The Best First Language For a Young Programmer
snydeq writes "Fatal Exception's Neil McAllister questions whether Scheme, a dialect of Lisp taught as part of many first-year CS curricula and considered by some to be the 'latin of programming,' is really the best first language for a young programmer. As he sees it, the essentially write-only Scheme requires you to bore down into the source code just to figure out what a Scheme program is trying to do — excellent for teaching programming but 'lousy for a 15-year-old trying to figure out how to make a computer do stuff on his own.' And though the 'hacker ethic' may in fact be harming today's developers, McAllister still suggests we encourage the young to 'develop the innate curiosity and love of programming that lies at the heart of any really brilliant programmer' by simply encouraging them to fool around with whatever produces the most gratifying results. After all, as Jeff Atwood puts it, 'what we do is craftmanship, not engineering,' and inventing effective software solutions takes insight, inspiration, deduction, and often a sprinkling of luck. 'If that means coding in Visual Basic, so be it. Scheme can come later.'"
I always considered Pascal (or Delphi) a great language for beginners. Powerful enough, structured, type safe and very elegant. From there, jumping to other languages is quite easy.
It's time to realise that Abble's products are the biggest abomination these days. Just say NO to the dumb iAbble way!!
No, it's *my* favourite language. Your favourite language is awful.
for the professors, that it. By removing all the syntax, etc, you can be introducing functions, lexical scope, binding, etc in the first week. Data structures and recursion in the second.
Result: most students quit by week two, and you are left with a fairly teachable remainder.
Just don't start with VB, PHP, Java or C# as it will screw the person up for lifetime.
I strongly disagree with this. Refusing to learn a new way of doing things will screw the person up for a lifetime. But the blame for that is on the person who is now screwed up, for being lazy.
It seems like TFS is dealing with the fact that there are at least two distinct, and at times temporarily opposed, aspects of being educationally good.
The one is engagement/excitement/comprehensibility: If somebody is disinterested in, or hugely frustrated by, a subject on first contact, they will have minimal motivation to continue. Unless you simply plan to beat it into them, introductory material needs to grab the audience(this doesn't mean that everybody must be your audience, of course). In many cases, this means a (temporary) sacrifice of rigor or correctness; think of intro physics, where you start with simplified Newtonian scenarios, or math, where you generally start by talking about addition/subtraction/multiplication/division, not sets and number theory.
The second value is that of being correct and rigorous, or at least not impeding later development in completeness and rigor. You obviously cannot learn everything all at once; but there are some simplifications that make it easy to fill in the gaps later and others that actively retard that effort. This can happen either because the simplifications are hugely different than the reality, and harden people in the wrong ways, or because, in an attempt to be "engaging" the intro stuff promises that the subject will be "fun", "relevant", and "exciting" to everyone, every step of the way. Fact is, that isn't true. Most subjects have, at some point or another, patches of sucky grunt work. Promising people that they are precious flowers who will never have to soil their hands with such is a good way to make them drop out when they hit those patches.
First, learn assembly, it teaches you how the machine works. (You should probably also learn electronics and digital logic)
Then learn C, it is the most widely used in both commercial and open source.
Then learn C++, it is a better C.
Then learn Java, it rules the web.
Then learn Python, it has some very clever ideas.
Finally...never stop learning
In my opinion, two issues compound each other. The first is that because functional programming is seen as very pure and simple, there is a myth that Scheme programs do not need much documentation. The second is that Scheme functions do not declare return or argument types. This means that in order to read someone else's code, if structures of any complexity are used, you can have to manually walk the full call depth of each function, possibly many calls deep, just to know what kind of structure it returns in the end. That makes it painful to work with someone else's code, compounding the other well-known problem with computer science education: in the course, you're usually writing your own code from scrach; in the real world you usually have to deal with code your colleagues (or even third party projects) have written.
http://www.alice.org/index.php?page=what_is_alice/what_is_alice
"Alice is an innovative 3D programming environment that makes it easy to create an animation for telling a story, playing an interactive game, or a video to share on the web. Alice is a freely available teaching tool designed to be a student's first exposure to object-oriented programming. It allows students to learn fundamental programming concepts in the context of creating animated movies and simple video games. In Alice, 3-D objects (e.g., people, animals, and vehicles) populate a virtual world and students create a program to animate the objects.
