6 ms·

Simon Peyton-Jones on computer science in school curriculums [video]

147 points by francescoc 7y ago

hackermailman 7y ago
Shriram Krishnamurthi of Brown University already accomplished this and gave a great talk about the problems of creating a new course, when you can just shove it into the existing math classes without needing to hire thousands of specialist teachers or worry about yet another mandatory class where schools reduce the rigor, or worry about students not even being able to take the class because they are in remedial classes and never even make it to a computational elective https://youtu.be/5c0BvOlR5gs https://youtu.be/5c0BvOlR5gs
- serpent_skis 7y ago
  What a coincidence, I just started watching his Programming Languages Theory lectures. Seems like a wonderful professor.
- C1sc0cat 7y ago
  Ah the take over by the Maths department - just what we need
  - jrockway 7y ago
    I was required to take CS107 in school, you couldn't test out of it. It was taught by a chemistry professor. He knew nothing about programming except that he thought it was cool, and often taught us a lot of incorrect information. Being a jackass 18-year-old, I corrected him on all of this. I got 100% on every lab, assignment, and exam... and got a B in the class ("participation!"). It was about this time that I realized college might not be the thing for me.
    - excessive 7y ago
      > It was about this time that I realized college might not be the thing for me. Frequently, there are higher caliber instructors in colleges than high schools. If your goal is to learn something you don't know, you may have made the wrong choice.
      - jrockway 7y ago
        I had one good class: https://cr.yp.to/2004-494.html https://cr.yp.to/2004-494.html I will also say that I went to a good high school (IMSA) and a mediocre university (UIC). High school prepared me to be very bored in college. I had already taken calculus through "multi-variable calculus", number theory, 4 years of Japanese (including living for a year in Japan), physics, organic chemistry, learned 4 programming languages, wrote a tiny OS, etc. Most of college was just rehashing all this again, and then in your fourth year you get to learn new stuff. I wasn't willing to be bored for 3 years so that I could spend one year learning new things, so just quit. It helped that I found a security vulnerability in the school's registration system as part of that MCS494 class (100% sanctioned by the professor) and they decided that was a violating of their computer use agreement, so I was no longer allowed to use any computer systems, including the registration system. I am sure I could have fought it, but I decided it was a waste of time and got a job instead. 15 years later... I wrote a book, I've taught classes, I worked at Google for 6 years... so I am not sure I missed out on much. I would never tell anyone else not to pursue higher education, but it wasn't my thing. If I'm interested in something, I can spend a week teaching it to myself. Not everyone is like that, so college is great for them.
        musicale 7y ago
        Sounds like you were ready to start some graduate courses and research in those areas rather than rehashing undergraduate material that you had already mastered. Graduate courses and research can be pretty interesting and rewarding. Or perhaps ready to investigate some other subjects. For me many of the most beneficial and interesting courses and learning experiences are far afield from my primary work areas.
      - walshemj 7y ago
        It varies my first year Computing teacher at high school (15) was very good but she left and a maths teacher took over but by that point I was learning faster from books than he was. But because I was in the lower stream CSE (the exams for those supposed to leave at 16) they wouldn't let me take Computing at A level - even though I scored the max level in computing and maths.
    - graycat 7y ago
      In college, commonly can encounter junk. The best in college can be terrific stuff, much more difficult to get elsewhere. In college: (1) You will be around people, both students and teachers, who are in college instead of a public high school. On average those people will be in intellectual maturity several steps up from public high school and will help you raise your game in intellectual maturity, e.g., develop a good sense of quality and direction in intellectual travels, how to read a road map of the intellectual landscape -- stay with the good stuff and avoid bad turns and swamps. (2) From (1), you will meet such people, and they can change your life. E.g., there was a guy in middle school who got a Christmas present of a simple radio receiver kit. His family knew nothing about elementary electronics, but next door was a retired radio engineer. In a few years, the guy had his first class commercial broadcasting license, a really good ham radio transmitter in a shack out back, and each summer was the part-time, fill-in transmitter engineer for local TV stations -- three summers, each of the three TV stations. He met a guy with a family fortune interested in the music business and wanted a recording studio. The radio guy helped out. I met the radio guy in freshman physics in college. On the first test there were four questions, and I got all four. The best of anyone else in the class got was two. The prof called three right answers 100%, so I got 133%. Got 100% on the rest. So, I led the class. In particular, I beat the radio guy. When I was out of college, Dad was working at the