place now that Oracle bought it. I know that it is the language of choice for Android and that IBM and other people have their own JVMs but will Oracles lawyers kill it.

I'm not sure that you know FORTRAN or COBOL very well or you wouldn't be comparing them to Pascal.

Pascal and, it's less popular cousin, Modula-2, were meant to be general purpose programming languages.

FORTRAN is primarily a programming language mean for engineers and scientists because of built-in high precision mathematics. It's still quite popular in both fields.

COBOL was designed to be a business language that accountants and business people could use to write reports, etc. Java did not replace COBOL, nor was it meant to. COBOL has largely been replaced by SQL.

The point is that it would be difficult to argue that Pascal is a "better" language than FORTRAN or COBOL. Both FORTRAN and COBOL have unique features which allows them to be better than Pascal for certain functions. It's also why Pascal, which can be replaced with C, etc., will die out long before either COBOL or FORTRAN.

Re:Thinking back to my undergraduate days (late 70

[Halo1]

Pascal was/is a much better language than Fortran or Cobol.
I would be shocked if it completely died out.

Me too. Especially since I've been contributing for 17 years to the Free Pascal Compiler, and it supports more platforms than ever. I also don't see any particular declines in our download statistics or the bug reporting rate. Whether Borland-Inprise-CodeGear-Embarcadero Delphi will survive, that's another question. If they'd disappear, that would however be unfortunate for us too though, since many of our users use both products (Delphi for its polish and commercial support, ours for the multi-platform support).

LISP

[tekrat]

LISP
LOGO
PASCAL
BASIC
PL/1

And maybe that's it. Most other languages are still in use by various organizations and developers.

Re:LISP

[wol]

Nope. Some of us old people still use lisp. Quicklisp.org solved the library problem and stockpiling parentheses is cheap - they stack nicely. We do, however, have a not-invented here syndrome problem which may never go away. Google bought ITA Software in 2011 for just under $1 billion and ITA's main product is largely driven by lisp. So still around, just not talked about much.

Re:LISP

[pigiron]

LISP isn't going away. In fact there are new dialects being invented all the time. I prefer an interpreted LISP 1.5 like newLISP.

Re:LISP

[rivercityrandom]

As long as there is emacs, there will always be a LISP.

Re:LISP

[pigiron]

Truth.

Not BASIC

[pubwvj]

In the end we'll all get Back-to-Basics with BASIC so it may be the last programming language standing, much to the chagrin of so many snobs. :)

Re:Not BASIC

[ceoyoyo]

How is a high level interpreted language "basics"?

We'll all eventually decide assembly programming is best.

Re:Not BASIC

[pubwvj]

I'm so sorry to hear you're humor impaired. Don't worry, Bill Gates is working on a cure.

Re:Not BASIC

[cwsumner]

How is a high level interpreted language "basics"?

We'll all eventually decide assembly programming is best.

I programmed for a few months in 6502 Hex code only. Thats basics.
Well actually, the PDP machines and the MITS Altair, with the rows of toggle switches for Binary, were the real basics...

Compared to that, the arguments about what high-level language is better, is sort of like the medieval arguments about how many Angels could dance on the head of a pin. 8-P

wrong, LISP going strong

[iggymanz]

Delphi is a dialect of PASCAL, still around.

You're not into CADD are you, LISP is programming language for AutoCAD and also included as one of programming languages in Bricscad and IntelliCAD to be able to run code made for AutoCAD . Yes, I'm former AutoLISP developer.

The article is on dice.com.

[QuietLagoon]

The purpose of the article is to make dice.com (/.'s owner) appear to be a place where people can go to read articles about job skills and such.

.
The purpose of the article is not to convey any manner of knowledge on the subject.

It's chewing gum for the job seeker, no more, no less.

Re:The article is on dice.com.

[Trailer Trash]

It's chewing gum for the job seeker, no more, no less.

Yes. Good, tasty, onion-flavored chewing gum.

