Scratch Jr

I’ve spent a lot of time in the past few years talking to students and teachers about learning to code. I’ve also spent a lot of time observing classrooms where coding instruction is being given. I’ve had a lot of “aha moments” and been inspired by many things I have seen. In all of those experiences, the thing that really stands out is seeing Scratch, a free visual programming environment developed at the MIT Media Lab, being used effectively by all ages and abilities. I have come to believe that almost anyone can learn to use Scratch and thus start down the pathway of learning to code.

The Scratch project started in 2003 and since then over 5mm projects have been built and posted on the Scratch website. You can browse the projects here, you can find one you like, you can fork/copy it, and you can make something yourself. It is this forking/copying thing that is so powerful in my mind. You don’t need to start with a blank canvas in Scratch. You can find a project you like, you can look at the code to see how it was made, and you can then modify the code to change the way it works. That’s actually how I learned to code too (by initially modifying someone else’s code).

I just found this flappy birds style game that was posted to Scratch yesterday (use the space bar to flap, hit the flag to start).

If you click on the link that says “posted to Scratch” right above the game embed, and then click on the button that says “see inside”, you will see the code that was used to create this game. You will see how visual and inviting that code is relative to most coding systems.

Anyway, this post is not about Scratch. It is about Scratch Jr. The one thing that you need to be able to do to use Scratch is read and write. So kids who are still learning to read and write can’t use Scratch. A few years ago some researchers at Tufts started working with the team at the Media Lab to create a version of Scratch that younger kids can use. It is called Scratch Jr. Scratch Jr is getting close to commercial release, which will be on tablets (iOS first, Android next).

Since Scratch and Scratch Jr are free to use and supported mostly be research grants, there isn’t a lot of money to commercialize Scratch Jr, particularly the development of easy to use iOS and Android native apps for tablets.

That’s where all of you come in. Scratch Jr posted a Kickstarter project this week. I backed it yesterday. In a few days, they have already raised their initial target of $25k, but I am certain they could use a lot more. The more money they raise, the faster they can get iOS and Android out and the more they can do to get Scratch Jr into classrooms all over the world.

Here’s the Kickstarter page and here is the video they posted. It’s only four minutes and the bits where Mitch Resnick (the “father of Scratch”) talks about learning to code and coding to learn is really good. You should watch it. And please consider backing the project too.

#hacking education

Comments (Archived):

  1. aripap

    Love Scratch and just donated to Scratch Jr. The biggest problem I’ve got is that I want to play with it instead of letting my son take the reins. Here’s “Pong” that he and I built on a snowy weekend day:

    1. fredwilson

      i just played your pong game. that’s great. the thing the scratch embed needs is a link to the project page so you can see the code you and your son created.

      1. Robert Holtz

        Thanks for pointing out this Kickstarter, Fred. Of course I backed it immediately. Great cause. You know? One day, quite a few years ago, when my mother asked me about what code was and how it worked, I fired up Scratch. It was amazing how instantly it conveyed the basics of programming and logic. She grasped it immediately. So when @jasonpwright:disqus says Scratch Sr. is next, he may be onto something. But Scratch will always have a special place in my heart because what could have been a very boring and long-forgotten conversation became a treasured memory.

    2. ShanaC


  2. Ana Milicevic

    Love it – Scratch Jr & Scratch look like software Lego to me.

    1. fredwilson


    2. markslater

      perfect analogy. lego is analog – scratch is digital

    3. Robert Holtz

      It IS software Lego. 🙂

  3. jason wright

    Kids Kode!Kool.

  4. jason wright

    …and next Scratch Srthey’re out there waiting to learn

    1. Jim Canto

      Indeed they are. I’m NOT one yet…just to be clear ;-)I have an idea I’m cultivating within this realm. My goal is to encourage similar empowerment, without writing code. I envision this first level of empowerment as a “gateway drug” towards actually writing code (which I hope to have time to learn some day.)

    2. Robert Holtz

      Scratch Sr is actually a pretty great idea on your part.See my related comment to @fredwilson:disqus.I’d happily back that version too if they ever Kickstart it. 🙂

      1. jason wright

        it is the new literacy, and everyone should have the opportunity.

        1. Robert Holtz

          You’re quite right.

  5. CJ

    I’m on this, my son is four now and right in the sweet spot for this.

    1. fredwilson

      it will be out in the next 3-6 months on iOS i think. so perfect timing for you

      1. markslater

        Me too – stella is gonna be all over this – she’s 4.5

  6. awaldstein

    Collaboration and iteration are how kids learn generally in school with many projects in public schools built around Gdocs and sharing.Coding should be no different.I’ll support this one.

  7. William Mougayar

    Wow, this beats model trains and lego. I wished this existed when I was a kid. We won’t know the real impact of this for probably 5-15 years, but this has a generational type of impact.

    1. fredwilson

      well scratch has been around for a decade now. but mostly for the past five years

    2. markslater

      model trains and lego were amazing analog learning experiences…

      1. Emily Merkle

        they still are.

