I am really curious if this actually makes things better..Maybe I am just getting old..
But this seems like everything else. Let's dumb it down so people can get it. But if you really want to be a programmer, you will get it. You will spend hours debugging and feeling like you know nothing. That is what makes you grow.
On the other hand....I ran into some Senior CS students who were incapable of programming in javascript...so...maybe it is not so bad.
I read this more as a nod to the fact that Stanford is the feeder school for SV. And pretty much all SV churns out these days is webapps.
Javascript is just a very mediocre language in virtually every way. There is only one thing it's good at, webapps, and it's only good at that by virtue of the fact that it's the only game in town for complicated front-end stuff.
The only thing that could be said in Javascript's favor as a teaching language is that it has basically no standard library, so teaching the entirety of the language can be done quickly.
Learning to program is hard enough without having to worry about hoisting, truthiness, function vs block scope, prototypes, the many rules of this, "classes" that don't behave like classes, and other quirks unique to JS.
Do you need to know all of the corner cases in order to write a simple script? No. But good luck trying to get someone to debug a JS program without a solid foundation. JS is a pedagogical disaster.
Prototype based inheritance vs class based inheritance will confuse exactly zero new people, since they will have zero prior experience with class based inheritance. Heck, if the goal is to teach algorithmic thinking and imperative programming (Stanford classes are 10 weeks long) then different forms of OOP don't matter at all.
Hoisting is weird, but is one of those things that can be worked around by teaching proper patterns and is no more mysterious to new people than whitespace errors in python or the magic word "static" in java or operator overloading in C++ or nightmarish type error messages in haskell.
They could have chosen a thousand better languages for starters like Python or Ruby to mention some useful ones or Scheme for example which has literally no syntax. They chose an abomination instead. Now we'll have even more influx of javascript-only coder hussars who have never seen an ecosystem which is not rotten to the core. Congratulations.
I went to a completely unranked school, not even a CS major. But took Programming I anyway. It was for freshmen, and was in C++.
IMO, for an intro language, you need either something close to the metal, so that students can work their way up the abstraction chain, or something very high-level with flexible constructs if it's desired to teach things from the algorithmic level.
I also think that any dynamically typed language is a bad choice for a first language. We're all used to holding the types in our heads, but beginning programmers don't even know what they are, and it's a really core construct.
I agree. When I first started out I learned java on my own. But when I first started taking college level CS classes it was all C++. And to be honest I prefer it.
If you have 1M bad programmers 1% will make terrible libraries. In 10-20 years at least a few of those you'll be forced to port to some other shitty library made by that 1%. I'm doing this right now...
Disagree, I think it is the exact opposite. I say that as someone who did their CS education in the early 90s when that was the case. We used C for everything unless the class material implied something else. We got into the down and dirty details quite fast.
IMO that's the wrong way to do it. Using myself as an example it took me far longer to learn things like data structures than it should have. Why? Because I still hadn't wrapped my head around pointers. The fact that the two are coupled is an implementation detail IMO. It turned out that I never had a problem understanding the structures, I had trouble writing code which didn't segfault.
I'd prefer to see CS education start with the more theoretical constructs and then start diving towards the metal as the basic concepts are known. In other words, treat it more like computer science and less like computer engineering.
I keep hearing that people have (or had) trouble understanding pointers. As someone who has never had that problem or met such a person, I am struggling to understand why...
In retrospect I don't understand either. It's such a simple concept. But it really befuddled me for a while.
My suspicion is that since everything involved pointers in C, even things like string manipulation, it was just an extra layer of confusion in a sea of new to me stuff.
It’s possible to use pointers successfully without understanding them fully - (having a sense of) incomplete knowledge can frustrate the learning process. Spending time dropping down to assembler to learn a few variations of MOV or looking at compiler output would make pointers achingly clear.
It's not so much as trouble with pointers as trouble with visualizing algorithms and data structures. Try implementing some moderately complex one in C. It reads like a monster. A dynamic language like Python would have a simpler implementation but at the cost of performance (Why do beginners need performance).
> IMO that's the wrong way to do it. Using myself as an example it took me far longer to learn things like data structures than it should have. Why? Because I still hadn't wrapped my head around pointers.
