One day, we'll all be programmers
zamfi.net
zamfi.net
haven't people been saying that since the 70s? even SQL was designed with the intent that accountants could use it. but programming is really hard, especially if you want something that's done well.
i've also yet to see any general purpose robots that can do anything useful. i still think it's a long way out, if at all possible. there are tons of considerations that go into just about anything and if there's no magic AI there's really no way to program every useful (nevermind adaptable) constraint into everything.
"Everyone needs computer programming. It will be the way we speak to the servants."
-John McCarthy, creator of Lisp and coiner of the term Artifical Intelligence.
But I don't think the age of the claim is in itself a reason to not believe it.
The more you know about programming, the more you can do with general purpose devices when they arise, and more importantly, the more devices you can make into general purpose devices.
A simple recent example is Kinect. It isn't a general purpose device. It was a gaming doohickey. Thanks to programming, you can turn it into a (more!) general purpose device.
One some level, we've already reached that point - though we're not programming directly to communicate with McCarthy's servants we are already on that path.
Also, have you seen what non-programmers can do with Excel? I'm often blown away by the technical skill demonstrated by people who don't see themselves as technical.
On robots, I suspect general purpose robots are not so popular for the same reasons that the industrial revolution didn't happen in Roman times - human labor is too inexpensive. Factory labor in China and other developing nations is much too adaptable and inexpensive for robots to be worthwhile. But that will change.
I agree with you that it's a pretty long way out!
Pushing a car three miles back to your house is hard. So is beating Kasparov at chess. But clearly they are two very different kinds of endeavor, and I forget how to unambiguously refer to the two types of effort.
And I don't just mean "mental" vs "physical" - I'd place learning a martial art closer to beating Kasparov than pushing a car home.
Pushing a car is physically exhausting but doesn't require much training and practice. Programming requires training and practice, but I don't know that it requires more training and practice than, say, becoming a lawyer, doctor, martial artist, or other engineer. And the opportunities for practice come much earlier in programming training than lawyer or doctor training!
Though now that I think about it, farming involved more than just physically exhausting work - knowing what to plant, when, how to handle the soil, etc. - those all seem "hard" to me in a "training and practice" sense.
Farming of course involves physically exhausting work, and yes, that's the easy part (at least from an economic standpoint.)
("Take a right at where the barn used to be", "Go until you see the house with a big tree in front of it, you'll know what I mean, it's pretty big", "Take a left at the dog")
For what it's worth, whenever I'm in that area, I always notice the lighthouse mailbox now.
I don't think we'll all be sitting down in front of a C compiler, but I think a lot of people will spend their days configuring software to do things that it is not pre-programmed to do, using logic that is in line with the programmer mindset.
I like to look to agriculture, because computers are becoming quite prevalent in the industry and is a good indication of where the technology is headed in all industries that have not been traditionally computer dominated. The modern tractors are doing a lot of automation that is programmed by the farmer, depending on their own requirements. Like: "When the tractor gets to this position in the field, raise the implement. Lower it when it gets to _this_ position." That is programming.
I do agree we are a long way off from having general purpose robots, but we already do have robots everywhere we look. These agricultural tractors I speak of are robots that are being programmed, not operated, by farmers. That is just one of many examples.
Over time, interfaces become more abstracted. Who could have imagined Ruby on Rails in 1980? jQuery in 1995? Further abstraction opens programming to a wider audience. Programming will become increasingly diverse as programmers cover an increasingly wide spectrum of technical expertise.
We commonly talk of a 10x difference in ability between programmers (and I won't go into the truth of that here) but it's not unreasonable to expect that the difference between the best and the worst programmers will widen by another order of magnitude in the future.
Really the argument is semantics. Technology (and software programmed by genuinely skilled programmers) is making it easier and easier for people who would otherwise be poor programmers to achieve a result - even if it is a strange sort of chimera cookie-cutter application that only sort of meets their needs. (This is what Microsoft does with software for the enterprise.)
We'll all become programmers in the sense that we're all now omniscient due to the introduction of mobile phones with web browsers built in and the availability of search results on tap through said web browsers.
1. Dexterity is NOT the "last mile" gap to having robots in minimum wage jobs. I don't care if the robot can flip a burger, can it recognize a grease fire (and put it out, and clean up afterwards?) Can it recognize a moldy bun or if the special sauce is spoiled? While low wage retail jobs don't require years of specialized training, they do take advantage of a lot of innate human skills. Most of the mechanical aspects of these jobs are already automated.
2. Good software is not created by people sitting in cube farms programming. In projects of any size it's more about communication among developers and between developers and other business functions than writing code in isolation.
3. Not everyone can be a programmer, and most that could be won't want to (and I predict won't have to). I'm sure you can invent trivial examples that anyone can do, but then you're not saying much. There are large classes of people that can't program a VCR or do long division, let a alone construct a slightly involved program. As someone who's been involved in research that attempts to teach average people how to "program", I feel confident that we will never see a large portion of people become programmers.
1: One human is enough to put out grease fires, clean up, and find moldy buns. Most fast food joints have multiple humans, and dextrous robots would change that.
2 & 3: I don't think I made either assumption. Specifically, I didn't say we'll all be good programmers or produce good software. (Even today, not all programmers are good or produce good software.) I suspect what we think of as "programming" will change significantly, and with it, the number of people who program.
People's dumb blogs get the job done for those people, and the people who choose to read them.
Automation will drive more creativity, but not necessarily more programming jobs. The need for more automation jobs will drive needs for more programming skills. I think we are already there, we call it "technology". It's anything that makes use of the digital world to get improved results.
Maybe one day we'll all be unemployed and bored out of our minds. But we certainly won't all be programmers.
Second, many people are not skilled enough in math and logic to program well. If programming became as pervasive as math education, lots of people would forget it after high school or never really learn it. If it became as pervasive as writing (as some commenters mention), then it would exist in a similar state: a collection of highly talented authors and a mass of untalented writers (with some in between). So you'd have to have newer programming environments that let unskilled people get some useful tasks done reasonably well. We have a lot of work to do before programming environments get there, but some people can make interesting things with mangled Excel spreadsheets or spaghetti code in PHP or copy-paste-modify JavaScript hacks.
Third, also as some people mention, labor is cheap in certain parts of the world, so robots aren't economical. Contrary to what some claim, I don't think that will go away. There will always be some portion of humanity that is not sufficiently educated to be programmers. This could be due to socio-economic circumstances, or it could simply be because they don't want to learn it. Some people go through high school refusing to learn, others drop out. How do you fix that? If they can't get customer service jobs, then they have to sell their manual labor cheap enough to compete with robots. You'd only replace all that if you eliminate poverty on a global level and have excellent education globally too. Then there won't be cheaper labor, so those jobs will move to robots, and then there won't be enough people wanting to hire humans for labor and so finally that market would cease to exist. I'm not counting on that happening, not even in a hundred years and possibly not ever. With a concerted effort we might have a chance at a utopian-like society world-wide in a couple hundred years, but that's plenty of time for wars, natural disasters, apathy, tyrants, terrorists, or any number of things to mess it up.
Conclusion: Sure let's make programming environments more accessible to less-nerdy people, and let's improve education about programming, since these things will improve lives. But we won't end up with all programmers.