  8. ShanaC

    Has anyone tested an alpha version with Scratch Jr with kids yet? If it works, it probably is as big of a breakthrough for adults. The reading part of code can be the most difficult – and using big fun shapes may get a lot of adults over the hump too 🙂

    1. fredwilson

      yes, several hundred kids have been beta testing it, mostly in schools in Boston

    2. Chris Phenner

      I have a four-year-old who I’ll film using Scratch Jr. This clinches my having to upgrade to iOS7 (kidding).

  9. Twain Twain

    Always loved Scratch. There is also a project out of Carnegie Mellon that I love called Alice which teaches kids how to program in 3D:*

  10. Brad Dickason

    Hey Fred – I’m trying to visit both links to the kickstarter project from the browser in the twitter iOS app. Both links redirect me to a blank page. Not sure if the issue is on your end, twitter, iOS, or kickstarter but here’s the link for anyone else struggling with this:

    1. fredwilson

      that’s a problem. thanks for highlighting it. they work for me in my chrome browser on MacOS. I will check on tablet and phone in a bit.

  11. Paul Sanwald

    Fred, you might know this, but others might not.We use scratch (not jr) to teach computer science to high school kids as part of the TEALs program ( I found out about this program through avc, and volunteering for it has been really, really rewarding. Our kids are able to build games and truly have fun learning about programming in a really short amount of time, and scratch has been a huge part of this.My sister has young kids, and Scratch Jr looks like just the thing to get them programming. So, I’m excited about the whole ecosystem, and glad to see this post.

  12. iggyfanlo

    This is amazing.. truly making programming available to everyone… don’t want to overstate it but it’s like Gutenberg and bringing print to the masses…. this is killer

  13. Emily Merkle

    saw it touring the Lab a couple years ago. very cool.

  14. Brandon Burns

    Not to hijack today’s thread, but can we talk about the new Moto 360 (Google / Android watch)? Its fucking gorgeous! I’m absolutely floored at how perfect it is (or at least seems).Also, Android as a platform is really showing its strength here. There are only so many phone makers in the world, so the mobile OS war has different stakes there. But there are *many* watch makers, most not “tech” companies. To think we could have watches made by anyone from Rolex to the maker next door, and have one awesome OS to go with myriad options for brand and design, is nuts! Apple could never compete with that. Phones are one thing; everyone may be cool having the same one, but you definitely won get a critical mass of people all wanting the same watch. Once you get into fashion, i.e. wearables, the game changes.Considering Fred’s love of Android, I was kinda shocked this wasn’t today’s topic! Sorry for the hijacking. 🙂

    1. Dave W Baldwin

      …and there is the one thing they need to put in all of the phones that will make the watch a killer.

      1. Brandon Burns

        what is that one thing?

        1. Dave W Baldwin

          Do the Chromecast to television. I bet that’s what they’re up to due to the little fella is being marketed on television now and the ability to have House/Cards playing on the tube while you do something else with the tab/phone would be useful.Plus, Google is wanting to see how many points in your room they can place the Virtual Assistant.

        2. Dave W Baldwin

          Sorry, missed the write up and it referring its interaction with TV. The watch interacting/directing other devices will be big thing.

    2. LE

      Its fucking gorgeous! I’m absolutely floored at how perfect it is (or at least seems).So what do you see as the market potential for this? I’m not seeing it.

      1. Brandon Burns

        What is the market potential for this? Seriously?!For starters, anyone looking for a new watch.From there, who knows. What was the market potential for tablets before tablets existed? Many people said none. Those people were wrong.

        1. LE

          Seriously?!The main reasons for a watch is to tell time. Sure people buy them as jewelry fashion statement as well but by and large the main reason is to tell time. Similar to why people buy cars. Main reason is transportation. Secondary reason is as expression of oneself.So along the lines of time there is much less reason to own a watch when you have a smartphone, tablet or desktop computer with the time. Which are larger and can do many things.While it’s very possible that there will be functions that end up being the killer app on the watch that we have not thought of and it’s possible that it will be a short term fad and many will sell when they first come out (if priced correctly) I don’t think it a slam dunk by any means.There is not enough screen real estate on a watch to make it useful for reading (especially if you are over 40 when many need reading glasses_).I just watched the “we want to be like Steve and Jony” video that they did for the watch:…Marketing like that works for sure in some cases just to me it’s a turn off for this type of new product. Especially because it’s heavy on the design and jewelry factor as reasons for wanting the product. When Apple did this type of marketing (or Porsche) they are trying to take something that people already want and need and do a better iteration of it and create more desire. I’m not seeing it fit with watches. Wall street type is still going to buy super expensive get laid watch. Not enough guys in Portland, Brooklyn etc. to support this type of watch.That said I’m not a person who wears watches so this should be interesting.Do you know what the cost is going to be? (Not finding anything that tells me that).

