I'm torn about all of these types of apps. On one hand, I think it's great to invite people into the world of coding and doing so in a user friendly matter. No doubt Grasshopper and a handful of similar apps are well designed, but is it really the right way to go about things?
I remember vividly learning to code at age 11. I had an old computer, windows would constantly crash and the only thing I could access really was QBASIC.
Alongside a book with code snippets, I would simply write a line of code, hit run and see what happens. Then, go back to the code and 'go rogue' (meaning: change the numbers a bit and make the line 5pixels instead of 2 for example). This is where the magic happened, because I actually coded something. It wasn't part of an educational application, it was the real deal inside the real editor with a real output.
The learning always came from trial and error and in a way is very similar to how I still code things (when I'm not designing SaaS products for awesome B2B companies)
The book might be replaced with Google & Stackoverflow, but the principles are the same.
The magic of coding, I believe is in writing something that is real.
Great thoughts! I completely agree that learning to code is a lot about trial and error. One of key things that we decided to do with Grasshopper is make it possible for people to write real code, but on their phones. From the very beginning, it's possible for users to "go off the rails" and write code, run it, and see what happens. We didn't want to do fill-in-the-blank stuff, since there's only one right answer. We constrain the code editing environment to be relevant to the puzzle at hand, but that's mostly so we don't overwhelm beginners.
We also have a Playground (available in the left-hand menu) where users can just play around using the Grasshopper coding environment.
Would love to keep getting your feedback on how we can make everything more real and keep that magic. :)
I think it's an important signal that in every thread like this rather than say "wow, I wish I had something like grasshopper when I was a kid", we start waxing nostalgic about QBASIC. (I'm right there with you btw).
A lot of us grew up learning with tools, not toys (except you, logowriter, you're cool). Then we went and made the tools totally unapproachable. So now kids get the opposite approach: toys, not tools.
Teaching software like grasshopper is an experiment, and we're about to see the results. The first version of Scratch came out in 2003. Obviously this is super fuzzy, but for the sake of argument call 2003 the border between the QBASIC/getting a Sam's teach yourself C book at B&N era and the Scratch era. So a hypothetical 8 year old who started learning with scratch in 2003 is now 23 years old. The kids who grew up on this stuff are about to start showing up in adult life, and it'll be interesting to see how they turn out.
FYI, the first public release of Scratch was in 2007 [1], so that's really a more accurate border. I'm aware of this because I started using Scratch in 2008. I'm a computer science student in college now, and I've had a couple of internships. From my experience, Scratch was a great start. I learned a lot from it, and I really credit it with getting me into the field, although I likely would have irregardless of Scratch. Either way, I don't think that it necessarily will be bad for toys to replace tools.
That’s really encouraging to hear. Thanks for correcting the date, too.
Something I’ve been thinking about since posting: imagine yourself in a sort of Victorian/steampunk world that runs on gears etc. What would you give young people to start getting them literate in the machines?
I figure it would be a toy: something like a meccano set. You’d want composable, easily understood parts that are simple to connect, and interact in predictable ways.
What you wouldn’t do is drop the kids off at a half-working tool and die shop and tell them to go make some gears. Maybe that’s qbasic.
> What would you give young people to start getting them literate in the machines?
Erector set[1]. It's a toy, but not super obvious how the parts are supposed to work together. (unless now they give advanced instructions for constructing things..)
Yup, same thing as meccano. Now that I think about it I had an erector set growing up, beats me why meccano came to mind first. I missed out on lego robotics, but those look really cool too.
At 15 I sat with a book about QBASIC and got to a certain point, got stuck, and had nobody around me to ask for help and no idea what to do. I decided I wasn't smart enough for programming and stuck to basic html and CSS for years and years. I avoided CS in college. I turned out ok, but: I can code! I was wrong when I gave up, but I didn't know that because the resources weren't approachable enough for me. And maybe my determination and self-belief wasn't as strong back then. I wanted to be a game developer really badly growing up. And then I found a book that taught me I couldn't program.
If you get stuck following a book or a tutorial it's warranted to send an letter or e-mail to the author telling them where you got stuck. They will be happy to know where their book can be improved.
This is very true once you start producing content for people yourself (if you ever do), but it's probably something that almost no one reading a book considers doing.
A lot of us also grew up with tools/toys like The Games Factory, which got us excited and left us wanting more. And then we got into QBasic (which I sure do feel nostalgic about) and Borland C++ (which, on Windows 3.1, I really do not feel nostalgic about :P ) and MS Visual C++ (Visual Studio 6 was awesome!)
