Ideally such functions will contain several pages of incredibly complicated looking code that will be completely optimized away in release builds, but exhaustively execute for debugs and traces, and require months of careful human analysis before anyone else can be completely certain of that fact.
Depends on what exactly you're doing. As a general rule I prefer to avoid deeply nested code, but I've also written some code where a large block of code all interacted with a large amount of data, so that there were no natural "separation planes" to decompose it into smaller blocks without creating subfunctions that would themselves take dozens of parameters that might (or might not) be modified, making the whole even more error-prone and difficult to understand.
Not a common occurrence I'll grant you, but sometimes the task at hand really is that ugly.
I've also employed deep nesting in special purpose situations code where it could be naturally decomposed into subfunctions, but those subfunctions would themselves be extremely brief with near-zero chance of reuse. At that point the overhead of function decomposition can rival the time to actually get it working, so unless there's a dramatic improvement in clarity or I've got time to spare, I'm unlikely to bother.
Quite. Reminds me of the definition of "undefined behavior":
It compiles perfectly.
It debugs perfectly.
It works perfectly throughout all functionality and QA testing.
It explodes in the worst possible manner when proper functioning is actually important.
Certainly they do - one has a vested business interest in improving the perception of Autopilot's safety, and one just avoided facing manslaughter charges because of a feature of his expensive new toy. That doesn't mean either is being objectivly honest.
Clearly Musk is including the entire computer control system, including long-common safety features under the "Autopilot" umbrella, which probably makes sense from an implementation perspective, but means Autopilot is not synonymous with self-driving, and is being given credit for safety features that other cars have had for a decade.
Let me ask you this - aside from various degrees of self-driving, what other "convenience features" does Autopilot have that can be engaged while driving down the street?
Unfortunately for your argument, there's not really much evidence that raising the minimum wage causes problems - in fact pretty much everywhere it's been increased we've seen quite the opposite effect. Minimum wage salaries are typically only responsible for a tiny percentage of the cost of the products or services being sold, and the cost of doubling them can easily be passed on to the customer. Meanwhile, doubling the wages of a large fraction of the population considerably increases the size of the customer pool, and thus the available profits for the business.
As for "harming the rights of the employer to operate his business in an equitable manner", what world are you living in? The vast majority of the US economy consists of corporations that have pocketed 99% of all the productivity gains of their employees for the last several decades, while worker wages have remained largely stagnant. That's hardly operating a business in an equitable manner.
If you want to talk about the government immorally imposing demands on society - how about we start with the completely artificial strong property laws that allow a tiny percentage of the population to concentrate the vast bulk of societies wealth into their own hands over the course of generations? Without those laws individuals can only accumulate as much wealth as they can personally defend, insuring a far more equitable wealth distribution. If we're going to have such laws, then we need counterpoint laws to prevent the terrible excesses that they enable.
Of course slavery requires consent of the slave. A slave who refuses to work is no slave at all, only a victim of abuse and probably eventually murder.
That the violence in a rigged economy is homelessness and starvation not directly imposed by any specific individual does not fundamentally alter the fact that consent is often coerced.
Nope, they're just going to know a MASSIVE amount about your daily routine. There's a reason that putting GPS trackers on your car traditionally required a warrant.
Autopilot wasn't enabled - they're giving it undue credit by including under the same name the completely unremarkable active collision avoidance while under human control
I wouldn't say NEVER requires a human to take over, but like an aircraft autopilot, it should never require human input without allowing at least many seconds for a distracted operator to be alerted and assess the situation in order to decide on a course of action.
Of course, given the much higher rate at which problem scenarios develop for automobiles, that does mean that the autopilot should be able to handle pretty much anything you can throw at it at least well enough to stop and ask for further guidance. Which pretty much just leaves the really unusual scenarios for human input - the road is flooded, a bridge just collapsed in front of you, a herd of buffalo is wandering past, etc.
While I'm inclined to agree on grounds of being technically right, and have even argued the case previously, I do think there's a good argument to be made that an autopilot worthy of the name should be capable of safely handling all normal navigation challenges, including adapting to sudden unexpected problems in a manner that avoids making things worse long enough to give the presumably distracted human pilot time to assess the situation and take control. After all, a certain level of distraction must be assumed in any pilot whose attention is not actively required for normal operation.
For an aircraft, you've got lots of empty space around you to work with, and correspondingly large amounts of time to make decisions, and so the demands on the autopilot are relatively modest - maintain altitude and heading as best as possible, and don't do anything that would send it spinning out of control in the event of sudden turbulence, engine failure, etc. Any obstacles are going to be visible long enough for even a rather severely distracted pilot to notice them long before they become an imminent issue. Even automated landing systems, which are rarely used, have the luxury of assuming they're entering a well-controlled environment that has been adequately prepared for their arrival.
For a car though - the normal operating environment is a close-packed with chaotically moving obstacles constantly within a few seconds of impact. Most potential collision scenarios will have already resolved within seconds of emerging, long before a distracted driver can even hope to assess them. As such, the demands on an adequate autopilot are *far* greater. Even if you limit it to only highway usage, the number of expected "problem scenarios" per mile is radically larger than for an aircraft.
Eh, I suspect that, if they weren't totally oblivious, the pedestrian would be able to locate the emergency vehicle long before the driver - between not being encased in a sound-distorting canopy of glass and steel, and being far freer to swivel their head, they were probably able to dismiss it as an imminent threat almost immediately - after all, locating potential threats is one of the primary things our ears have spent millions of years evolving for.
They may not have considered the effect the siren would have on nearby traffic though. And considering they stepped out directly in front of a car at night, whose fast-approaching headlights were presumably clearly visible, I wouldn't rule out "completely oblivious" either.
I understand not reading TFA, but come on, even the summary specifically mentioned that Autopilot was disabled.
Okay, so they actually said the "convenience features" were disabled, but since they're giving credit to Autopilot for the basic collision avoidance system, I think it' safe to assume that actual self-driving is included in the "convenience features".
That would have to be a well-armed skateboarder, or ferocious bad luck.
If the normal behavior was, like the buffet, to purchase a package now and then, I wouldn't argue the point. But the normal behavior is a subscription style plan with automated monthly renewal, and all customers who have the plan have been consistently renewing their subscription for years. In that case, inserting an automated termination if you actually partake of the advertised services certainly seems a breach of good faith.
Yes, like the all you can eat buffet, they are presumably within their legal rights to refuse future business with those they feel they're getting a bad deal from - but I have never actually heard of a buffet doing that either - and I know people who will take a book with them in the morning and spend all day eating. You advertise "unlimited", you have to expect a few people to take advantage of you and factor that into the cost. If you're not willing to do that, then either clearly mention the limits in the ad, or expect public backlash for false advertising.
I mean seriously, it's not like replacing the Unlimited* plan with a "100GB" plan would inconvenience the vast majority of the customers subscribing to it - the only change for them would be losing the psychological appeal of the word "unlimited".
!07/11 PDP a ni deppart m'I !pleH