          1. Brandon Burns

            Mobile phones have been around for two decades now. I’m not positive, but I don’t think the watch market has yet to be affected.When it comes to shopping, especially for things as useless as accessories and gadgets (watches fit both categories), the logic is simple: people buy things they like, are pretty, and/or they think are cool.Will people think the Moto 360 is cool? Will people think it’ll be worth the price (which I don’t have details on, either)? We’ll see!

          2. LE

            but I don’t think the watch market has yet to be affected.While I was not able to quickly uncover reliable data to prove my point (I could throw data of course but would rather preserve my, um, integrity by making sure) I was able to find this with the search of “do people still wear watches” that is interesting. (Attached graphic)Note some of the story titles. I say where there is smoke there is fire.Click to enlarge.

          3. Aaron Klein

            I’ll make you a bet right now that the wearable computing market with a small watch-sized display will be huge when it is properly executed.Moto 360 looks like it might be and I can’t wait to see Apple’s take.

          4. LE

            when it is properly executed.I’ll consider taking that bet. But you would have to define a time frame and “huge”. [1][2] This isn’t a sports event. It doesn’t easily match the pregnancy or death standard.I had someone want to make a bet after the market crashed (to 6k the last crash) that the market would go back to 14,000. I said “ok by when?”.[1] Imagine taking a bet years ago that “real estate in Brooklyn will be very expensive”. Obviously a time frame is necessary otherwise anything is possible.[2] In deal making there is a saying. You can name the price if I can name the terms.

          5. Aaron Klein

            I think I was being more rhetorical when I said “bet.”You said “the screen size is too small.” That part is not going to change.With well-executed software, a smart watch that is integrated with my phone, can alert me and let me see previews without pulling my phone out, can monitor sleep/activity and can be a silent alarm…that will be a huge hit.Could be in 2014 or 2015.

          6. Emily Merkle

            Chip implanted in the brain will do the same thing. And you won’t have to give up your Rolex.

          7. Aaron Klein

            What’s a Rolex?

          8. Emily Merkle

            Your Swatch, then?

          9. awaldstein

            Bingo–it’s not a watch at all.Will be curious to see whether the name stays and the function changes under it or whether it gets renamed to suit new uses.

        2. Emily Merkle

          It’s not a “watch”. It defies the definition of “watch”.

      2. Emily Merkle

        Chalk one up for slick advertising.

      3. SubstrateUndertow

        Continuous/transparent monitoring of AI empowered biometrics like- blood-glucose via non-invasive laser based sensors = Billions$- head end for insulin pumps- blood pressure / pulse / arrhythmia / pacemaker monitoring- cheap halter monitor replacement- head end for bio-marker toilet sensors- automated cancer sensors- automated nutritional intake metrics via AI/Camera/food-scale- cheap programable/intelligent medical alert bracelets- other home blood-marker testing- etc. . . . . etc. . . .The healthcare revolution is the obvious next big thing for Apple. It offers endless blue sky opportunities to disrupt/innovate on a large set of mega-$-bloated incumbent processes/costs.It requires deep-pocketed, persistently-patient, deeply-interrogational innovation built atop an exceptionally-simplified perceptualy-ergonomic user interface that is consistent enough to be mutually accessible to patients, doctors, labs, hospitals, pharmacies, and other health support systems.I can’t spot any other under-served mobile-device-based mega-$-value target industry capable of fuelling Apple’s voracious hardware-based profit food tube. Google not so much ?

        1. Emily Merkle

          Implants. Agree on your articulated future view, but healthcare is not a UI or design play. It is a measurement and reporting/relaying play at scale.

          1. SubstrateUndertow

            Very true but that is the easy part.EDITED:A simplified UI smoothly integrated with an effective/user-transparent hardware sensor array is the hard part that really speaks to the more difficult issue of usability and mass participation.

        2. LE

          Good points but: blood-glucose via non-invasive laser based sensorsThere are existing devices that already do this. And if there was a big enough demand there should already be a dedicated product to do this.blood pressure / pulse / arrhythmia / pacemaker monitoringExisting products that already do this. How many people need constant monitoring of bp and pulse anyway? Ditto for pacemaker monitoring size of market and already a solution to this that seems to work.head end for bio-marker toilet sensorsWhat is that exactly?automated cancer sensorsWhat are you saying that you would wear a device to be notified of impending cancer? Would a device that you didn’t have to have attached that is larger that you hook up periodically solve that same problem? Not sure I get this one.automated nutritional intake metrics via AI/Camera/food-scaleYou’re saying some kind of feedback loop that will make people pay attention to how they eat?cheap programable/intelligent medical alert braceletsCouldn’t this be done already with existing technology?In any case moto’s marketing seems to be targeted to a) young people who need to be notified of the next party and b) people who work in corporations who need to be reminded of their next meeting c) People who want to make a fashion statement.All of what you say is possible of course.One you seemed to have missed though is “you’ve had enough to drink”.