I'm more worried about the tablet/phone generation. This was when our machines fundamentally shifted to toys not tools. We lost the input methods necessary for tools and things like the file system are abstracted away. We'll have a generation of programmers that have never seen a command prompt and no concept of what a file is.
My nephews (10 & 12) are learning programming at school for which they both had to have an iPad for some shitty programming interface. They didn't learn anything about logic because the animated graphics were too distracting, they learned how to manipulate sliders to adjust things like speed and that's about it.
It would be interesting to know if anyone on HN in their early 20s or lower were first introduced to software development via scratch or some other means.
I'm 21, and have a fair bit of programming experience now (did some internships, have a job, know Python/C++ well and JavaScript/Go/Ruby/Lua reasonably well) and I learned to program using Scratch.
I had a lot of fun making games with Scratch, and I think it taught me a lot about logic and gave me a great way to express myself. I definitely got way more into the weeds about it than everyone around me, but one of my favorite things was opening up a game I admired on their website and trying to figure out how it works. This skill has definitely transferred into my broader programming career.
I have really high praise for it! The only thing looking back I wish we had was a way to export it to a .exe self-contained runtime to share with friends. I just used to pull my friends over at lunch.
I'm 24, so maybe a little old, but I was first introduce to programming with Game Maker. Later I did a little bit of Flash if you count that, and then moved onto Visual Basic.NET when I did 'Software Development' at high school.
For a long time (since mid-primary school) I had been toying with HTML, mainly from a book from the local library, and by just looking at the source code of websites that interested me.
I think it's clear that 'toys' like this, and like Game Maker, that make programming simple and user-friendly have been around for a long time, and don't necessarily stop people from learning more advanced concepts or languages, except if they lose interest. But even if they do, hoefully they will have packed up some skills that they can apply to other endeavours.
For me (26), part of what kept me interested in learning to program was the sense that I was working through something convoluted and solving some personal problem. Overcoming some sort of adversity. In this case, it was writing hodgepodge pascal to automate Runescape and Neopets, then also Flash for fun and profit.
I now have friends that are interested in the concept of programming, but have no application for it. I'm happy to aid in their journey, but when they ask me "What should I program?" I have no idea what to tell them. "Pick something arbitrary" I say, only to hear "What's that mean?". I feel like programming is something most people either come to organically or don't follow though with, regardless of challenge involved.
If we are to think about what would make a successful programming class, would it look more like math such that "Today we'll learn math, because math will totes be useful later for some reason" or would it look like "Make a computer do something interesting or useful with minimal human interaction".
+1 for Game Maker, what an awesome educational tool. You could start with building blocks, Scratch style, then move over to code when you needed a little more power -- with a 1-1 relationship between lines of code and the blocks (IIRC), and the ability to run that code as "just another block". A wonderful gateway to the real stuff.
I'm 23. Unlike many people here, I started programming only after graduating from high school and before going to college.
I always thought programming was cool but was never able to learn it (because of I bought some bad intro books to programming). Then one day I found out about the book C++ Primer Plus, it was wonderful :)
My first attempt at a real program was a console based repository management program for my dad's apparel company. Considering my dad didn't even know how to send email, he probably tried very hard to give me a good feedback. I didn't even know the existence of database, and just stored the data in binary files (I thought it was secure that way)
Funny thing is that after 2 years in college, I learned some web, database related stuff and developed the second version using Tomcat, Spring and MySQL. My dad was again forced to use it :)
23. First programming experience was with a C++ book my dad gave me when I was 10. Didn't learn too much from that. When I was around 12 in school we used Logo (MicroWorlds) in technology class. I had a lot of fun with that. I finally fully got into programming with C# when Microsoft XNA came out, was around 14 at the time.
At my school, we teach Scratch in 5th grade (US, age 10) and I teach Python in 6th grade. I think Scratch and the like are good starts, so I'm glad to see more offerings. Even so, stepping up to Python is a big step from the blocks languages.
Last week I started the 6th graders on MicroPython on the BBC micro:bit. So far it is going well, and I think it will capture some of the aspects of the QBASIC on DOS experience. Today I will show them how to do images on the 5x5 LED matrix, which will be much simpler than trying to do graphics on Windows. Even so, they will still learn about X-Y coordinates and pixel brightness levels.
Qbasic was never quite the professionals' system like C and Assembly was, but looking back, I was at least using the same system that non-technical managers who regretted firing the developers were clumsily banging out solutions with.
> Alongside a book with code snippets, I would simply write a line of code, hit run and see what happens. Then, go back to the code and 'go rogue' (meaning: change the numbers a bit and make the line 5pixels instead of 2 for example).