Pentagon, and I found a job near there in physics, applied math, and computing, and my career improved quickly. Soon my annual salary was six times the cost of a new, high end Camaro. I got a phone call from the radio guy I'd known in college. He flew to DC, and we talked in vague, hush-hush terms about transportation. He wanted the name of a transportation consulting firm, and I gave him a name I'd heard but knew nothing more about. A year to two later, he asked me to come to Memphis. He sent me an airline ticket. He met me at the Memphis airport; it turned that he had a pilot's license and had a private plane he shared with the recording studio guy, and he flew us to Little Rock, Arkansas and to an office in the First National Bank building there. I learned about a startup to do US nationwide, over-night, high-priority, small package delivery. Long Lady Van de Geld could go shopping downtown, charge to her account, and ask for delivery. That evening UPS gathered all those packages from all those stores, did a central hub sort for the city, and the next day did the deliveries. So, yes, there in Little Rock was the start of Federal Express, now known as FedEx. The recording studio guy was Frederick W. Smith, founder, CEO, COB. His idea was much like that of UPS but to use a 500 MPH truck and do the central hub sort for the US in Memphis. For his offices, airplane maintenance, and hub sort, he had leased some land on the Memphis airport with some old WWII aircraft hanger space. For his 500 MPH trucks, Smith had started with a popular business jet, the Dassault Fanjet Falcon DA-20, maximum gross takeoff weight of 28,660 pounds and maximum cargo weight of 6000 pounds. The plane was from France, especially rugged, with US electronics and engines. Smith had worked out an FAA approved modification of the Falcon to carry cargo. Smith had several of the Falcons and buying more. One afternoon, Smith tried scheduling his fleet, for an initial IIRC 11 planes but also for his planned 33 planes covering 90 major US cities. When the afternoon was over, he walked out of his office tired and frustrated and announced "we need a computer". The radio guy, then, called me. I got flown to Little Rock to see about scheduling the fleet. I joined FedEx. I had been consulting in applied math and computing at Georgetown University and teaching computer science, gave up the consulting, continued the teaching, got a time-sharing terminal to an IBM VM/CMS (then CP67/CMS -- right, likely the first good virtual machine) system, and called the local IBM office for the full set of IBM manuals on PL/I (I'd been using there in DC). Soon an IBM Marketing Representative hand delivered the PL/I documentation and was all ears on what new company could want PL/I documentation? Six weeks later I was done with the teaching and the software for scheduling the fleet and drove my high end Camaro to Memphis. [Only for car guys -- 396 cubic inches and a 2.56:1 rear axle ratio, really good for fast travel on the Interstates from DC to Memphis.] Some Members of the FedEx BoD had a lot of airline experience and were concerned about the fleet scheduling problem. Crucial funding was in question. One evening another guy and I used my software to schedule the fleet for the full planned 33 planes and 90 US cities. Copies of the schedule were passed around, and two representatives of BoD Member General Dynamics went over the schedule in detail and announced "It's a little tight in a few places, but it's flyable." At a senior staff meeting, Smith's reaction was "Solved the most important problem facing the start of Federal Express.". A few months later the BoD was concerned again, this time about revenue projections. We knew the current revenue and what we had in mind, call it b dollars a day, for the full service with 33 planes. So the projections were in a sense an interpolation between those two. I At each time t, in days, let the revenue be y(t) and assume the growth rate is directly proportional to the current number of customers (or revenue y(t)) using, and, hence, generating publicity about the service, and also directly proportional to the number of target customers (b - y(t)) not yet served but hearing about FedEx via the people already using the service, that is, word of mouth advertising (e.g., maybe receiving packages from FedEx, virality). Then for some constant of proportionality we have d/dt y(t) = y'(t) = k y(t) (b - y(t)) a first order, linear, constant coefficient, ordinary differential equation initial value problem with a closed form solution essentially the classic lazy 'S' logistics curve. We picked a good looking value of k, and drew a graph. The BoD was happy again, and I was Director of Operations Research with my office next to Smith's. The promised stock was late; my wife was still in DC in her Ph.D. program, and I wanted something for my financial security no one could take away from me, either stock in FedEx or a Ph.D. With the stock late and my wife in DC, I left FedEx and got a Ph.D. Lesson: Even if know a lot of good, useful stuff, it STILL can be from important to crucial "who you know", and one of the best places to meet such people is in college. (3) At times college can teach you useful stuff much more difficult to get elsewhere. And college can give you a good start, road map of the knowledge landscape, tough to get elsewhere, that can let you learn more WITHOUT lots of bad turns and wasted time from a poor road map. From such a road map, I did quite a lot before my Ph.D.: I entered with a dissertation problem in