In Alice's interactive interface, students drag and drop graphic tiles to create a program, where the instructions correspond to standard statements in a production oriented programming language, such as Java, C++, and C#. Alice allows students to immediately see how their animation programs run, enabling them to easily understand the relationship between the programming statements and the behavior of objects in their animation. By manipulating the objects in their virtual world, students gain experience with all the programming constructs typically taught in an introductory programming course."
kulakovich
My first year as a CS major I took "symbolic logic" to supplement to required Pascal, Fortran, and Assembly Language courses. After all that, I always thought of the symbolic logic class as the "Latin of programming". Personally, I think any language which is free and gives quick results would be suitable for beginners...Python, for example.
Floating in the black seas of infinity without a paddle.
Its not consistent, its not even well designed I expect, but its a remarkably easy way to learn to manipulate a computer. Learn a bit of HTML first, some CSS, then work on OO PHP and you can accomplish a lot. People will dismiss PHP but there are a lot of very large websites built using it - ones that lots of kids will be familiar with.
Follow it up with a second language once you have gotten the basics down pat - Python is likely a very good choice.
"The first time I got drunk, I got married. The second time I bought a chimpanzee, after that I stayed sober" Arian Seid
Nowadays I would suggest $_language_of_choice as firs language as it is $_reasons[0], $_reasons[1] and $_reasons[2] language. Then extend it with $_arbitrarily_superior_language.
Just don't start with $_other_language[0], $_other_language[1], $_other_language[2] or $_other_language[3] as it will screw the person up for lifetime.
Please, I get so tired of arguments like this.
As long as:
who cares what languages they learn? If they enjoy it and it allows them to learn how to program why should it matter what language they start out with?
God, schmod. I want my monkey man!
In before old people telling you about punch cards.
Come to think of it, is there a way to do calculations with kids on/off your lawn?
The best first language for a young programmer is english with possibly a little bit of boolean logic, because then he could search Slashdot and find one of the Ask Slashdot stories about what the best first language for young programmers is that appear every couple of months or so.
People that knock the hacker ethic are a bunch of MBA drones that could never really build a damned thing themselves.
You learn to program by diving in and doing it. The more you practice and study, the better you get at it. GM was very good at shackling some very brilliant engineers and turning them into process drones. Look at where it got them. Great things are built by individuals and the more steps you have in the way of people being individuals, the worse you will get. Products have to be owned by the engineers that make them and they are personal works of art.
At the end of the day, the managers, bean counters, and all of these other people with their measurements, metrics and fancy charts are so much fluff, a tax on the capable in society... by really a bunch of leaches that could barely feed themselves as they lack the mental self sufficiency to do anything other than to try and ride the labor of others. We condemn socialism in society there's no real difference between the PM in a three piece suit and the lowest of the homeless people. Neither add any real value to society, its just that, the PM knows how to use PowerPoint and the homeless guy does not.
This is my sig.
ANY first programming language introduces new concepts. When you're starting out even something like the concept of a variable takes a little getting used to. Maybe you can relate it to memory store/recall on a pocket calculator, but with a name. Later you can introduce arrays of variables, non-numeric variables, etc.
You seem to have forgotten what it was like in the beginning to know *nothing*.
It teaches you how a computer really works. That way you can become a 'real' programmer instead of an IDE user.
Really? Seriously?
I don't think assembler is the best way to instill the magic and excitement of getting the most complex machine in your house to do what you want it to. And, that's what a fifteen year old newb needs. If you start with assembler, you're assuring that it will be months before he has learned enough to be able to take a program that he's written to a friend or parent, and have that person say, "Cool!". And, it will be even longer before he can use even a fraction of the modern technology that computers now have; things like GUI's, and networking. More often than not, it will only cause frustration.