          1. SubstrateUndertow

            Some of these exist in the rough and some are just around the conner. Apple’s challenge is to wirelessly integrate them into a bio-sensor eco-system with an accessible UI.”blood-glucose via non-invasive laser based sensors”People have been working on these sensors for decades. The only one that claimed to work with repeatable accuracy is an Israeli product (don’t remember the name) that a year or so ago claimed to have some sort of government approval to go to market. They and their claims seem to have vanished. If you are familiar with such a working product I’d love a link.As for demand, diabetics can spent hundreds a month on test strips, so the market is huge! And with cheap continuous glucose monitoring and good AI management software the healthcare system could save a fortune!

    3. SubstrateUndertow

      Not so sure about”Apple could never compete with that”As bio-sensing/computing/communicating gadgets become miniaturizes down onto our wrists, style will be redefined by necessity around elegant ergonomically-tight integration between forum and function.After all smart-watches are not really watches any more than smart-phones are really phones and as such they will bring new style imperatives unlike your grandfather’s watch.That is right in Apple’s wheelhouse !

    4. jason wright

      does this work with the Moto G and Moto X? it looks kinda nice. what will it do for me though?call it the Motch.

  15. Chris Phenner

    While on the topic, an activist post.Five states (AZ, CA, IL, NY & OK) have proposed legislation to make computer science count towards high school math and science graduation requirements.…Hit the above petition for your state if you want access to CS to increase.

  16. pointsnfigures

    awesome. if the schools won’t teach it because they can’t (resources) or don’t want to-interested kids will find a way. Posted this to the Hyde Park Angels Facebook page. Great idea.

    1. Chris Phenner

      Nice and thank you. gets credit for alerting me. It sounds hokey but they really know how to send tight, actionable emails.

  17. Matt A. Myers

    If I break my spacebar key will someone replace my laptop for me? I’m already on my temporary one. :/

  18. Josh Haas

    I’m working on a sort of “scratch sr”: a way for adults trying to build apps to program things without coding: http://bubble.isWatching people get excited by discovering they can create things is really awesome 🙂

  19. Jim Canto

    I love the moment in the video here… 2:28… “developer” says; “it makes it wait 50 seconds.” Interviewer asks; “Wow… how did you know?” “Developer” answers; “It has a clock.”I’m pretty sure the developer is maybe 7yrs old? (Note the slight pause before he answers. I imagine he’s going through “why” she may have asked that question when, to him, the answer is simple and obvious.)Looks like the gap between PR and Engineering will still take time to close.Lastly, thanks for sharing this, @fredwilson:disqus. I have two Grandchildren who I need to introduce this to. Grandson is 13 and Granddaughter is 7. Both have smartphones as good or better than mine. And, both are way better with a touch screen than I am. I would love to see them get hooked on Scratch.

  20. Josh Haas

    I’m working on sort of a “scratch sr”, geared at helping adults program without having to code: http://bubble.isI've played with Scratch (and watched kids play with it) and it’s super-exciting the moment someone realizes they can actually make things….

  21. vadimoss

    Everyone who has a kid should donate! Those without a kid should double the contribution and learn to code themselves:) Gen Z will use coding to express themselves whether it’s for the sake of art, problem solving or environmental challenges.

  22. Elia Freedman

    Just donated. I’ve been thinking about introducing my 8 and 6 year old to programming but didn’t know where to look. I do now! Thanks!Yesterday afternoon my six year old sat on my lap and wanted to help me code. We changed a couple of lines of code together and she could see how they changed on the screen. Her eyes grew quite big that she could “make things happen.” I told her that making things on computers (programming) is like making things out of paper except my tools are keyboard and mouse while hers are scissors and glue.I think both my girls would be excited to play with Scratch and ScratchJr.

    1. Ruth BT

      They will love it. My 10 yr old son currently goes to a school where most learning is done on a tablet. The kids sit at lunchtime and code using scratch. There is a girl who is particularly good and all the boys wait for her to code something and then fork it. The girls seem to make great games that are a bit more visually pleasing – the boys then try to make them a bit more violent!He was 8 when introduced to scratch – perfect age. Scratch Jr looks great too.

      1. Elia Freedman

        Thanks, Ruth. I had someone else say basically the same thing. I’m really excited about this! Hope the girls are, too.

      2. fredwilson

        oh man, the school you are describing is the school of my dreams. i am trying to make that a reality in NYC

        1. Ruth BT

          You would think so but when they removed all the books in the library to make way for a paperless learning centre and didn’t give parents or teachers adequate learning support the shininess soon rubbed off. If is fantastic in some ways but there are massive inefficiencies just trying to use the technology. I would caution any school to spend as much time on teachers and parents as they do the kids to make the project a feasible reality a lot faster. Additionally, I fundraise to build libraries in developing countries and I sure as hell didn’t think I would have to be lobbying for a real library with books at our school.