mind and with a good, first intuitive solution. On the five Ph.D. qualifying exams, I did the best in the class on four of them, and for three of those the success was from what I'd learned on college or taught myself from the road map. In my first year in grad school, from some good coursework, SUPER tough to get otherwise, in the summer I was able independently to complete my research for my Ph.D. except for the illustrative computing. Later I did the computing, typed in the dissertation, stood for my oral exam, and graduated. Lesson: A good road map for independent study can be crucial. Otherwise can waste much of a career in just bad turns in the landscape. Summary: It can be (A) how you matured intellectually in the environment of college, (B) WHAT you know from college or independent study from following a road map you got in college, and (C) WHO you know, and maybe met in college. Net, college is not always a nice place, but the best of it can be pretty good.
  - threatofrain 7y ago
    That, or schools would need to hire more teachers who specialize in early CS education. This is all assuming that CS-for-all is a good thing.
- throwaway6734 7y ago
  Thanks for sharing this!
- chrisseaton 7y ago
  > when you can just shove it into the existing math classes This is how it's already done in the UK, and has been for decades. When I was in high-school (or rather the UK equivalent) about twenty years ago we did discrete maths, algorithms, and data structures in the regular maths classes. I don't think computers were even mentioned - it was all described as maths topics. And this was a bog-standard state school. So when I want to university interviews I was already able to describe for example how to implement a hash table.
  - MarcScott 7y ago
    Not any more. I think you'll find that the maths specifications have been trimmed down a little.[1] Now we have discrete Computer Science at GCSE and A-Level. [1] https://filestore.aqa.org.uk/resources/mathematics/specifications/AQA-7357-SP-2017.PDF https://filestore.aqa.org.uk/resources/mathematics/specifica...
    - OJFord 7y ago
      Within mathematics, the 'D' (e.g. D1) modules are 'discrete' at A level, and are mostly (basic) graph theory/algorithms. Cf. 'S' (Statistics), 'M' (Mechanics, i.e. mathematics of physics), 'C' (Core), and 'FP' (Further Pure). You're right that Computer Science/Computing is separate, but it's not a particularly common and certainly not a required choice for people pursuing CS, many view it as unworthwhile/a joke either in advance or retrospect. Certainly when I went through they'd far rather you took as much mathematics as possible (which can be three full A levels - Mathematics/Further/Additional if you take all of the modules) followed by sciences. A faculty member at Cambridge expressed an explicit preference for Latin over Computing at A level.
  - closeparen 7y ago
    I’m thankful every day for getting into programming early enough to know that the adults saying “programming is like math” were full of shit. No career could be less appealing than cranking arithmetic, polynomials, integrals, and derivatives for 8 hours day in and day out, which is of course what “math” means when you’re in K12. If it were introduced to me in that context and by teachers with that mindset, I wouldn’t be here.
the_af 7y ago
I really enjoy listening to SPJ on almost any topic. He seems so friendly as well as knowledgeable. And humble! I already like Haskell, but I bet he could convince me to like almost anything else.
- CraigJPerry 7y ago
  It's funny - i also share your views but i don't remember enjoying his lectures at Glasgow Uni. That was a wee while ago now but i really enjoy his modern stuff.
- kreetx 7y ago
  Here is another one which I've really liked over the years: https://www.youtube.com/watch?v=3NRrUvqQlnY&t https://www.youtube.com/watch?v=3NRrUvqQlnY&t He is so warm and spontaneous, yet smart and obviously can invent exact and concrete things!
  - the_af 7y ago
    Yes, I love that talk too! Thanks for posting it.
hogFeast 7y ago
Just my 2c but this is one of the worst ideas. First, this was designed by politicians who I am sure everyone on this sub would hate (basically, the Brexit boys). The stated aim was to be more elitist: make subjects much harder, encourage smart children...and dumb children...well...sorry, if you are dumb when we test you at 11 then you will be digging ditches. Second, the logic for this came from people who had never taught in schools. Again, the people above declared war on teachers. And then took advice on how to educate children from people who had only worked in tertiary education (I actually support a lot of the principles but the implementation has been beyond dire, and loaded with corruption). Third, try explaining programming without computers to a child. There are so many abstract concepts...it is just insane. I understand why an academic thinks this is a good idea. More jobs for the boys. But it still makes no sense. Fourth, this feeds into the aspect of British culture that reveres irrelevant knowledge (and despises practice). Nowhere is this more evident that in CS departments. Example: the UK has great CS depts but no innovation within business. The university local to me has a top ML department, they have been doing speech recognition since the 60s...all the PHds leave, there is only one speech recognition startup in the city...and that is govt-funded afaik. Taking advice from people like the OP is suicidal.
- DavidTNcl 7y ago
  The 11+ is long gone.
- Feint1 7y ago