Grasshopper does pretty much this, but with a nicer UI and adapted to the constraints of a tiny phone screen. It gives you a code window and a graphics window and you get to write code and see what the code does. It's is a JavaScript environment which is modern-day web-world filler of BASIC's niche.
It is worth pointing out that your style of learning might be different than others', and gender has been found to be one such difference [1]. There are quite a few studies that have found tinkering to be done far more often by males, but is not necessarily more effective.
The myth of the "myth of learning styles" is also discussed there. The 'debunked' "learning styles" was a very specific hypothesis about audio vs visual vs tactile modality, not a disproof that there are any differences in how people learn.
Actually the core of what I was trying to say is not about learning styles at all. You can use a real-world editor and a completely different learning style to create a real world application (by following along tutorials for example). The big difference, and that was the point I was trying to make, is that in a real code editor, you're learning while building in a real-world environment. Even if you stop in the middle of your journey, you will have built something that you can share with friends. Whereas these types of apps will take you through a journey which is great, but I wonder how many take the step of going to their computers, opening an editor and building their first real application.
How far does it make sense to take this? Take it too far, and all tutorials are bad.
IMHO there needs to be a balance between guided learning and self-directed exploration.
Too little guidance and you can't get your bearings, causing you to get stuck and unable to cross what should be relatively minor hurdles.
Too much guidance (or too little interest exploration) and you end up not really understanding why things are the way they are or what is possible beyond the very narrow range of things you've been taught.
Exactly. And the thing with old computers is they were kind of limited in scope because of hardware limitations. That in itself kind of gave you an idea of where to start because there wasn't much else on there. OP even mentions basically being stuck with QBasic. Now there's an endless field of where you could possibly start at. It's so useful to have some guidance of where to start.
Imo it's also much faster. Yes I can trial and error figure out a way to fix my car, but why would I not use a manual?
My experience with coding was using the old BASIC books (like the ones at https://www.atariarchives.org/ ), and stubbornly typing my way to make these games work.
When they didn't, I found it enormously satisfying to find out how to work around the problem being expressed in print; there were enough variations in BASICs for different machines that the listing for a TI 99/4A program wouldn't work in straight AppleSoft BASIC, for example.
But the time between typing all the code and finding out if it worked was grueling. I'm not sure if the students of today would be satisfied with that large a report-back loop.
The magic of coding, especially early on, is in realizing that you, a user, can deterministically control a computer (like you stated in going rogue). This has nothing to do with writing something that is real, it has to do with cause, effect, action, reaction.
I also started with Qbasic. I learned from the examples in the on-line help system. The first one I vividly recall running and changing was the CIRCLE example; probably the first code I ever wrote.
I tend to agree with Bret Victor on this one (http://worrydream.com/LearnableProgramming): changing numbers randomly is a terrible way of learning programming. We can, and and should try and do better. The microwave analogy in the 'Read the Vocabulary' sections is very apt.
The app does! Or at least, instead starting off with a `drawRectangle` function that takes a full set of x-y coordinates and 3 numbers for RGB for fill color, it starts off with "drawBox(red)" (remember, this is an introduction to programming) and builds from there.
Check out the user video someone else linked on this page!
We're super excited to be announcing Grasshopper today! We're hoping that it'll make learning to code accessible to more adults that are interested in applying coding skills to their careers, hobbies, or just for fun.
Please let us know what you think, ask questions, all that jazz. We're still in the process of learning and making changes, so all feedback is super helpful.
If you try to install the app, does it actually list those permissions? Or are they just listed on the Play Store listing?
We think there's a bug that's hitting some folks on Play Store that's related to an old manifest from our Beta 9 months ago... but we don't require those permissions anymore. Or at least we shouldn't. If you do see something when you actually try to download it, could you reach out to grasshopper-support@google.com? We're trying to find instances of this occurring and trying to convince the Play Store team it needs to be fixed :-/
Thanks for verifying. Just got off the phone with the Play Store support folks. Apparently it's the Play Store policy on play.google.com to list every permission you've ever asked for forever. We pushed a build many moons ago while we were in beta that asked for permissions by accident, but fixed it very quickly. We no longer ask for those permissions.
This isn't a great experience for us or other developers, so I'm escalating internally. Thanks again for the report.
Good question! We're building Grasshopper while we're within Google's Area 120, so the easiest way for us to ensure Google's privacy best practices are followed is if we use Google Authentication.
It might be possible for us to convince the powers-that-be to use another form of authentication, but for right now, we're sticking with just Google Auth.