    2. sigmaalgebra

      I’ll give you a Christmas story: My sister in law had two daughters and tried as hard as she could to get them interested in music and, in particular, learning to play piano. She had no progress at all. At Christmas my wife and I were visiting her family farm. I brought my violin and was upstairs practicing, likely some part of the Bach ‘Chaconne’. Tough piece to play, and some parts of it can be seen as ‘exciting’. Well my sister in law’s oldest daughter, about 9, walked upstairs and watched. She saw and heard something from me she didn’t from her mother — I was really involved; I really liked it. So, I put my violin on her left shoulder and under her left chin, showed her how to hold the violin with her left hand, and how to hold the bow and draw it across the strings. The next day her father asked me: “How much is a violin going to cost me?”.”Are we learning yet?”.Now, let’s see: Your six year old has long seen you working intensely at a computer with your keyboard and mouse. So, suddenly you park her in your lap, give her a lot of smiles, attention, caring, affection (she’s already ‘programmed’ to notice such things!), help her and do some little thing and see the results. You get excited.And now you say “I think both my girls would be excited” and you want all of us to believe that what your daughters are getting “excited” about is some computer thing?Hint: Starting in the crib, almost certainly your daughters were very, astoundingly highly interested in people and not things. Don’t bet more than 10 cents that that has changed or that suddenly they are interested in some computer thing! In particular, at or near the top of the list of what they were interested in, from the crib on, is you, Daddy, and there they want a lot of smiles, attention, caring, affection, praise, approval, caretaking, security, etc. They want to please Daddy. Indeed, quite likely they have already for years had you wrapped tightly around their little finger and had you so ‘programmed’ that you could never say no to either of them — they are really good at such things.Uh, their ‘interest’, ‘excitement’, and activity is not really about the computer thing and, instead, is about pleasing YOU. And you thought that it was the computer thing they were excited about! They’re not stupid, you know!It’s still true: If a girl is smart, and apparently you have two that are, they know that they don’t have to have ‘brains’ and, instead, can find a man, now their father, to do such things for them!With your computer you have, indeed, taught them important lessons they have learned very well! E.g., you have taught them a lesson now totally burned into their brains for life — computers are guy things and they are girls!Yes, as they play with a computer thing and you keep showering them with smiles, attention, caring, affection, praise, approval, caretaking, security, etc., for a few more years they may continue playing with a computer thing.But in about 10 years they will apply another lesson you have taught them and look for young men who are good with computers! And, they will look meek and sweet and say to the young man, nerd, “You’re so smart! You can do computers! I could never do anything like that!”.You expected something else? You want weak, sick, or dead limbs on the tree? Remember: “It’s not nice to try to fool Mother Nature.”. Mother Nature was there many hundreds of thousands of years before some computer thing and filtered out any girls who could be so easily distracted!Once I tried to fool Mother Nature and failed. Then I learned and accepted the quite traditional lessons I’ve outlined here.A girl can get much farther in life as a girl with a sweet smile, a little black dress, and a computer than with a computer alone!Without your smiles, praise, etc., for a computer and a dime they wouldn’t give even 10 cents.I mean, you do want the pitter patter of the little feet of grand kids, right? Or do you want to eat Thanksgiving dinner alone at McDonald’s, for a few decades?

  23. Tereza

    I have a kid (7) who can read (foundationally) and write…but she sucks at spelling. Is this a blocker?

    1. Rob Underwood

      Not at all.

      1. Emily Merkle

        Who needs to learn how to spell, anyway…

        1. SubstrateUndertow

          Up voted because I can’t spell shite !

    2. sigmaalgebra

      ALL the girls I knew in pretty dresses werereally good at spelling! So, put her in prettydresses! We’re talking full skirts, ribbons, bows,pastels, flowers, etc.Or, just any excuse to have girls be pretty again!If God wanted girls to pursue STEM subjects,be assertive, ‘lean in’, hack code, be moreinterested in things instead of people, and havemore math aptitude than verbal aptitude, thenhe would have made them boys instead!Am I going to win the award for the most offensive sexist comment so far on AVC?????

      1. SubstrateUndertow

        Expect no replays !

      2. Tereza

        She’s kicking ass at math…just got 100% on the state practice exam. She’s actually good with people (albeit charmingly sassy). Spelling…not her forte. {She can’t be bothered.} You’re probably right, I need to cut this off at the pass with full skirts, rainbows and unicorns.