  > I understand why an academic thinks this is a good idea. More jobs for the boys. This isn't fair. Academics usually have a true love of their subject and desire to teach that to the next generation. >Third, try explaining programming without computers to a child. There are so many abstract concepts...it is just insane. They teach many abstract concepts. The current curricula for many subjects are already full of abstract ideas that we expect children to learn. Programming is basically functions, logic and basic algebra. It's not remotely difficult to teach these to children. >Fourth, this feeds into the aspect of British culture that reveres irrelevant knowledge (and despises practice). Nowhere is this more evident that in CS departments. Example: the UK has great CS depts but no innovation within business. The university local to me has a top ML department, they have been doing speech recognition since the 60s...all the PHds leave, there is only one speech recognition startup in the city...and that is govt-funded afaik. Taking advice from people like the OP is suicidal. I think this is wrong. Developing strong theoretical understanding is not pointless. You can't have innovation without theory. I agree that maybe the country needs to apply that academic kowledge better to money making endeavours, but I don't believe that keeping the education system as it is now is going to help. We need more numerate people with detailed knowledge of scientific disciplines in order to drive innovation.
getaclue 7y ago
Thanks for sharing this!
willtim 7y ago
What a fantastic call to arms! I'm tempted to get involved myself after watching this talk. I couldn't agree more with everything SPJ said.
- MarcScott 7y ago
  Want a quick start - start a CodeClub or Coder Dojo. https://codeclub.org/en/ https://codeclub.org/en/ https://coderdojo.com/ https://coderdojo.com/
donjigweed 7y ago
SPJ is an absolute beast, though whenever I hear a talk of his all I can think of is the Impressive Clergyman.
ngneer 7y ago
Thanks for sharing. A great reminder of the beauty of the field.
gexla 7y ago
Great video, listening to it now. His accent is fun. He sounds like he could play Caesar in Life Of Brian.
0898 7y ago
I unknowingly met him once in 2002 or 2003, and I’ve just this second realised who he was. I was visiting Microsoft Research in Cambridge writing an article for PC Format magazine, a British computer mag. They took me to see a few people to show off Microsoft Research. There was this one guy who talked to me in his office. He got very excited about Haskell (which I didn’t know what it was) and he was also working on some Excel feature he was convinced of the brilliance of, but hasn’t persuaded the product team of. His theory was that the only thing you know about an Excel spreadsheet is that most of your predications are wrong. My interpretation was that he was proposing a way of indicating a probabilistic range rather than a hard number. I don’t know if this ever made it into Excel, or if indeed this was what he was talking about. But it’s amazing to see this odd fellow I met was actually quite famous.
- gnufx 7y ago
  That work might have been https://www.microsoft.com/en-us/research/publication/a-user-centred-approach-to-functions-in-excel/ https://www.microsoft.com/en-us/research/publication/a-user-... Perhaps it's significant in context that a writer for a UK computing mag wouldn't know of SPJ. Would they know of any UK CS luminaries, like Hoare?
  - 0898 7y ago
    Yes, that’s it. Good find. I’m the first to admit I was an IT dilettante, despite covering it.
kenni 7y ago
I work in a similar organisation to SPJ in Australia, the Australian Computing Academy (https://aca.edu.au https://aca.edu.au). Everything he said mirrors quite closely what has been happening in Australia for the past decade. We also have a national curriculum (https://docs.acara.edu.au/resources/Digital_Technologies_-_Sequence_of_content.pdf https://docs.acara.edu.au/resources/Digital_Technologies_-_S...) which we're rolling out across the country. The main take away I've received from doing this job for the past two years is the same one that SPJ reiterates in his talk. Teaching is a very distributed discipline. There's no top down solution to implementing a new curriculum. If you work in the industry and volunteer to help a school you can make a huge difference!
MarcScott 7y ago
I work for the Raspberry Pi Foundation, and we are very involved with the NCCE which SPJ talked about. If any of your are even slightly inspired by what was talked about in this video, and would like to get involved with teaching the next generation of hackers, then I'd like to suggest that you think about starting or supporting one of the clubs SPJ mentioned. CodeClubs are school based clubs that you can run in a local school. We provide all the resources and guidance you need, and you can get involved here. https://codeclub.org/en/ https://codeclub.org/en/ Coder Dojos are clubs that are run at weekends, in libraries, offices and community centers. Again, we provide educational resources to help you out. https://coderdojo.com/ https://coderdojo.com/ Both of these clubs are international. So it doesn't matter which country you are based in. Also happy to answer any questions people have. Oh, and if you have kids and want to get them started, checkout some of our projects. These are the Scratch ones. https://projects.raspberrypi.org/en/projects?software%5B%5D=scratch https://projects.raspberrypi.org/en/projects?software%5B%5D=...
agumonkey 7y ago
Next : sicm in physics class