Re:The article is on dice.com.

[Vitriol Angst]

Actually, having a Job Hunting website give the heads up on which languages are dying makes perfect sense. Who else is going to first notice; "nobody is asking for a Perl programmer?"

It's like the Housing market as a bell-weather for the economy. More building starts means more money in circulation. More jobs in a programming language means there is demand for that skill set.

It could be a puff piece purchased to get clicks and then resumes, but I just posit the notion that collecting data on job positions is not a bad way to read the tea leaves.

Apple Pilot

[TechyImmigrant]

Apple Pilot is dead. It's so dead that a Google search for "apple pilot" brings up nothing related.
Google for "apple pilot language" and the first hit is an Apple II history page.

It deserved to die, but it's not just dead, it's erased from the internet, but not completely
 

Re:Apple Pilot

[david_thornley]

I had a Pilot implementation on my TRS-80. I played with it for a few minutes, discovered exactly what it was capable of, and never touched it again.

Perl 6 =! Perl 5

[jjn1056]

It would be incorrect to say Perl is dying because Perl6 has not delivered on its original goal of being a worthy successor to Perl5. Although Perl6 was originally slated as that replacement, the two languages (Perl5 and Perl6) have drifted significantly and there is no syntax compatibly between them (although there are some projects around who goal is to make it easier for Perl6 programers to load and use Perl5 libraries, and vice versus). As a result the Perl community no longer sees Perl6 as the replacement for Perl5. Now there are two separate development teams and the Perl5 team has recently release an update to version 20, and work on version 22 is in the works. The Perl6 team continues to work on Perl6 but many people in the Perl community see it more as a hobby and as a lab rather than as a language one would expect to use on the job. Nevertheless lots of ideas have emerged from the Perl6 'lab' and has influenced the more iterative development of Perl5.

Languages

[Thanshin]

Still pissed off about the fact that there is not one single language for programming. Someone has to make an interface that seamlessly translates anything to C++ bidirectionally on the fly.

Or, even better, an AI that translates binary to C++ organizing it in a human readable correct structure. With procedurally generated comments based on cyclomatic complexity. For example:
"DEV1: Yeah, the next part is a big fuck-up. I'm sorry."
"DEV2: Dear DEV1, after two days correcting your shit I swear I'd give my car for the chance to meet you in a dark alley."

Re:Languages

[Dutch Gun]

Not every task has the same requirements, and thus, a single language can't possible be the "best" tool for every type of job. Languages are often created to solve very specific problems. From a technical perspective, any Turing-complete language can solve any programming problem, but certain languages are often much more elegant at expressing domain-specific programming solutions.

A programmer complaining that there isn't one single master language for programming is like a carpenter complaining that he has a box full of tools instead of one tool that can do it all. It's much better to have multiple tools that excel at what they were specifically designed to do rather than a single tool that does a mediocre job at everything.

Re:Thinking back to my undergraduate days (late 70

[Frederic54]

I was programming in Ada last month, so this one will not die too...

Forth

[istartedi]

Nobody mentioned Forth. It must be dead. Half-joking here. The first user that leaps to mind is NASA. IIRC, they're known to have written some very robust Forth for probes. It's been a long time since I've looked at the early boot process of a BSD or Linux distro. Is it still there? The reason I think Forth might die is because I come across people who think something with a gig of RAM is a "tiny embedded system". With that kind of power you don't need Forth.

A few people are playing with other languages like Joy, which also use RPN. These are academic "esolangs" though. Can they ever be said to have been alive?

Re:Forth

[crunchy_one]

Of course, FORTH lives on and on as Postscript.

Re:Forth

[istartedi]

Oh wow, good one. Totally forgot about it. That's the thing about Forth. It's in the type of code that just sits there doing jobs so well that you forget about it. One might mistake this for death, but it's really maturity.