        1. sigmaalgebra

          Congrats to you and your daughter.I like girls, think they all are cute, sweet, pretty, darling, adorable, precious, should never be told “no”, etc. and hate to see them suffer. And too many of them have suffered. At times I’ve suffered and hate to think that a girl would ever suffer. Maybe that’s my being ‘protective’; Mom said that if I’d had a sister, then I wouldn’t have had such a high opinion of girls!I hope you and your daughter don’t suffer.I know if I had to do anything that ‘girls do’, then I’d likely fail and if successful would no doubt have worked so hard at something I just was not built to do I would suffer. So, if a girl tries to do ‘boy things’, then maybe she will suffer, but I can’t say for sure because to me girls are like something from Venus for someone from Mars.For some aspects of ‘girl thinking’, math is cruel, unforgiving stuff. Math has it’s own way of doing things independent of and oblivious to what any human feels.Math totally saved my tail feathers: In the early grades, I was so socially awkward the teachers apparently wished I’d go out with the trash paper. But in the ninth grade I discovered math and that I could do it, and then no teacher on this planet, no matter what the heck she thought about me, could refuse to give me credit.In college, once I got out of the required ‘humanities’ courses, I took only math and physics, and the math was nothing but fun and games all day; I got “Honors in Math”, a good GRE math score, etc. Early in my career in mostly DoD work around DC, math saved me again. My current project has some original math at the core, and the math prerequisites are beyond all but a tiny fraction of full professors of computer science. I know the core math is solid because I did it with theorems and proofs (essentially all math taken at all seriously by the math community is, when polished and presented, just theorems and proofs).So, if can do math, then it can be powerful stuff. How powerful? My view is that it is the main key, with no alternative in sight, to most of the powerful future of computing, first by humans and later by ‘real AI’.So, I like math. But, then, I’ve been able to do it.If your daughter can do math and likes it, good. I wish her well. If she has to strain to do well in math, say to achieve some recent ‘feminist’ goals, then later she may, did I mention, suffer?But even “good” has to be qualified: Even if your daughter likes math and is good at it, so far the world of math, say, from trigonometry and solid geometry in high school on, is mostly a ‘man’s’ subject. In some colleges, there is a second track for ‘teacher training’, for K-12 math teachers to be, and there the books are easier and there are more women. Else math is a severe discipline, and the people the best at it now are men with high natural talent and uncompromising, iron determination.Unless the world of math changes a lot by the time your daughter gets to college, what she will see in solid math courses are essentially 100% men. But, there, she can exploit the same situation I did: If she can prove the theorems, then no one can deny her. There is some opportunity here since some of the exercises are difficult enough that likely at most one student in the class will get it. At least in ugrad school, a student who gets a few such exercises will be “honored above all other” students.Starting now, it should be easy for her to race ahead if both of you want. Maybe a person ends up knowing the material the same if they learn it after the course or before the course, but the second is much more impressive during the course! So, get her started with high school algebra. Prerequisites? Essentially none. So, get her several highly recommended books and let her dig in — for each ‘lesson’, read the material and, then, work the exercises. Let her have several books and concentrate on the one she likes best but have the others available if they are helpful occasionally (no book is perfect or is the best at everything it does; if something is not at all clear, then try another book). If know a good college or university math prof for occasional consultation for you and/or her, then better. No, a K-12 math teacher won’t qualify! If for the books you can get the teacher’s manual with the answers to the exercises, terrific! Let her learn how to work her way through a math book this way, that is, essentially independently; if you wish, give her praise when she is successful and help where she struggles.Then, sure, do the same for plane geometry (gorgeous stuff; like eating caramel popcorn — can’t stop), second year algebra, trigonometry, solid geometry, a little analytic geometry (the conic sections, amazing stuff), and college calculus. Same way: For each subject, have several of the best recommended books, pick a favorite and make occasional use of the rest if wish, study the text, work the exercises, check results with the teacher’s manual, and have a good math prof for occasional consultation.Then continue with linear algebra; work through 2-3 highly recommended books; and there end with the classic Halmos, ‘Finite Dimensional Vector Spaces’ (a crown jewel of civilization; a finite dimensional introduction to Hilbert space theory; from the knee of von Neumann when he was at the Institute for Advanced Study). Continue with Rudin, ‘Principles of Mathematical Analysis’ (calculus done carefully, plus a lot). For the exterior algebra, use Cartan’s book now available in English (high end approach to some of the crucial math of general relativity). Optionally take a pass through Spivak, ‘Calculus on Manifolds’ (there is a typo in Spivak I can help her with if she gets that far!). If do well, then have done well with the Harvard Math 55, sometimes called the most difficult ugrad math course in the country since at least at one time Halmos, Rudin, and Spivak were the standard texts. Next take a good pass through ‘abstract algebra’ — get some really good facility with set theory and then groups, rings (do some coding theory in high end electronic engineering), some number theory (right, for cryptology if nothing else), fields, vector spaces (a little more general than saw in linear algebra), maybe some category theory, maybe some group representations (e.g., part of how to identify molecules). Take a pass through ordinary differential equations (e.g., Coddington, elegant, polished, balanced); touch on partial differential equations (warning: can get lost here, so don’t; the level of precision will likely lower than in, say, Halmos or Rudin and, otherwise, way too difficult to read, and a lot of ‘engineering intuition’ is justified).Now can knock off quite a lot of serious mathematical physics as footnotes or small exercises. Would be right at the head of the class in anything about ‘machine learning’ or ‘big data’ if can stand the horribly low quality of that stuff!Take a first college course in probability and statistics (don’t take it very seriously; do that later).If wish, take a pass through optimization as in old ‘operations research’ and now part of computer science, machine learning, much of now classic mathematical economics, and wherever else it can be used now or in the future (there stand to be some uses) — linear, network linear, non-linear, dynamic.Then, right, heck, take the math GRE. If do well, then pick a university with a good math department, e.g., Princeton, get the materials they recommend for their Ph.D. qualifying exams, and dig in. Last time I read what Princeton said about that, no courses were offered for preparation for those exams, and students were expected to prepare on their own! So, your daughter will be fully ready to dig in. Likely she will have to dig into measure theory and functional analysis as in Royden, ‘Real Analysis’ and the first half of Rudin, ‘Real and Complex Analysis’ — gorgeous stuff; in places beyond belief. ‘Measure theory’? The grown up version of freshman calculus.With measure theory and functional analysis, have what is now something of a special opportunity so far not heavily pursued by US pure math departments: Take a pass through a few of the high end texts on ‘graduate probability’. The main authors are M. Loeve (the whole two volumes are a bit much), J. Neveu (succinct and just brilliant), L. Breiman (super nice to read), K. Chung. Then take a pass through stochastic processes, say, Cinlar’s little introduction (it has a lot of implicit monotone class arguments and, thus, often can’t be taken too seriously by a student on a first pass), and then, say, Karatzas and Shreve and/or Chung and Williams (Williams is a woman!). Then are set to do some high end work in some parts of mathematical physics, electronic engineering, and machine learning (the top end of that sad field). E.g., can do power spectral estimation on the data from the 3 degree Kelvin background radiation! Also will have an excellent start on high end versions of math for Wall Street.Now, pick some problems, pure or applied, in math, physics, engineering, computer science, the social sciences, etc., do some research to get solutions, largely just as math exercises, publish a few papers, and show up at a university for a year and get a Ph.D., possibly with the published papers as the dissertation. Social science? Maybe find a good problem in ‘social media’ and/or virality!A ‘good problem’ in academic research? First, a problem can solve where the solution will be ‘new, correct, and significant”. Second, for business, the solution will be powerful, valuable intellectual property can convert to software and run hidden in a server farm!Or, if want to learn more about how to do research from some real masters, then take some graduate courses (at Princeton supposedly they will all be introductions to research) and do some supervised research. From such on-campus work can get some of the ‘values’ and ‘socialization’ available there, if want such.Then get on with life. “Look, Ma, few or no classes!”, tuition, dorm life, drunken frat parties, etc.The description above is essentially what I did, that is, what worked for me, largely independent study, although I was dragged down by the classes. Tough to study when the teacher is exercising their vocal chords! In grad school, far too much of the week was spent driving to/from the university and sitting in classes, thus, leaving too little time for the crucial studying.For how to do research in applied math, I essentially discovered that on my own. My view is that the exercises in Halmos and Rudin are good starts. A famous recipe for rabbit stew starts out, “First catch a rabbit.”. Well, my recipe for good applied math is, “First get an application.”, that is, a good problem to solve. For more, have a lot of intuition, wild guesses, and evaluating wild guesses with lots of intuition; the careful theorems and proofs are mostly or all just at the end — no one ever told me this; maybe you first saw it here! Some really good math profs may occasionally have some more good ‘meta’ lessons to share. Note: I’m not such a “really good math prof”!Some advice: Don’t regard the books as perfect. While the level of quality of the best books is astoundingly high, there may not be a perfect book yet. So, each book can have some errors, obscure passages, poor explanations, out of place or way too hard exercises, wrong answers in the back of the book, a section that is way too hard considering the level of the rest of the book, etc. So, don’t get stuck. Don’t spend two weeks on a single hard exercise. If study well and still something is obscure, then look at alternate sources, go on and maybe come back later, use Stack Overflow, talk to a good math prof, etc. It is common to assume that need to understand every point, topic, theorem, proof, exercise or may be missing something crucial; instead, nearly never will be, and even if so then can patch it up later. So, don’t let some such chuckholes in the road cause giving up on the whole effort.Three big topics to skip over quickly, i.e., don’t get stuck, (1) the axiom of choice, (2) the Navier-Stokes equations, and (3) the question of P versus NP. Take one of those problems seriously and can go deep under water and never come up again.Most of what I have outlined here is ‘pure’ math and is by far the most highly polished, exact, and reliably correct part of civilization; so, if move into applications in science, engineering, technology, etc., then be prepared to face a lot of really sloppy work!In nearly all high end work in science, engineering, technology, and computer science, the workers are seriously handicapped by not having enough in math and, thus, struggle terribly. With what I’ve outlined here, your daughter will find a lot of such work as easy as shooting fish in a barrel.Before having your daughter go very far along what I have outlined here, get some second opinions!Generally, be careful, be very careful.Most important of all, don’t let her suffer. The girl who would get damaged is your daughter.In the end, you might prefer that she concentrate on finding a good husband and then giving you some good grand children instead of a stack of research papers.I’m failing to see why a career in math is better than being a good wife and mother.