147 points by francescoc 7y ago

Shriram Krishnamurthi of Brown University already accomplished this and gave a great talk about the problems of creating a new course, when you can just shove it into the existing math classes without needing to hire thousands of specialist teachers or worry about yet another mandatory class where schools reduce the rigor, or worry about students not even being able to take the class because they are in remedial classes and never even make it to a computational elective https://youtu.be/5c0BvOlR5gs https://youtu.be/5c0BvOlR5gs

serpent_skis 7y ago
What a coincidence, I just started watching his Programming Languages Theory lectures. Seems like a wonderful professor.
C1sc0cat 7y ago
Ah the take over by the Maths department - just what we need
- jrockway 7y ago
  I was required to take CS107 in school, you couldn't test out of it. It was taught by a chemistry professor. He knew nothing about programming except that he thought it was cool, and often taught us a lot of incorrect information. Being a jackass 18-year-old, I corrected him on all of this. I got 100% on every lab, assignment, and exam... and got a B in the class ("participation!"). It was about this time that I realized college might not be the thing for me.
  - excessive 7y ago
    > It was about this time that I realized college might not be the thing for me. Frequently, there are higher caliber instructors in colleges than high schools. If your goal is to learn something you don't know, you may have made the wrong choice.
    - jrockway 7y ago
      I had one good class: https://cr.yp.to/2004-494.html https://cr.yp.to/2004-494.html I will also say that I went to a good high school (IMSA) and a mediocre university (UIC). High school prepared me to be very bored in college. I had already taken calculus through "multi-variable calculus", number theory, 4 years of Japanese (including living for a year in Japan), physics, organic chemistry, learned 4 programming languages, wrote a tiny OS, etc. Most of college was just rehashing all this again, and then in your fourth year you get to learn new stuff. I wasn't willing to be bored for 3 years so that I could spend one year learning new things, so just quit. It helped that I found a security vulnerability in the school's registration system as part of that MCS494 class (100% sanctioned by the professor) and they decided that was a violating of their computer use agreement, so I was no longer allowed to use any computer systems, including the registration system. I am sure I could have fought it, but I decided it was a waste of time and got a job instead. 15 years later... I wrote a book, I've taught classes, I worked at Google for 6 years... so I am not sure I missed out on much. I would never tell anyone else not to pursue higher education, but it wasn't my thing. If I'm interested in something, I can spend a week teaching it to myself. Not everyone is like that, so college is great for them.
      - musicale 7y ago
        Sounds like you were ready to start some graduate courses and research in those areas rather than rehashing undergraduate material that you had already mastered. Graduate courses and research can be pretty interesting and rewarding. Or perhaps ready to investigate some other subjects. For me many of the most beneficial and interesting courses and learning experiences are far afield from my primary work areas.
    - walshemj 7y ago
      It varies my first year Computing teacher at high school (15) was very good but she left and a maths teacher took over but by that point I was learning faster from books than he was. But because I was in the lower stream CSE (the exams for those supposed to leave at 16) they wouldn't let me take Computing at A level - even though I scored the max level in computing and maths.
  - graycat 7y ago
    In college, commonly can encounter junk. The best in college can be terrific stuff, much more difficult to get elsewhere. In college: (1) You will be around people, both students and teachers, who are in college instead of a public high school. On average those people will be in intellectual maturity several steps up from public high school and will help you raise your game in intellectual maturity, e.g., develop a good sense of quality and direction in intellectual travels, how to read a road map of the intellectual landscape -- stay with the good stuff and avoid bad turns and swamps. (2) From (1), you will meet such people, and they can change your life. E.g., there was a guy in middle school who got a Christmas present of a simple radio receiver kit. His family knew nothing about elementary electronics, but next door was a retired radio engineer. In a few years, the guy had his first class commercial broadcasting license, a really good ham radio transmitter in a shack out back, and each summer was the part-time, fill-in transmitter engineer for local TV stations -- three summers, each of the three TV stations. He met a guy with a family fortune interested in the music business and wanted a recording studio. The radio guy helped out. I met the radio guy in freshman physics in college. On the first test there were four questions, and I got all four. The best of anyone else in the class got was two. The prof called three right answers 100%, so I got 133%. Got 100% on the rest. So, I led the class. In particular, I beat the radio guy. When I was out of college, Dad was working at the Pentagon, and I found a job near there in physics, applied math, and computing, and my career improved quickly. Soon my annual salary was six times the cost of a new, high end Camaro. I got a phone call from the radio guy I'd known in college. He flew to DC, and we talked in vague, hush-hush terms about transportation. He wanted the name of a transportation consulting firm, and I gave him a name I'd heard but knew nothing more about. A year to two later, he asked me to come to Memphis. He sent me an airline ticket. He met me at the Memphis airport; it turned that he had a pilot's license and had a private plane he shared with the recording studio guy, and he flew us to Little Rock, Arkansas and to an office in the First National Bank building there. I learned about a startup to do US nationwide, over-night, high-priority, small package delivery. Long Lady Van de Geld could go shopping downtown, charge to her account, and ask for delivery. That evening UPS gathered all those packages from all those stores, did a central hub sort for the city, and the next day did the deliveries. So, yes, there in Little Rock was the start of Federal Express, now known as FedEx. The recording studio guy was Frederick W. Smith, founder, CEO, COB. His idea was much like that of UPS but to use a 500 MPH truck and do the central hub sort for the US in Memphis. For his offices, airplane maintenance, and hub sort, he had leased some land on the Memphis airport with some old WWII aircraft hanger space. For his 500 MPH trucks, Smith had started with a popular business jet, the Dassault Fanjet Falcon DA-20, maximum gross takeoff weight of 28,660 pounds and maximum cargo weight of 6000 pounds. The plane was from