How do you expect to understand anything about data structures if you fail to wrap your head around one of the most basic data structure mechanism around: referencing a memory address?
In fact, how exactly do you expect to implement a data structure if you fail to understand the very basics of accessing and managing memory?
I don't see how that's a problem caused by a particular pick of programming language. In fact, that seems to be a major failure to understands the very fundamentals of CS.
Because a memory reference isn't necessary to be completely understood to understand the concept? It's just a low level implantation detail. I can draw a structure on a whiteboard and indicate that element a points to b. That doesn't need to imply a "pointer". I'm. It saying avoid the topic altogether in saying one doesn't need to understand memory management to do structures
If anything learning things this way would likely make learning about memory references easier?
I'm sure I'm biased, but I'm a huge fan of how they taught our first two semesters where I went to school. Semester 1 freshman year was imperative programming in C (learning pointers, bits, the stack and heap, memory, ...). Semester 2 was functional programming in SML (learning recursion, types, function composition, formal reasoning, ...).
When it comes to other programming language features, there are few that can't be understood as some combination of imperative and functional features. Introducing them freshman year, students take these lessons with them to every subsequent language they learn.
I teach at Stanford though not in CS. I think this headline is a bit off. There is a CS 106J class that follows the CS 106A curriculum using JS instead of Java, but as far as I know, the CS 106A Java curriculum has not been tossed out for the academic calendar of 2017-2018 (syllabuses aren't up yet, so I can't confirm).
As the article states, CS 106J was spearheaded by Professor Roberts this past year. I would imagine it'd take some more time for it to become a curriculum that all of the profs teach, though JS is being used in several other CS classes. For example, CS 142: Web Applications, used to be taught with Rails:
I've TA'd all of Stanford's intro programming classes, including the first course in programming CS 106A, so a few things:
- unlike a lot of other schools, CS 106A is mostly taken by non-majors (90-something percent of undergrads take at least one CS course before graduation, and for non-majors that means 106A)
- the follow up course, 106B, is in C++ and focuses on OOP, pointers, memory, data structures, etc.
- 106J was crafted as a replacement for 106A by in JS. It had limited enrollment, and word on the street is 106J's first offering this past spring was an experiment and it didn't go all that well, so who's to say if it'll still be around next year.
Maybe they should switch from Java to something hipster and trendy, but there's about 10+ years worth of assignments that will need to be rewritten, including autograding infrastructure for the 600-800 people who take it every quarter.
I personally feel like JS is an awful language with few opinions or guard rails, and thus a bad language to teach people who are looking for programming principles. It works for the web but I think for a course almost entirely consisting of non-majors it's not a good first language.
Thanks for the insight. With your experience, what would be a good language for non-majors so that they can make use of their experience later? Python, Go, C?
Python would be nice if everybody uses Python 3. Simple dynamic and most of the Open source stuff is scriptable using python. So daily automation is a real possibility. Atleast that's what I am doing.
I would definitely say Haskell. Second is python, but i feel Haskell teaches concepts better and more efficiently. Everybody has used spreadsheets before, and I think most people can draw a m mental link to spreadsheets with Haskell.
That's completely out of the question at the moment. True, it could potentially be most useful to know haskell for a lot of programmers, but at the moment, the opportunities to use haskell in the work place are actually rare and not to a small part because an understanding of haskell is barely achievable in an introductory semester, while the steep learning-curve of the language is kind of intended as it's usefulness only really plays out at a higher complexity. Whereas, python has a focus on scripts, shortness and quickness are even symbolized in it's name. Script languages with introspection and reflection are best suited for REPL, especially with good integrated documentation.
Javascript in the browser suffers at that because it's opinionated towards web, but OP sees that as a plus and web programming certainly is an applicable skill with a future. Not everyone would have to use facebook as homepage, for example
Haskell used to be the introductory language at UT Austin a few years ago IIRC. So it's been done before.
The point of an introductory programming class is not to teach a language for vocational purposes, but to teach the basics of computation. Someone who understands computation can reasonably be expected to pick up most languages.
It's funnny how, when java first came out, everyone was writing applets and not doing a good job of it (incl myself). This was due to it being way easier to code and deploy as opposed to C/C++ at the time. This gave ppl the impression that java was slow, when in fact the code was terrible (and ms was constantly submarining sun).