Back in the day, assembler might have been the right option. But today, I think that's a recipe for killing that spark of creativity and excitement that draws people into programming, and gets them to slog through the nitty-gritty stuff.
... and that's when the C.H.U.D.'s came at me.
One not-so-obvious candidate: JavaScript and HTML.
Pretty much every browser in existence supports JavaScript, so with nothing more than a simple text editor and your browser of choice you can be off and running. As far as beginning programming is concerned, JavaScript easily encompasses any programmatic constructs you'd need.
The best part is that the students can easily display the results of their test programs in HTML, either dynamically generated or just by manipulating some divs, textboxes, tables etc that they've written on their page. Additionally, an instructor could write a 'playground' bit of HTML and JavaScript, so all output variables are bound up and easy to access. At that point the student is free to focus on what really matters, his/her first logic routines. When the student has created his first masterpiece, sharing the accomplishment with parents/peers is as simple as sharing a link to their HTML file.
I think this has the potential to engage students much faster than observing console output or fighting with a front end like windows forms in VB or Swing in Java.
Sauer
I don't associate Scheme with 'the hacker ethic'. I don't strongly associate any language with hacker sensibilities. I do associate Scheme with the intellectual rigor required for a programmer who really has a clue about programming.
I think Scheme is an excellent language to teach college students who think that they know how to program because they managed to smoosh together a bunch of working PHP code and make a website. I think it is a poor language to teach high school students who are learning their first language.
The association of technical knowledge with competence irritates me. A competent painter needs to know about brushes and mixing paint, the difference between oil and acrylic, and a whole host of other technical details. But that's not how you get a good painting.
One of my reasons for feeling that Scheme is a good language for people who think they know how to program is that such people frequently know all about paint but do not have the depth of understanding to be a good painter. Scheme is a language that forces you to think about programming differently than you did before. And if you understand it you are on the path to being a good programmer rather than just a code monkey.
But I would not recommend it as a first language. I would recommend Python for that. Clean, concise, expressive and powerful. It's my favorite language for a reason. :-)
Need a Python, C++, Unix, Linux develop
Since we're talking about learning languages here, I think we need to remember to balance "excitement of programming" with actual learning.
Assembler is a terrible first language because it doesn't really teach programming so much as learning how a particular CPU works. You could theoretically get away with something like MIX because it's just a simple emulation of assembly, but real assembly language programming is something that really shouldn't be attempted until normal programming is learned.
On the other hand, using a language that allows a student to create a "cool" application very quickly can't possibly teach them very much about programming and computer science. When there is too much "magic" and "gee-whiz" language support, exacerbated by a heavy-handed IDE, the student learns only how to press the right buttons to make their project work. They don't learn what is going on in enough detail to take those concepts to other more useful languages.
So a balance definitely needs to be struck. Python, as many have said, is a good language, though I would actually prefer one with static typing. BASIC is another good choice, though it's been much maligned by many. Personally, I think C is a great learning language, though it too has shortcomings.
Like anything, though, you can't satisfy everyone. But you might as well try to satisfy someone.
The problem with assembler, little Anonymous Coward, is that it doesn't let you do anything without a significant amount of work, and what you can do is unlikely to impress a fifteen-year-old kid just getting into programming.
Being able to print a few lines to the screen won't impress that kid and make him want to keep programming. Give him a language that can easily create GUIs, so he can see his stuff in action. To do this, I'd recommend an object-oriented language, maybe Python (though I personally detest it) or C# (which is a very nice language with very nice tools).
"You can either have software quality or you can have pointer arithmetic, but you cannot have both at the same time."
In Germany, researchers into didactics (teaching) of computer science (Informatik) have done some work on this topic. I recently found it when I was looking into materials for the computer science course in the Netherlands (seeing if I could do better).
Based on 15 criteria, they ranked 27 languages, ranging from Scheme to Haskell, ADA to Ocaml. The worst language for teaching was, by far, APL (scored a 5, which is the worst), followed closely by Perl. The best language for teaching was Algol 60 (1,50). Second best Python (1,66), 3rd place Ruby (1,88) and scraping in at a 4th spot was Pascal (2,14).