Perl6==perl-dev. 6 : Perl : : Fedora : Red Hat

[raymorris]

Yeah Larry has long said that Perl6 is the development playground. Things are tried out in 6 and what works well get applied to Perl 5.x. We may never have a Perl release called 6.x, and we don't need one. In some ways, it's like Fedora and Red Hat - try things out in the development branch and put the best stuff in the stable branch.

Complaining that Perl6 isn't stable "yet" is like complaining that Fedora isn't stable "yet" - it's not supposed to be stable.

What about those dead before being born?

[ramirodt]

I agree completely; what a wonky, flimsy thing JavaScript is, but it seems it's got a fast VM and now it's going to rule us all...

An easy metric

[Opportunist]

How much code is written in the language that satisfies these criteria:

- Mission critical.
- Impossible to port.
- Impossible to replace.
- Impossible to redo.
- Works as intended with no/minimal maintenance.
(note: "impossible" also includes "costs enough to sink the company if we tried")

More then 3 on "yes"? It's going to stay. Less than 2? Probably not so.

In other words, Cobol is still here because it satisfies ALL of these requirements. Ruby isn't because the only companies where anything done in Ruby was mission critical are dot.coms that folded before 2005. And those that didn't, porting/replacing/redoing Ruby code is far from impossible.

Re:Thinking back to my undergraduate days (late 70

[pigiron]

"COBOL has largely been replaced by SQL."

SQL is not even an actual Turing-complete programming language! You don'y know WTF you are talking about.

Polish programmer's are great (proof inside)

[Anonymous Coward]

2006 ACM Programming Contest Poland 2nd and 7th in field of 12 http://it.slashdot.org/comment...

FROM THE RESULTS POSTED ON THE FRONT PAGE, FINAL SCORES/PLACEMENTS:

1. Saratov State University (Russia) - 6 problems
2. Jagiellonian University - Krakow (Poland) - 6 problems
3. Altai State Technical University (Russia) - 5 problems
4. University of Twente (Netherlands) - 5 problems
5. Shanghai Jiao Tong University (China) - 5 problems
6. St. Petersburg State University (Russia) - 5 problems
7. Warsaw University (Poland) - 5 problems
8. Massachusetts Institute of Technology (USA) - 5 problems
9. Moscow State University (Russia) - 5 problems
10. Ufa State Technical University (Russia) - 5 problems
11. University of Alberta (Canada) - 4 problems
12. University of Waterloo (Canada) - 4 problems

*AND*

Iirc, it was even BETTER for them in 2005 in the same contest.

2007 too -> Polish Students won ACM Programming Contest http://pbarut.blogspot.com/200... + Polish students prove best programmers in the world http://www.naukawpolsce.pap.pl...

Recently (2012) same deal "Russian, Polish Universities Take Top Spots in ACM ICPC Programming Contest" http://www.acm.org/press-room/...

TOTAL STATS FOR THAT PROGRAMMING CONTEST ARE HERE -> http://en.wikipedia.org/wiki/A...

APK

P.S.=> Yes, "I have a dog in this hunt" (I'm polish by descent, but 1st generation U.S. Citizen) & regarding Object-Pascal itself, I most *definitely* do (for great technical reasons shown here (yes, "shameless plug") http://developers.slashdot.org...

... apk

VB.Net advantages

[Tablizer]

VB.Net has some nice features over C sharp. 1) type name is to the right of the variable name in declarations, which is how it should be for modern type names. 2) No semi-colons, nuf sed. 3) End-X block enders are better self-documentation and reduce loop and IF mix ups after fat fingering.

Re:VB.Net advantages

[Tablizer]

oh, and 4) cleaner CASE statements with no need for BREAK.

Note that if you are using an IDE, then #1 does not matter as much. But trying to find the variable name among long type names is an eye killer if there is not colorized differentiation.

Re:VB.Net advantages

[Tablizer]

oh, and 5) more consistent blocking syntax conventions across authors than curly braces because it imposes a certain line spacing that curly braces do not.

Ruby dying?