Now I'm sitting here with 16GB of really terrible Javascript code running in my browser using 5% constantly of the fastest intel chip you can buy today. No one says it's slow probably because no one knows it's running.
If chrome and ff put a task in the process list with all the Javascript CPU usage I'm guessing things would be way different.
I really do think K&R C should be taught to university level students. It's important to understand pointers and mallocs, even if you don't use it often.
You assume that the course is for CS majors. For students who'll only have 1-2 CS courses in their life, teaching some practical skills could be the better choice.
Every language with loose typing has some edge cases.
Once students know that JS has type coercion, and how to avoid this entire class of problems, they never need to know or think about what "0" == false evaluates to.
Or Haskell if you want to go the FP route while CL or Clojure for the Lisp. Or python for dynamic (atleast more popular than Ruby which is more concentrated on Rails ecosystem).
I don't like Python, but I have to admit that it ticks most of the boxes of an ideal starter programming language, unlike CL, Clojure, other Lisps, Ruby, or C/C++.
I'd say Python would've been the 'safe' and 'sane' choice, but maybe it's just my bias against Java and JavaScript talking.
Just curious, has anyone ever heard off choosing to teach something like bash / PowerShell? It might seem less intimidating than downloading a lot of stuff and students can play around with stuff already in their computer. Hell, maybe even something like VBA (as much we might dislike it) since non-majors will likely encounter it in MS Office anyway.
In the department I work, 5 people recently started using Python who never wrote a line of code before (except some limited VBA).
While the coding itself was relatively smooth, we spent a majority of our time consolidation which Python version works with which packages and how to create code that can be executed by computers that use Python 2 vs Python 3. Some people had to use packages that only worked with P2, others only worked with P3.
There's still not really a solution for this. We'll probably just switch to Julia once it gets enough support in areas we need..
So I disagree, Python is actually not a good language for beginners. The language is great but the ecosystem can be extremely frustrating.
The mistake here is not just making a choice but instead having people pick at random. Hardly pythons fault, every programming language has multiple incompatible versions.
Could you elaborate more about why you don't like JavaScript? Asking out of curiosity -- I just started learning it yesterday, although most of my experience is with C/C++/Python/Java.
49 comments
[ 2.1 ms ] story [ 95.5 ms ] threadBut this seems like everything else. Let's dumb it down so people can get it. But if you really want to be a programmer, you will get it. You will spend hours debugging and feeling like you know nothing. That is what makes you grow.
On the other hand....I ran into some Senior CS students who were incapable of programming in javascript...so...maybe it is not so bad.
Javascript is just a very mediocre language in virtually every way. There is only one thing it's good at, webapps, and it's only good at that by virtue of the fact that it's the only game in town for complicated front-end stuff.
The only thing that could be said in Javascript's favor as a teaching language is that it has basically no standard library, so teaching the entirety of the language can be done quickly.
And unlike C it is easy to get started programming in it. Thus a good choice for an intro language if you ask me.
Do you need to know all of the corner cases in order to write a simple script? No. But good luck trying to get someone to debug a JS program without a solid foundation. JS is a pedagogical disaster.
Prototype based inheritance vs class based inheritance will confuse exactly zero new people, since they will have zero prior experience with class based inheritance. Heck, if the goal is to teach algorithmic thinking and imperative programming (Stanford classes are 10 weeks long) then different forms of OOP don't matter at all.
Hoisting is weird, but is one of those things that can be worked around by teaching proper patterns and is no more mysterious to new people than whitespace errors in python or the magic word "static" in java or operator overloading in C++ or nightmarish type error messages in haskell.
IMO, for an intro language, you need either something close to the metal, so that students can work their way up the abstraction chain, or something very high-level with flexible constructs if it's desired to teach things from the algorithmic level.
I also think that any dynamically typed language is a bad choice for a first language. We're all used to holding the types in our heads, but beginning programmers don't even know what they are, and it's a really core construct.
Agreed. They aren't just learning to program but they are learning how computers work so the less abstraction the better.
On the plus side, I imagine this will be great for our salaries in 10-20 years.