So to summarize: better dust off your Algol 60 books and compilers :P
Failing that, Python and Ruby are nice as well for just teaching programming (although if you want to show the distinction between imperative and functional programming I'm not altogether sure that Ruby would be enough).
------
This was found in a (Dutch language) PDF: http://www.utwente.nl/elan/huidige_studenten/overig/OvO/OvO-inf/Eindverslag%20INF.pdf (see page 8 for the German criteria, and page 9 for the results). See the original research (*) here: http://subs.emis.de/LNI/Proceedings/Proceedings22/GI-Proceedings.22-12.pdf (German language document)
(*): [LH02] I. Linkweiler, L. Humbert. Ergebnisse der Untersuchung zur Eignung einer Programmiersprache fÂur die schnelle Softwareentwicklung â" kann der Informatikunterricht davon
profitieren?, Didaktik der Informatik, UniversitÃt Dortmund, 2002.
Therefore, by the (faulty) logic you're using, you're just a cow with a keyboard - osu-neko (2604)
in fact Lisp itself is built in C
Errr...no. Lisp originally dates back to the late 1950s; C didn't emerge until the early 1970s. The first working Lisp implementation was writtien in IBM 704 machine language; A Lisp compiler (itself implemented in Lisp) was implemented in 1962 - fully 10 years before the birth of C.
Not everything that can be measured matters; Not everything that matters can be measured.
The point of IntroTo Programming courses is to instill a comprehension and sense of awe at the ability to control the actions, operations, and functions of physical machinery by using symbols, that are non-physical. It is by using your brain to amplify your body (robotics) or by using your brain to build and control a machine that can vastly amplify your brain (a pocket calculator).
Intro To Programming needs to skip language and process at the beginning and first teach how electricity can be used to create and manipulate symbols. This is multi-stage process that teaches how to use electricity to represent binary numbers, then using binary numbers to represent decimal numbers, using decimal to represent CPU instructions, using instructions to make programs, and using programs to control machines that amplify the users physical and mental abilities. And finally, how to use the imagination to create new structures of symbols to create programs.
Install the sense of god-like awe at the ability to manipulate physical reality by rearranging symbols, and the mundane details of language structure are of minor importance to both the student and the teacher.
Programming in Scratch helps kids
During the directed learning that takes place in a Scratch-oriented curriculum, the teaching team can introduce another programming language to show how syntax-oriented programming languages can perform the same tasks as the graphics-oriented systems. Any programming language can serve as that second language.
I find it a bit ironic that the best language for teaching programming languages isn't a language at all.
I agree, anyone who says that learning language X will screw up a person for lifetime is an idiot. The only people who would be "screwed up" are people who aren't good at programming in the first place. And on another note, Java and C# are pretty damn good languages, and even VB.net is supposedly pretty good if you actually give it a chance. I haven't tried it myself, but it's orders of magnitude better than VB6, which is what most people think of when they diss VB.
All your base are belong to Wii.
I agree that Visual Basic is as bad a choice for a first language as any other complex programming platform.
What made old skool BASIC good was that it was limited in ability. Admitted data structures were limited to arrays, which was a problem. However a medium-complexity basic like Blitz Basic 2 on the Amiga allowed the creative side to be expressed, without having to wade through complex APIs like you would with a modern language.
And the best way to learn programming to a young person (under 16) is to allow their ideas to be expressed and implemented, be that writing your first football league tracking application, to a simple game, to a text adventure, and so on. If that means using BASIC, e.g., RealBasic, then so be it. It needs to be pick-up-able.
I bet there are people saying Haskell and ML on this thread, for some academic reasons. The last thing a young person wants to be doing is learning how to manipulate data structures, functionally, with all the brain-fuckery that involves, and only to get a sorted list at the end. That isn't exciting, it's not even something to be slogged through, it's tedious and will actually put them off, totally.