    3. fredwilson

      try Scratch. i think she’ll be OK

  24. sigmaalgebra

    I see three issues:(1) Better Use of Available Time. Roughly my guess is that it would be better to have the target age group, say, 3-9, out playing ball, running, etc. So in this way they (A) would get some exercise and skills at sports and (B) from being with their peers make some progress in ‘socialization’. Considerations (A) and (B) are especially important now in the US where the economy pushes people too hard toward sedentary work alone or where ‘socialization’ plays less of a role.Or, there is limited time for this age group, and a question is, how to allocate the time. My strong suspicion is that, compared with computing, out playing ball with peers is better use of time.Or, people in ‘education’ have long had a very long list of things all the children really, really, really should do with their time. Bummer. Mostly these ‘education’ people are selling, mostly snake oil, and should quit ‘molesting’ the children.(2) What to Teach. The field of education is old, going way back. It was in, what, Chaucer that “Of the writing of books there is no end.”. The ancient Greeks were teaching the Pythagorean theorem? What was it, 200 years or so ago high end education in England wanted to teach Latin, Greek, and plane geometry as ‘intellectual discipline’ or a ticket to the upper classes or some such?So, there is a huge, old question what the heck to teach. I mean, sure, we should, of course, have a huge industry that keeps real estate property taxes high, provides stable employment for a huge fraction of the women who do have jobs, and tie up all the time of all children through at least age 18 and more likely 22, would you believe 25 or even 30, with some likely huge costs, cash or opportunity, to parents and students? Of course that’s what we should do, right? I mean, if the students had more freedom in the use of their time, then they might not understand all the details of the binomial coefficients, the hidden meanings in parts of English literature, properly patriotic sanitized versions of history, and English grammar as a special case of Latin! Shame!!!So, there is a huge, old question of what to teach, for culture, citizenship, prerequisites for more in academics, preparation for life, technical training, etc.Addressing this question and proposing adding some new material are challenging.(3) Computing. Scratch is heavily about computing, but it’s not at all clear just what will be important in computing in 15 years or what the heck would be good to teach now to improve results in 15 years.In simple terms, so far for much of the history of, say, commercial computing the ‘core’ has remained surprisingly stable and is, essentially,Allocate-FreeExpressions and assignmentIf-Then=ElseDo-WhileCall-ReturnTry-CatchSo, looking ahead only 15 years, assuming it is worthwhile to teach computing in K-12 at all, it’s probably safely worthwhile to teach that approach to computing.Summary. For something like Scratch, or other very easy to use ‘programming environments’ aimed at children, e.g., Logo, I still have to expect that there isn’t much value there and that playing baseball with peers would be better use of time.But the NSF has been supporting Scratch, and like most of NSF research maybe eventually the work will be considered worthwhile. Still, for now, for kids, I’d recommend sports.Later, for computing for 15 years from now, I’d recommend math instead of computer science. Sorry ’bout that.

    1. Emily Merkle


  25. SubstrateUndertow

    I want the – Scratch Sr. Edition !That lets old farts like we visually program audio/visual/text mash-up sequences by reusing all the audio/visual/text resources on the web as base components.With the programming/presentation layers being embedded as standard component in all browser engines no less 🙂

  26. Conrad Ross Schulman

    A big win for education and coding!



  28. Jack McDermott

    Interested to hear whether you think that ScratchJr–as a university-backed and developmentally-appropriate (with the research to support it)–holds a distinct product advantage compared to more market-focused products (Kodable, Hopscotch, etc.)?

  29. Jeff Magnusson

    I got tired of waiting for Scratch Jr (they’ve been talking and testing for a long, long time…) and after backing the great Hello Ruby on Kickstarter – (I have a 4yo and 7yo I’m getting started with), I learner about these other resources that are already live. Hopscotch and Kodable have been a fun intro on iPad with my boy, and Hopscotch seems to match Scratch Jr. has been interesting, it’s a bit simplistic but might be the perfect introduction to creating an instruction sequence to solve a problem.Also, Lego has the WeDo programmable construction set, where you use a visual style lifted from Scratch to control motors and actuators. My son has done classes with this stuff at schools and community centers, thinking about buying one of the educational packs for home. The physical crossover, introducing pullies and lever arms with a kid-friendly control language on the laptop is great.Some more resources are available on this post from Hello Ruby, I’ll be posting our experiences on my blog as we go.

  30. RobotEnomics

    The founder of ‘scratch’ also worked on the development of Mindstorms with Seymour Papert. Lego Mindstorms have helped millions of people develop an interest in coding and building product. Hardware development is also essential knowledge. For some background see – Training workers for the 21st century, with help from Lego robots

  31. Brandon Burns

    Ha! Fine, fine. Fred is bound to make this a topic in the coming days. I can wait a bit to have this discussion. 🙂

  32. pointsnfigures

    Scratch his comment-heh

  33. Brandon Burns

    Um… its cool!Personally, I’m over the teaching kids to code conversation. Nothing against the initiative, just not personally interesting to me.Hit me up when we start teaching kids proper design, and I’ll have much more to say. 🙂

  34. Brandon Burns

    Ha! I literally laughed out loud on that!