France, especially rugged, with US electronics and engines. Smith had worked out an FAA approved modification of the Falcon to carry cargo. Smith had several of the Falcons and buying more. One afternoon, Smith tried scheduling his fleet, for an initial IIRC 11 planes but also for his planned 33 planes covering 90 major US cities. When the afternoon was over, he walked out of his office tired and frustrated and announced "we need a computer". The radio guy, then, called me. I got flown to Little Rock to see about scheduling the fleet. I joined FedEx. I had been consulting in applied math and computing at Georgetown University and teaching computer science, gave up the consulting, continued the teaching, got a time-sharing terminal to an IBM VM/CMS (then CP67/CMS -- right, likely the first good virtual machine) system, and called the local IBM office for the full set of IBM manuals on PL/I (I'd been using there in DC). Soon an IBM Marketing Representative hand delivered the PL/I documentation and was all ears on what new company could want PL/I documentation? Six weeks later I was done with the teaching and the software for scheduling the fleet and drove my high end Camaro to Memphis. [Only for car guys -- 396 cubic inches and a 2.56:1 rear axle ratio, really good for fast travel on the Interstates from DC to Memphis.] Some Members of the FedEx BoD had a lot of airline experience and were concerned about the fleet scheduling problem. Crucial funding was in question. One evening another guy and I used my software to schedule the fleet for the full planned 33 planes and 90 US cities. Copies of the schedule were passed around, and two representatives of BoD Member General Dynamics went over the schedule in detail and announced "It's a little tight in a few places, but it's flyable." At a senior staff meeting, Smith's reaction was "Solved the most important problem facing the start of Federal Express.". A few months later the BoD was concerned again, this time about revenue projections. We knew the current revenue and what we had in mind, call it b dollars a day, for the full service with 33 planes. So the projections were in a sense an interpolation between those two. I At each time t, in days, let the revenue be y(t) and assume the growth rate is directly proportional to the current number of customers (or revenue y(t)) using, and, hence, generating publicity about the service, and also directly proportional to the number of target customers (b - y(t)) not yet served but hearing about FedEx via the people already using the service, that is, word of mouth advertising (e.g., maybe receiving packages from FedEx, virality). Then for some constant of proportionality we have d/dt y(t) = y'(t) = k y(t) (b - y(t)) a first order, linear, constant coefficient, ordinary differential equation initial value problem with a closed form solution essentially the classic lazy 'S' logistics curve. We picked a good looking value of k, and drew a graph. The BoD was happy again, and I was Director of Operations Research with my office next to Smith's. The promised stock was late; my wife was still in DC in her Ph.D. program, and I wanted something for my financial security no one could take away from me, either stock in FedEx or a Ph.D. With the stock late and my wife in DC, I left FedEx and got a Ph.D. Lesson: Even if know a lot of good, useful stuff, it STILL can be from important to crucial "who you know", and one of the best places to meet such people is in college. (3) At times college can teach you useful stuff much more difficult to get elsewhere. And college can give you a good start, road map of the knowledge landscape, tough to get elsewhere, that can let you learn more WITHOUT lots of bad turns and wasted time from a poor road map. From such a road map, I did quite a lot before my Ph.D.: I entered with a dissertation problem in mind and with a good, first intuitive solution. On the five Ph.D. qualifying exams, I did the best in the class on four of them, and for three of those the success was from what I'd learned on college or taught myself from the road map. In my first year in grad school, from some good coursework, SUPER tough to get otherwise, in the summer I was able independently to complete my research for my Ph.D. except for the illustrative computing. Later I did the computing, typed in the dissertation, stood for my oral exam, and graduated. Lesson: A good road map for independent study can be crucial. Otherwise can waste much of a career in just bad turns in the landscape. Summary: It can be (A) how you matured intellectually in the environment of college, (B) WHAT you know from college or independent study from following a road map you got in college, and (C) WHO you know, and maybe met in college. Net, college is not always a nice place, but the best of it can be pretty good.
- threatofrain 7y ago
  That, or schools would need to hire more teachers who specialize in early CS education. This is all assuming that CS-for-all is a good thing.
throwaway6734 7y ago
Thanks for sharing this!
chrisseaton 7y ago
> when you can just shove it into the existing math classes This is how it's already done in the UK, and has been for decades. When I was in high-school (or rather the UK equivalent) about twenty years ago we did discrete maths, algorithms, and data structures in the regular maths classes. I don't think computers were even mentioned - it was all described as maths topics. And this was a bog-standard state school. So when I want to university interviews I was already able to describe for example how to implement a hash table.
- MarcScott 7y ago
  Not any more. I think you'll find that the maths specifications have been trimmed down a little.[1] Now we have discrete Computer Science at GCSE and A-Level. [1] https://filestore.aqa.org.uk/resources/mathematics/specifications/AQA-7357-SP-2017.PDF https://filestore.aqa.org.uk/resources/mathematics/specifica...
  - OJFord 7y ago
    Within mathematics, the 'D' (e.g. D1) modules are 'discrete' at A level, and are mostly (basic) graph theory/algorithms. Cf. 'S' (Statistics), 'M' (Mechanics, i.e. mathematics of physics), 'C' (Core), and 'FP' (Further Pure). You're right that Computer Science/Computing is separate, but it's not a particularly common and certainly not a required choice for people pursuing CS, many view it as unworthwhile/a joke either in advance or retrospect. Certainly when I went through they'd far rather you took as much mathematics as possible (which can be three full A levels - Mathematics/Further/Additional if you take all of the modules) followed by sciences. A faculty member at Cambridge expressed an explicit preference for Latin over Computing at A level.
- closeparen 7y ago
  I’m thankful every day for getting into programming early enough to know that the adults saying “programming is like math” were full of shit. No career could be less appealing than cranking arithmetic, polynomials, integrals, and derivatives for 8 hours day in and day out, which is of course what “math” means when you’re in K12. If it were introduced to me in that context and by teachers with that mindset, I wouldn’t be here.