We've thought about putting the signup after the tutorial, just haven't had the time to get to it just yet. I'll add this as a +1 to us starting to work on it :)
Not trying to side-step, just trying not to get into the weeds too much, but since you asked, reposting a bit of my response from another part of this thread:
We need a user account to store your progress: what puzzles you've completed, what code you've entered in order to save progress if you didn't complete the puzzle yet, coding streaks, etc. We're not collecting information about your device (or any other personally identifiable information other than the Google Account we associate with your account).
It's been easiest for us to do this user account creation using out-of-the-box Google Authentication, so that's why we went down this route. Hope that makes sense!
I get the ease of use for login/authentication. However, one thing that I'm running into is that I have a child user account setup on my tablet so my kids can use it without having an email. I tried to get my kid to use this app this morning but this account doesn't allow authentication so I just get an error: "there are no accounts to select". If I had an alternate way to sign-in I could get my kid involved in coding.
That makes sense, but is Grasshopper really the kind of app you want to sync across devices? I sometimes work on mobile games, and we always have the issue of whether we should force people to login (and give us data) or play without the login. We obviously want the login to improve our chance to make money. But sometimes logging in causes users to quit the app. But we can let them play without logging in, and sometimes games do that to get players hooked and nag them to surrender some privacy later. But that is why we do it: to make money.
I personally wouldn't try this because it wants me to login. I don't care if it's Google authentication or Facebook or Yahoo or anything else. I don't see any reason why I should link any of my online accounts with this. Since this is an app, you could save all progress information on the device and if required, collect any analytics right from the device whenever it runs (if you do need usage data).
Please remove the need to create an account or login anywhere. Or make it optional so that those who do want to login, can. I would prefer the default be to not create/login to any account and an optional login flow for others.
What login system would you prefer? We're currently doing Google Authentication because it's the best way for us to uphold Google privacy best practices, but can chat with folks about what other types of authentication might be possible.
Why does it need a user account at all? You say it's to uphold privacy best practices, but I don't know what data it's collecting from my phone. I would hope none. But if no data is collected, I don't see why any authentication is required.
We need a user account to store your progress: what puzzles you've completed, what code you've entered in order to save progress if you didn't complete the puzzle yet, coding streaks, etc. We're not collecting information about your device.
It's been easiest for us to do this user account creation using out-of-the-box Google Authentication, so that's why we went down this route. Hope that makes sense!
So to sum up : you’re not collecting information, you just want the user to be able to retrieve what he has already done.
First, you should ask for an account creation only when the user have already started. As I said, I didn’t even see if it was interesting at all, because of the login wall.
Second, if there are no sensitive information collected, there is no need of using a mountain to hit the nail. A pseudo will be enough. No email needed. Not sure if it even need a password...
You store the pseudo locally and you are done.
(I agree I’m a little bit oversimplifying, but you see the point)
A username and a password. All this third party login nonsense needs to go. I just spent the better part of last week actively removing myself from all things google and I’ll be damned if I am going to throw that away over a coding app.
Off topic, but since your on the tram. Is there any way to allow a child's device using family link to access grasshopper? It would be great if on such a device one could add another account at least for the purpose of logging in to apps like this.
My teenage daughter showed me Grasshopper yesterday. She was very excited about her progress, showing off the puzzles that she had solved and the code she had written. She hates schoolwork, but she really loved the app. I think it's a fantastic way to approach that often daunting initial learning curve.
I tried looking for a video of someone using the app so I could get a more clear idea of the sorts of things it would teach. This is the first thing I came across: https://www.youtube.com/watch?v=lsA2bGdrkaQ
I found it interesting to hear the kinds of reactions the kid was having to it: if you skip around the video you'll see the most common reactions are jumping around between distress, frustration, and anger.
I don't think that's because this is a bad app.
Instead, I think that's probably a natural part of learning to program—and of programming even once you've learned. It's kind of a strange thing that something which can bet net enjoyable (oftentimes) can also so typically elicit those kinds of reactions (of course adults are better at suppressing them, but the impact must still be there).
You would think that someone at Google could use their own search tools to look up a name for a coding tool before using one that's already been in existence for years. http://www.grasshopper3d.com/
You're assuming they care. Google was aware of a startup called Inbox that did stuff with email, yet they went ahead with their own app called inbox, steamrolling the startup.
I have no reason to suspect the same thing isn't happening here.
Thanks so much for the feedback (both here and in the app). That's something we're looking into, and I totally understand it makes it hard for kids to use the app. Grasshopper is an Area 120 project (a part of Google) which is why that account system is currently used for syncing course progress and achievements between devices.
btw is there a login system you'd prefer? I'd love to forward any thoughts you have to the Area 120 leadership.