10 Print "I am god!" : goto 10
run
instant result.
It's sad that computer magazines don't have programming in them any more, unlike the 80s. Game type-ins promised rewards to typing, and learning was osmotic.
Most of the people posting here really haven't grasped (1) and (2).
Assembly as a first language is ridiculous, yet so many are arguing for it.
Not only is it irrelevant today apart from microcontrollers, which they might get a job programming in 10-15 years time (assuming they're young now), but it will be incredibly frustrating.
The student has to come to the decision to use C, Assembler, etc, themselves, when they decide they have to in order to realise their vision for whatever they're programming. I.e., they're not beginners any more.
In the 80s that would mean writing your game in BASIC, and then finding that performance sucked. You'd extend BASIC with some extensions you found in a magazine or bought (e.g., Sprites Alive! on the Amstrad CPC) to see if that would help, and it would for the first few projects.
What's the modern equivalent? They've already done their MySpace profile, and so on. Maybe they've done basic HTML and CSS by themselves. Now they want to go further ... or write games online? So is Flash the answer? Java for applets, despite the massive initial learning barrier (shared by most of the languages being recommended in this story)? Or do you scrape them away from the web and onto RealBasic, especially if they look at PHP, ugh.
I would add in (2a) ability to get quick results, however primitive they are. 10 INPUT "What's your name? "; a$: Print a$ + " is a fag!" : REM: Instant fun.
I agree with (4) but it shouldn't be overriding.
Also many young people will never get turned on by programming, whatever you do. Maybe you should get them into repairing cars, or building things, or whatever. Or disown them.
Start them with Objects... I had a hard time getting into OO programming because I started with a very low level language.
I started on Fortran. It was horrible. Then I got a home computer with BASIC and advanced to assembly language.
25 years later, I still am at the assembly language stage for programming. But I use different processors and tools. The language is not as important as the tools that support that language. Visual BASIC is great because it gives a simple easy-to-use way to create programs in a Windows environment. Its structural limitations are irrelevant. It is the cost and sophistication of the development tools that is more important.
Now that you can buy a microcontroller for $1.50 that has more power than the original IBM PC, development tools like IDEs are the most important consideration. Computer science was important when computers cost a million dollars: it is meaningless today when they cost a few dollars.
I detest C because I can't debug it on the IDE that controls my $1.50 microcontroller. I can read it and write it fine. I can work with it fine. But I hate it because it's too abstract. I have no idea of what exactly the CPU is doing.
OOP is just science fiction; it's advantages are imaginary. If your application is so advanced and complex that you need OOP to create a program to do it, then it's time to completely rethink the idea of what a computer does.
Computer science is the process of reconfiguring complex concepts to fit into the limitations of the machine. Computer science becomes irreverent when you realize that the more complex the problem, the easier it is to solve by redesigning the computer to fit the problem. Not reducing the problem into small enough processes that will fit into the machine.
It is cheaper and faster to design a custom arrangement of 1000 $1.50 microprocessors to match the needs of complex problem than it is to write and debug the software that will 'solve' the problem on a $5000 standard Von Neumann computer. Microcontroller programmers are cheap and easy to recruit: OOP software development teams are expensive.
This is the new reality of the 21st-century. OOP is the last gasp of the 'big iron' boys.
a good programmer should know all the chain, from Java/Python/Scheme/Whatever down to the machine code.
Yes, but you don't start with assembly language. You start with something conceptually simple, like Python. I started with Basic on the Commodore 64. Before a year was up, I was doing shit in 6502 assembly because interpreted Basic was too slow. Not a chance in hell I could've picked up assembly straight off without some understanding of a higher level language. Throwing assembly at someone is like throwing a pile of parts and fasteners at someone and telling them to build a combine harvester.
If a job's not worth doing, it's not worth doing right.
Personally I would recommend C++ or Fortran since that should quickly kill their interest in programming. And I really don't want more competition from bright young people.
They are new programmers.
If they want to understand the logic of computers, then they should write pseudo code first.
Do this until they are not new programmers.