[l0ungeb0y]

FFS -- what a crock of shit. I personally don't recommend Ruby to clients any more in favor of NodeJS, but there's still plenty of startups going with Ruby. A hell of a lot more than Perl. But with that said, if it were up to me I'd happily see Perl and PHP die. While Perl is definitely a dying language, PHP seems to be holding on -- with nasty frameworks like Yii (which feature views where anything goes) gaining popularity.

Delphi has been dead for a long time

[halfdan the black]

Has anyone seriously used this for anything other than writing Windows shareware in the 1990s?

I really have no problems with the language, even used it in the 1990s, actually pretty decent. But what an idiotic company behind it. Basically Borland's company strategy was: "Lets go muzzle to muzzle against Microsoft with our closed source language that only runs on Microsoft Windows, and lets charge a lot of money for our language thats completely incompatible with, but sort of resembles Microsoft C# when Microsoft lets you have Visual Studio for free".

If they would have open sourced so it could be cross platform, it or had some way of using your existing C++ code they might have had a chance. But yet again, a half way decent environment ruined by a completely idiotic company.

* yes, evidently the new version is "cross platform", but it a joke. The thing only runs on Windows which generates a binary that looks like it has some sort of emulation environment in it, then copy and hope that it runs on a Mac. And, they don't even remotely resemble Mac apps, looks similar to an Gnome/Qt app running on Linux with an OSX theme.

Re:Delphi has been dead for a long time

[Windowser]

Has anyone seriously used this for anything other than writing Windows shareware in the 1990s?

Lots of big programs have been written in Delphi, the most famous one being Skype (before MS bought it and rewrote it. Note how it became huge and slow).

or had some way of using your existing C++ code they might have had a chance.

Delphi can actually compile C++ code.

Re:Delphi has been dead for a long time

[shutdown -p now]

Delphi can actually compile C++ code.

You're thinking about C++Builder. They are two different products, even if they happen to be owned by the same company, and have some compatibility in terms of using libraries across two languages.

What about the children? Think of the children!

[plopez]

The primordial ooze of sed and awk from which sprang forth Perl whose ugliness and intractability spawned Python and Ruby. LISP whose offsring are a multitude; Scheme, Close, Scala, R, S+ and more. COBOL, assembly, and Fortran whose bastard offspring was BASIC, the black pit of which sprang forth VB, VB script and a host of shambling imitators. C which gave forth C++ and Objective C. Pascal which brought into the world Delphi and UCSD PASCAL. UCSD Pascal which when grafted with C and C++ gave us Java and C#.

They were fruitful though and so will never truly die.....

Re:What about the children? Think of the children!

[mr_mischief]

You forgot Ada, a child of Pascal.

Re:What about the children? Think of the children!

[mvdwege]

No, they forgot Algol, which begot both Ada and Pascal.

Useless article with no information content

[WaffleMonster]

There is no useful or objective information anywhere in the article it is all childish name calling and appealing to what the cool kids are doing.

TFA is what I hate about this industry too many people have their heads in what's cool and getting suckered by marketeers rather than thinking about what they are doing and investing necessary effort to research and arrive based on objective criteria the best tool to get the job done.

dice.com clickbait

[idontgno]

Complete with unsupported (unsupportable?) assertions ("perl is dead! DEAD!") on completely wrong-headed propositions ("A programming language can die," which apparently means that "the cool kids don't want to get their precious genius minds dirty with it because it's too mainstream.")

And yet here I am, participating in the conspiracy by commenting. I hate what Slashdot has made me become. I should have walked away months, maybe years, ago.

PL/M

[nateman1352]

There is actually one language that I can think of used to be popular and significant that is actually now dead: PL/M

CP/M was written in PL/M (the OS that MS-DOS is based on.) Later versions of CP/M had most of the code rewritten in assembly for speed reasons. When Microsoft converted it from the 8080 to the 8086 for PCs after version 1.0 one of the things they focused on was replacing the remaining PL/M code with C code. It didn't take much time before MS-DOS was completely free of PL/M code.