IMO that's the wrong way to do it. Using myself as an example it took me far longer to learn things like data structures than it should have. Why? Because I still hadn't wrapped my head around pointers. The fact that the two are coupled is an implementation detail IMO. It turned out that I never had a problem understanding the structures, I had trouble writing code which didn't segfault.
I'd prefer to see CS education start with the more theoretical constructs and then start diving towards the metal as the basic concepts are known. In other words, treat it more like computer science and less like computer engineering.
My suspicion is that since everything involved pointers in C, even things like string manipulation, it was just an extra layer of confusion in a sea of new to me stuff.
How do you expect to understand anything about data structures if you fail to wrap your head around one of the most basic data structure mechanism around: referencing a memory address?
In fact, how exactly do you expect to implement a data structure if you fail to understand the very basics of accessing and managing memory?
I don't see how that's a problem caused by a particular pick of programming language. In fact, that seems to be a major failure to understands the very fundamentals of CS.
If anything learning things this way would likely make learning about memory references easier?
When it comes to other programming language features, there are few that can't be understood as some combination of imperative and functional features. Introducing them freshman year, students take these lessons with them to every subsequent language they learn.
As the article states, CS 106J was spearheaded by Professor Roberts this past year. I would imagine it'd take some more time for it to become a curriculum that all of the profs teach, though JS is being used in several other CS classes. For example, CS 142: Web Applications, used to be taught with Rails:
https://web.stanford.edu/~ouster/cgi-bin/cs142-winter14/inde...
Now it's taught with MEAN (Mongo, Express, Angular, and Node): http://web.stanford.edu/class/cs142/info.html
- unlike a lot of other schools, CS 106A is mostly taken by non-majors (90-something percent of undergrads take at least one CS course before graduation, and for non-majors that means 106A)
- the follow up course, 106B, is in C++ and focuses on OOP, pointers, memory, data structures, etc.
- 106J was crafted as a replacement for 106A by in JS. It had limited enrollment, and word on the street is 106J's first offering this past spring was an experiment and it didn't go all that well, so who's to say if it'll still be around next year.
Maybe they should switch from Java to something hipster and trendy, but there's about 10+ years worth of assignments that will need to be rewritten, including autograding infrastructure for the 600-800 people who take it every quarter.
I personally feel like JS is an awful language with few opinions or guard rails, and thus a bad language to teach people who are looking for programming principles. It works for the web but I think for a course almost entirely consisting of non-majors it's not a good first language.
Javascript in the browser suffers at that because it's opinionated towards web, but OP sees that as a plus and web programming certainly is an applicable skill with a future. Not everyone would have to use facebook as homepage, for example
The point of an introductory programming class is not to teach a language for vocational purposes, but to teach the basics of computation. Someone who understands computation can reasonably be expected to pick up most languages.
Now I'm sitting here with 16GB of really terrible Javascript code running in my browser using 5% constantly of the fastest intel chip you can buy today. No one says it's slow probably because no one knows it's running.
If chrome and ff put a task in the process list with all the Javascript CPU usage I'm guessing things would be way different.
I really do think K&R C should be taught to university level students. It's important to understand pointers and mallocs, even if you don't use it often.
Students who start by learning Javascript won't learn solid principals.
Everything depends how the course is taught, not on the language used.
Usually this statements holds true, but not in this case. I mean, after all we are talking about language where `'0' == false` is `true`.
Once students know that JS has type coercion, and how to avoid this entire class of problems, they never need to know or think about what "0" == false evaluates to.
Just curious, has anyone ever heard off choosing to teach something like bash / PowerShell? It might seem less intimidating than downloading a lot of stuff and students can play around with stuff already in their computer. Hell, maybe even something like VBA (as much we might dislike it) since non-majors will likely encounter it in MS Office anyway.
While the coding itself was relatively smooth, we spent a majority of our time consolidation which Python version works with which packages and how to create code that can be executed by computers that use Python 2 vs Python 3. Some people had to use packages that only worked with P2, others only worked with P3.
There's still not really a solution for this. We'll probably just switch to Julia once it gets enough support in areas we need..
So I disagree, Python is actually not a good language for beginners. The language is great but the ecosystem can be extremely frustrating.