the_af 7y ago

I really enjoy listening to SPJ on almost any topic. He seems so friendly as well as knowledgeable. And humble! I already like Haskell, but I bet he could convince me to like almost anything else.

CraigJPerry 7y ago
It's funny - i also share your views but i don't remember enjoying his lectures at Glasgow Uni. That was a wee while ago now but i really enjoy his modern stuff.
kreetx 7y ago
Here is another one which I've really liked over the years: https://www.youtube.com/watch?v=3NRrUvqQlnY&t https://www.youtube.com/watch?v=3NRrUvqQlnY&t He is so warm and spontaneous, yet smart and obviously can invent exact and concrete things!
- the_af 7y ago
  Yes, I love that talk too! Thanks for posting it.

hogFeast 7y ago

Just my 2c but this is one of the worst ideas. First, this was designed by politicians who I am sure everyone on this sub would hate (basically, the Brexit boys). The stated aim was to be more elitist: make subjects much harder, encourage smart children...and dumb children...well...sorry, if you are dumb when we test you at 11 then you will be digging ditches. Second, the logic for this came from people who had never taught in schools. Again, the people above declared war on teachers. And then took advice on how to educate children from people who had only worked in tertiary education (I actually support a lot of the principles but the implementation has been beyond dire, and loaded with corruption). Third, try explaining programming without computers to a child. There are so many abstract concepts...it is just insane. I understand why an academic thinks this is a good idea. More jobs for the boys. But it still makes no sense. Fourth, this feeds into the aspect of British culture that reveres irrelevant knowledge (and despises practice). Nowhere is this more evident that in CS departments. Example: the UK has great CS depts but no innovation within business. The university local to me has a top ML department, they have been doing speech recognition since the 60s...all the PHds leave, there is only one speech recognition startup in the city...and that is govt-funded afaik. Taking advice from people like the OP is suicidal.