Fast forward to today and there isn't a single modern PL/M compiler out there. Pretty incredible really considering that today all it takes is 1 guy deciding to spend about 6 months writing a LLVM frontend. The last one was PL/M-386, which dates to the 80's, everything newer than that focuses on converting PL/M code to C code. I would be surprised to hear about a single new software project being started today in PL/M, and I expect that the number of programmers actively writing PL/M code is a 2 digit number.

Amazing when you think about it that a language used to implement an OS which the world's most popular OS is descended from is dead now.

Re:PL/M

[squiggleslash]

PL/M is indeed dead, but I'm not sure where you get your history of MS DOS from.

MS DOS is based upon Q&D-OS/86DOS. 86DOS has nothing in common with CP/M beyond implementing the same API for backward compatibility reasons - there's no shared code, the user interface (command line shell) is entirely different, and probably the most complicated bit of "that type of operating system", the file system, is completely different in virtually every way with the superficial exception of using 8.3 filenames.

86DOS was written in 8086 assembler, not PL/M or C, and was never available for the 8080. It was written as a stop gap because the manufacturer of an 8086 card for (IIRC) S100 systems (that bit I may have wrong) was waiting for DR to release CP/M86 and DR was taking rather a long time to get around to it.

Later Microsoft probably did rewrite some of 86DOS in C, but I believe MSDOS 2.x and 3.x were still primarily written in assembler.

CP/M itself has a development history and it's possible DR rewrote parts of it in C as the software lurched slowly to Concurrent DOS in one direction, and DR DOS in the other via DOS Plus, via CP/M Plus (CP/M 3.) But MS DOS? Totally different history. Patterson's decision to implement the CP/M API has led to a lot of confusion over the years, but anyone with familiarity of both operating systems can tell you they really aren't related.

Reversed order

[jtara]

The order is roughly reversed, the bottom two are already dead. And we can only hope that the third from the bottom would die but it won't.

No real justification is made, save for a couple of code examples for each language that the author finds somehow absurd. Oh, and he marched out the Twitter argument for Ruby. Yes, it's true: not every language is suitable for every purpose! No, Ruby was not appropriate to use for EVERYTHING to bring Twitter to scale.

While putting .net on the list won't make it die, I wish the author would put Java on the list. It won't die either, but we can hope.

not a worthwhile article

[jlv]

A sample of the writing in the article:

"Perl, which works as a CGI scripting language, found its most popular use in generating Web pages."

Clueless drivel.

Re:not a worthwhile article

[jlv]

Another one:

"Microsoft’s long love of the BASIC programming language extends all the way back to 1991."

I guess the prior dozen years just don't count.

Re:Thinking back late 70's Algol, SNOBOL etc

[shoor]

I was a CS major in the late 70s too. We learned to program at least a smattering of Norwegian University Algol on our Univac 1106, as well as SNOBOL (StriNg Oriented and symBOlic Language), APL and something called XL6 (a 'list processing language' so obscure it's not even in the wikipedia, but I'm pretty sure I got the name right.) In other classes I learned some specialty languages, Dynamo (Dynamic Models) and GPSS (General Purpose Simulation System).

Whatever happened to Algol anyway?

As I recall, SNOBOL was actually pretty cool, sort of like Pearl in that you could learn to do some useful things very quickly in it. I don't know that the code was all that readable even to the coder after 3 weeks though, since I never had to look at my code 3 weeks later. And I think even the professor who taught it complained about some of the choices for characters to use as operators. Still, I have this lingering feeling that somehow SNOBOL was a language that was unfairly passed over, maybe because of the comical name.

Re: 2/5 isn't bad?

[cyber-vandal]

Microsoft are not planning to toss the whole .NET ecosystem. I have no idea why people keep believing and repeating that shit.

What can replace Perl ?