How about an anonymous login system that doesn’t require an email address and just generated a random userid / password? I’m thinking of setting up a Code Club at a local school and it would be great if kids could use it without having to provide any personal information whatsoever.
Seeing the response on HN to these kinds of tools/toys makes me wonder if they're marketed incorrectly.
These things aren't really teaching "how to code" (what we associate with stuff like Java, Python, etc.), but "how to think like a computer" or "how to break down and solve problems." I wonder if there would be an equally pessimistic reaction if the headline and marketing material said that instead.
I think a lot of being a successful developer comes from being able to think a certain way, and I think that these types of games/toys can definitely help children learn to think that way. I remember the Dr. Brain games giving me some of my first glimpses into this kind of literal, step-by-step thinking.
I'm now tired because I played with it instead of going to sleep :(. I must say, it's well-executed and feels lightweight (which is a surprise both considering applications of this kind, and considering applications coming from Google).
I must say I love the attention to details! I was impressed by the effort that seems to have went to design the error detection - I start to mess with an example, and the messages tell me exactly what I'm missing at the right level of abstraction. Like, I forgot to to the next line when drawing blocks, and the popup explains I should use newLine(). Or, in another case, the popup telling me that yay, I have row 1 and 3 of the solution correct, and need to work on row 2. Etc.
Also the little details made me smile - like being able to replay the "jumping grasshopper" animation on success screen by tapping the grasshopper.
The keyboard model used in Grasshopper is something I first seen in Hacked (http://www.hackedapp.com/), and it seems to be catching on, which is great! Now I really wish someone would make a keyboard for Lisp like that; I could do some work while standing in a bus...
So here's the question: does this product reflect any of the research into teaching programming or is it just "somebody had an idea"? There is nothing on the website I could find that addresses this.
We know a reasonable amount about teaching programming, and in particular we know that traditional methods are really bad. Most people, when they teach programming, reproduce the poor traditional methods used to teach them.
There is already a large quantity of "learn to code" apps and sites. We don't need more. We need more quality.
Long ago I had an Apple Newton. The funnest thing for the Newton, by far, was a program that would evaluate statements or code snippets of NewtonScript on the device. It let you try little programs while you were waiting for something.
As far as I can tell, if a child is on a Google account with parental controls, it is impossible for them to log into the app. This kinda sucks for a nice educational app.
I would've recommended this to people I know that want to learn to code, but I wouldn't wish JavaScript to be my worst enemy's first coding language...
Really? I think JavaScript is the perfect language to learn with. Yes it has its quirks and weird edge case, but from someone with no programming background I think JavaScript is one of the easiest and most practical language to learn.
I think a lot of the "bad stuff" can be seen as an advantage. Like for example:
- Lack of strong typing
- User don't have to bother with the data type at first, they go directly into the logic
- Typecasting when comparing
- faster for quick comparison. E.g. if a user wants to check if the input of the user is equal to a certain number, he doesn't have to cast it as an int/float/whatever before, he just does "10" == 10 and it just works.
And also, its greatest strength I think is the fact that it's the language of the web, and web development has the lowest "barrier to entry", because all you need is a notepad and a browser and you can easily develop and test.
Anyway, as many people have pointed out, I don't think it matters that much what language you start with, because it makes you learn the logic of programming, which is then transferable to all other languages anyway.
No one is going to learn to be an effective programmer using a smartphone.
I do expect a lot of people to install it, use it for 5 minutes, and never bother with it again. It's a perfectly suitable way to push people out of programming because there's no way you can get a real feel for it with a smartphone. All of the 'learn to "code"' applications that don't involve writing actual software seem like a giant waste of time and effort to me.
Can they please just stop putting their fingers in every pie? I WANT Free coding lesson websites, but not from google thanks. They should stick to github with their contributions.
It's difficult to enjoy our lives, with these corporations strolling in like "IS THAT FUN WE SEE?!?!" (because they're giants)
I remember vividly learning to code at age 11. I had an old computer, windows would constantly crash and the only thing I could access really was QBASIC.
Alongside a book with code snippets, I would simply write a line of code, hit run and see what happens. Then, go back to the code and 'go rogue' (meaning: change the numbers a bit and make the line 5pixels instead of 2 for example). This is where the magic happened, because I actually coded something. It wasn't part of an educational application, it was the real deal inside the real editor with a real output.
The learning always came from trial and error and in a way is very similar to how I still code things (when I'm not designing SaaS products for awesome B2B companies)
The book might be replaced with Google & Stackoverflow, but the principles are the same.
The magic of coding, I believe is in writing something that is real.