DavidTNcl 7y ago
The 11+ is long gone.
Feint1 7y ago
> I understand why an academic thinks this is a good idea. More jobs for the boys. This isn't fair. Academics usually have a true love of their subject and desire to teach that to the next generation. >Third, try explaining programming without computers to a child. There are so many abstract concepts...it is just insane. They teach many abstract concepts. The current curricula for many subjects are already full of abstract ideas that we expect children to learn. Programming is basically functions, logic and basic algebra. It's not remotely difficult to teach these to children. >Fourth, this feeds into the aspect of British culture that reveres irrelevant knowledge (and despises practice). Nowhere is this more evident that in CS departments. Example: the UK has great CS depts but no innovation within business. The university local to me has a top ML department, they have been doing speech recognition since the 60s...all the PHds leave, there is only one speech recognition startup in the city...and that is govt-funded afaik. Taking advice from people like the OP is suicidal. I think this is wrong. Developing strong theoretical understanding is not pointless. You can't have innovation without theory. I agree that maybe the country needs to apply that academic kowledge better to money making endeavours, but I don't believe that keeping the education system as it is now is going to help. We need more numerate people with detailed knowledge of scientific disciplines in order to drive innovation.

getaclue 7y ago

Thanks for sharing this!

willtim 7y ago

What a fantastic call to arms! I'm tempted to get involved myself after watching this talk. I couldn't agree more with everything SPJ said.

MarcScott 7y ago
Want a quick start - start a CodeClub or Coder Dojo. https://codeclub.org/en/ https://codeclub.org/en/ https://coderdojo.com/ https://coderdojo.com/

donjigweed 7y ago

SPJ is an absolute beast, though whenever I hear a talk of his all I can think of is the Impressive Clergyman.

ngneer 7y ago

Thanks for sharing. A great reminder of the beauty of the field.

gexla 7y ago

Great video, listening to it now. His accent is fun. He sounds like he could play Caesar in Life Of Brian.

0898 7y ago

I unknowingly met him once in 2002 or 2003, and I’ve just this second realised who he was. I was visiting Microsoft Research in Cambridge writing an article for PC Format magazine, a British computer mag. They took me to see a few people to show off Microsoft Research. There was this one guy who talked to me in his office. He got very excited about Haskell (which I didn’t know what it was) and he was also working on some Excel feature he was convinced of the brilliance of, but hasn’t persuaded the product team of. His theory was that the only thing you know about an Excel spreadsheet is that most of your predications are wrong. My interpretation was that he was proposing a way of indicating a probabilistic range rather than a hard number. I don’t know if this ever made it into Excel, or if indeed this was what he was talking about. But it’s amazing to see this odd fellow I met was actually quite famous.

gnufx 7y ago
That work might have been https://www.microsoft.com/en-us/research/publication/a-user-centred-approach-to-functions-in-excel/ https://www.microsoft.com/en-us/research/publication/a-user-... Perhaps it's significant in context that a writer for a UK computing mag wouldn't know of SPJ. Would they know of any UK CS luminaries, like Hoare?
- 0898 7y ago
  Yes, that’s it. Good find. I’m the first to admit I was an IT dilettante, despite covering it.

kenni 7y ago

I work in a similar organisation to SPJ in Australia, the Australian Computing Academy (https://aca.edu.au https://aca.edu.au). Everything he said mirrors quite closely what has been happening in Australia for the past decade. We also have a national curriculum (https://docs.acara.edu.au/resources/Digital_Technologies_-_Sequence_of_content.pdf https://docs.acara.edu.au/resources/Digital_Technologies_-_S...) which we're rolling out across the country. The main take away I've received from doing this job for the past two years is the same one that SPJ reiterates in his talk. Teaching is a very distributed discipline. There's no top down solution to implementing a new curriculum. If you work in the industry and volunteer to help a school you can make a huge difference!

MarcScott 7y ago

I work for the Raspberry Pi Foundation, and we are very involved with the NCCE which SPJ talked about. If any of your are even slightly inspired by what was talked about in this video, and would like to get involved with teaching the next generation of hackers, then I'd like to suggest that you think about starting or supporting one of the clubs SPJ mentioned. CodeClubs are school based clubs that you can run in a local school. We provide all the resources and guidance you need, and you can get involved here. https://codeclub.org/en/ https://codeclub.org/en/ Coder Dojos are clubs that are run at weekends, in libraries, offices and community centers. Again, we provide educational resources to help you out. https://coderdojo.com/ https://coderdojo.com/ Both of these clubs are international. So it doesn't matter which country you are based in. Also happy to answer any questions people have. Oh, and if you have kids and want to get them started, checkout some of our projects. These are the Scratch ones. https://projects.raspberrypi.org/en/projects?software%5B%5D=scratch https://projects.raspberrypi.org/en/projects?software%5B%5D=...

agumonkey 7y ago

Next : sicm in physics class