[GuB-42]

What language is as good as Perl for throwaway scripts ?
If written countless perl scripts for countless problems such as extracting data from web pages, crude monte-carlo simulations, code generation, merging contact lists, etc... Most of this code is no more, in fact, a lot of it was one-liners directly written on the command line.
For this, I don't need a language that is clean, maintainable and readable. I need a language that is featureful, terse and lenient, and it's exactly what Perl is. As an added bonus it is installed by default in most UNIX systems.

Ah, and BTW, it is a legitimate question.
A lot of modern languages focus on maintainable code, which is a good thing. However, there are some cases where short term efficiency is more important. I thought about ruby (second in the list of dying lanquages...) but are there more recent alternatives ?

Re:I wish it was JavaScript

[narcc]

Try learning it. Your opinion will quickly change.

What a shit article.

[Xaemyl]

I've read my fair share of shit posted on /. over the years (including my own comments!), but this "article" has to rank up near the top.

Was it at the Windows 10 reveal?

[Anonymous Coward]

I'm pretty sure Microsoft said very recently that VB.net wasn't dead.

Re:Thinking back to my undergraduate days (late 70

[Wdomburg]

Java is going nowhere. In addition to being in most phones (Android, Jave ME), it is in every Blu-Ray player (BD-J), every cable box (OCAP), cash registers, ATMs and voting machines. And that isn't even touching on enterprise, web and desktop applications.

Take a look at most language rankings and Java remains near or at the top, whether you look at Tiobe, PyPL, RedMonk, IEEE Spectrum, or the various job surveys published by the likes of Dice.com and eWeek.

mentioned?

[mrego]

Has RPG been mentioned? Or languages such as BLISS ? BCPL ? WATFOR? SPITBOL? SAIL ?

... ducking now ...

[Anonymous Coward]

uh ... forth ?

Re:Adoption by large organizations limits extincti

[DamnOregonian]

Perl doesn't fade

That's partly due to the witch-craft and dark incantations inherent in the language.

I love Perl :)

Moose is an example

[raymorris]

I dodn't have an exhaustive list. Moose is one example http://act.yapc.eu/ye2013/talk...

Fortran's death long predicted

[Anonymous Coward]

Some 33 years ago, when the firsts computer magazines appeared and most articles were written by computer-science academics, who preached that Pascal was the only worthwhile language, I wrote an article of Fortran that the magazine (Practical Computing) re-titled "The language that refuses to die"

The most memorable of the critcisms in the letters column was "if it wasn't that it produced smaller programs that ran faster, no one would use Fortran"

It seems that, 33 years later, Fortran is still hanging in there, unlike Pascal

Ruby dying?

[slashdice]

Startups? 90+% of them are going to fail big time. Not exactly a ringing endorsement.

VB ~ Images Simple? Why would I use this....?

[TiggertheMad]

I like VB and VB.NET. When I see a coder willingly using it over another option, I instantly know they are an inept tool. Without it, I would have to speak with them possibly for several minutes in order to make this determination.

Re:VB ~ Images Simple? Why would I use this....?

[ConceptJunkie]

I had the unfortunate experience to be required to use VB for a project a few years ago, and in general I found it wasn't such an awful language. On the other hand, the application I was writing was for Excel 2007, and that aspect of it made me feel like I was doing Windows programming back in the early 90s, except with all the flaky OLE stuff added in.

Unlike the rest of Office, which I think is hideous and unusable, I always liked Excel, but after developing for it using VB for Apps, if I were in charge, I'd fire anyone who tried to depend on Excel for anything. It is literally, and by far, the most fragile and buggy development environment I've ever used in 25+ years of software development.

Re:Does "CBM Basic V2.0" count?

[__aaclcg7560]

As rare as IBM Basic is these days.

Article is pure clickbait.

[blerg]

Perl and Ruby and VBA (to name a few from the article) are dead and dying?

Author has their head up their arse.

Perl isn't going away anywhere close to soon. Citing perl 6's long development cycle is hardly proof.

