Forget 6-Minute Abs: Learn To Code In a Day 306
whyloginwhysubscribe writes "The usually excellent BBC 'Click' programme has an article on 'Why computer code is the new language to learn' — which features a company in London who offer courses on learning to code in a day. The BBC clip has an interesting interview with a marketing director who, it seems to me, is going to go back and tell his programmers to speed up because otherwise he could do it himself! Decoded.co's testimonials page is particularly funny: 'I really feel like I could talk credibly to a coder, given we can now actually speak the same language.'"
Re:language != logic (Score:4, Interesting)
Studiously refraining from teaching somebody any of that boosts their confidence in a way that only years, or even decades, of advanced study can hope to equal!
Incidentally, why doing you programmers just prove that your algorithms will never hang before shipping code? Are you lazy or something?
Re:language != logic (Score:1, Interesting)
They teach HTML, CSS and Javascript. Only need an hour and definitely no logic.
Re:language != logic (Score:3, Interesting)
"Incidentally, why doing you programmers just prove that your algorithms will never hang before shipping code? Are you lazy or something?"
Not sure if troll or not. If you are troll, color me an fool for biting.
As someone with formal comp sci training, and onward into graduate degree... let me give you a crude hint. Others can cite specifics if they want. But this is the answer:
In general, the problem you propose -- if solved, would immediately result in a million dollar reward for solving a millenium problem.
You could safely turn down this reward -- because you would have to publish your proof in an academic journal.
You'd probably also immediately get a fields medal and nobel prize.
A greedy person would take the process/result and encrypt it-- guarded very heavily as a trade secret. In fact, I'm pretty sure I'd try to hire mercs immediately and up front, and go into hiding. You would be able to use this to break virtually all modern encryption algorithms, most random number generators, and create the best compression system theoretically possible. Ever. In the known universe.
This problem is called solving L-Halt (halting problem). And if such a technique existed, and could be done in "reasonable" time -- such that it collapsed NP hard into NP into P ... it would be a breakthrough that truly redefines the meaning of "epic". The reasonable implications are truly earth shattering.
There *is* an algorithm to do it (not in reasoanble time or space). Sort of. The process actually grows faster than any computable number. I don't believe the answer is known (deterministically and in general) for any program with more than five instructions.
It's been an open problem since the 20's, and a decent group of scientists believe a proof of existence or non-existence is impossible. An even larger (possibly majority? Other experts can chip in) group believe if a proof is found, it will be that no such polynomial time algorithm can exist.
So... let's put it this way:
For specific pieces of code, such proofs may be done. I'm not an expert in it, but I've studied it very lightly. It's very nearly impossible for all but the most trivial softwares -- specifically, for anything but certain numeric algorithms. And if it isn't impossible... it takes a huge amount of time by hand. There's automated provers, but they are... basically infants.
The notion of "proof of correctness" is a formalism that would make modern software development practices very nearly impossible. And if it could be done, the costs would be so prohibitive that even hardware design would likely be cheaper -- because even chips don't have the same level of rigor in their analysis.
Are we lazy? Well frankly yes, a lot of us are lazy. But the proof you ask for... it's the holy grail of computer science.
And a real computer scientist would know that immediately. A real programmer -- they might dream of it.
Re:language != logic (Score:4, Interesting)
Yes, it's possible, it's always possible, it's a question of time and money.
Obviously, you've never had a marketing person ask for something that is so out of the ballpark that it would be an equivalent of solving "strong AI" problems ...
Heh. A team I was once on was asked to do a task that provably required an upgrade to the speed of light. It involved the time for getting messages between widely separated places on the planet. The managers couldn't accept that the universe imposes a speed limit on such things. It was clear that they understood this to mean that we weren't smart enough to solve the problem. We on the development team quickly updated our resumes ...