Ruby is being used by more and more devop style projects. Puppet and Chef are hardly small time in that sphere.

VBA will continue to be used in corporate environments.

Re:Adoption by large organizations limits extincti

[the_humeister]

Like Brainfuck?

Tabs FTW

[AmericaRunsOnDunkin]

This is why tabs and only tabs are the proper line indentation character. One tab = indent one level. Two tabs = indent two levels. Tab is the only line indent char. It's a semantic definition. No more confusion.

So, here's my problem with whitespace being syntactically significant ... everybody likes to see code with different levels of indent.

Not a problem. Everyone defines how much space to render for a tab in their own editor. Everyone wins.

The problem was his electric mode in emacs was thinking itself oh-so-clever, and instead of storing the *actual* number of tab indents or whitespace, it just stripped them in favor of a single tab that emacs would then know how to render later.

Misbehaving tools are bad. You correctly locked him out until he fixed it. Problem solved.

For any of us who have taken compiler classes, a context free grammar specifically ignores whitespace. That's how compilers have worked for a very long time, if the grammar productions for your language involve counting whitespace ... well, my compilers prof would have failed me. Instead of having a visible thing to define a block, oh, well, just indent a few more chars.

Tab = indent. Space = ignore. Problem solved.

You can't see what character whitespace actually is ... is that 8 spaces or a tab?

Beginning of line = tab = indent. Elsewhere = space = ignore. Problem solved.

I've seen someone debug a python program, and even though things were in the same column in the editor, some were tabs and some were spaces, which had the very bizarre effect of making it semantically different than it looked.

That man was an idiot. Run python -t. Tabs and spaces should never be mixed. Tab is the only proper, semantic line indent char. Problem solved.

I have several issues with whitespace defining block structure, but the ones you identified are trivial to fix. The real issues are:

• communicating code through sources that collapse or mangle whitespace (html, some emails)
• difficulty with quickly and efficiently changing block structure, e.g. to temporarily bypass a conditional or refactor functions

The second is properly addressed by better editing tools. The first is annoying, but not enough to outweigh the benefits in clarity and expressiveness from using python. It's really quite easy to adjust.

Clojure

[laughingskeptic]

Scala seems to be kicking Clojure to the curb.

Re:He was one of a team

[Old Aylesburian]

Wrong. He was *in* your team, but not *of* your team.

Re:Adoption by large organizations limits extincti

[LordWabbit2]

Which is why COBOL won't die. It does what it is designed to do really really well. That and the cost to "just rewrite it in java" is too high. When people tell me "just rewrite it in java" I just shake my head and walk away. They are the type of people who think a seventy page website is a "big" project. I worked at a company where just printing the working storage section took a ream of paper, double sided. Granted whoever let the program grow so big should have been dragged into the street and beaten to death with the printer, but there are BILLIONS of lines of COBOL code out their, doing your taxes, house mortgages, credit card interest, etc. etc. Boring fvcken language though, glad I only did 4 years of it before getting out.

Re:Thinking back late 70's Algol, SNOBOL etc

[14erCleaner]

Whatever happened to Algol anyway?

By 1980, it had turned into Pascal, never to return. Wirth's compiler was a big reason - it was easy to port to new machines, so spread like a virus.

If it is part of Windows or Linux, it won't die

[bearvarine]

For example, take awk. Please...

No way!

[kgish]

Perl I can understand, but Ruby? No way!

Forth as an alive language

[Mister Null]

Forth is a threaded interpreted dictionary language that might on the surface seem dead until you realize it's usefulness in machine control, both for servos and stepper motors. As long as there are robots there will be Forth.

Re:Don't believe it

[Joe U]

I checked, Netcraft has NOT confirmed that VB is dying.

Sorry everyone, hastily written apps will continue until the end of time (or Windows, whichever comes first, odds are that time ends first.)

Re:Thinking back late 70's Algol, SNOBOL etc

[david_thornley]

ALGOL gave birth to a large number of other languages, such as Pascal and C, with features seeping into FORTRAN. It appears to have died out more because of how things are named. Common Lisp and Scheme are recognizably descendents of McCarthy's Lisp 1.5, but have a lot of differences, arguably more than Pascal and C have with ALGOL. The latest Fortran is a recognizable FORTRAN IV descentant, but has a lot of changes. However, Fortran is still considered FORTRAN, Common Lisp and Scheme are still considered Lisp, while no current language is considered Algol.

As far as SNOBOL, it was incredible at some things, but the control structures really, really sucked. I found it far too frustrating that way to write anything significant in it.

Re:Thinking back to my undergraduate days (late 70

[cwsumner]

... Pascal and, it's less popular cousin, Modula-2, were meant to be general purpose programming languages. ...

Modula-2 was what Pascal should have been.

But then the new versions of Pascal incorporated the new ideas.
And then many other languages incorporated them.

And Modula-2 was not so far ahead any more...

Need For Maintenance Keeps Languages Alive

[LinuxLuver]

The need to maintain applications that cost millions to develop is a powerful driver for keeping languages alive. The banking industry (and others) invested HUGE amounts of money in applications based on COBOL. Those applications still do the job, every day. Occasionally they need to be enhanced or extended....and it's cheaper to do that than start from scratch with something else. The languages that will die are those that never resulted in applications that can't be easily / cheaply replaced by something else.

Re: 2/5 isn't bad?

[cwsumner]

Microsoft are not planning to toss the whole .NET ecosystem. I have no idea why people keep believing and repeating that shit.

The rumor is:
Microsoft intended to write Windows in dotnet, that's what happened to Vista.
But the people doing MS Office balked and delayed, it turns out rightfully.
They never got Vista all the way into dotnet, and began removing the traces in Win7.
Of course now it is just beginning to be reliable. But it is probably too late.

There is some evidence to back it up. Some is in Windows it's self...

APL?

[sd.patricia]

What about the really obnoxious APL? (Who names something "a programming language"...) I used it off and on at university in 1974-1976 for programming statistical analyses. I was happy to move on to stat analysis packages such as BMDP and SAS.

Well there are many "fad languages"

[Casandro]

In short you can expect most languages younger than 30 years to disappear soon.
If a language doesn't have it's special reason to exist, it won't.

In case of COBOL and FORTRAN that was a huge amount of business critical code, as well as the possibility to simply run it on your next computer. C is seen as a "smart assembler". LISP and its family are great for logical processing. Java seems to become the new COBOL.

And there are 2 languages on the List which do have those special reasons. One is Perl, which is just a great tool for dealing with strings in a quick and dirty way, the other one is "Object Pascal" which, in it's original form with Delphi is dead, but lives in in FOSS projects like FreePascal and Lazarus, it has the great opportunity of having native code plus platform independent native GUIs. In short their way of doing GUI means that you can write a program on Linux, compile it on a Mac and it'll look and feel like a Mac program.

What will die soonish is of course the .net world as it completely depends on a single company... which doesn't use it for their important products.

re: Welsh

[Hognoxious]

At least they aren't pushing to make it an official language of the EU like some people I could mention. Yes, Ireland, I did look at you.

Because if there's anything better than translating from 25 languages into 24 other languages it's translating from 26 into 25.

Whatever pays the bills.

[Necron69]

My company is wedded at the hip to a test automation system built in Perl that dates back to the early 2000s. I grumbled a bit about continuing to use this system after a review two years ago, but it isn't really that bad, and it pays me six figures a year. Perhaps, like COBOL, the rarer it gets, the more valuable the skills will be?

I'll probably get around to learning Python or Java one of these days. :)

- Necron69

It's ok!

[iq145]

i'd rather just click on an icon!

no logical data?

[Champaklal]

i'm surprised the author never mentioned cobol, fortran, ada, haskell etc? also there are no numbers to back his claim that users are dwindling away...