One of my most destructive bugs was because we were using a single variable to hold one value that was used to mean two different things. I was the lucky guy who wrote the line of code that modified the value to be correct for one use but not the other. When one is writing code that winds up in CNC machines, the results can be spectacular.

When micro-optimizing the code, you need to know what's good and what isn't. Old advice can screw up optimizations on more modern compilers.

For much of my youth, I was told that unrolling loops was good for performance, and quite a few years ago I sped up a slow section by rolling the loops up as tight as I could. My best guess was that my version was much more cache-friendly, but what I did know was that my profiling showed a very large speedup.

It's not just parameter evaluation, it's execution of side effects. If you write something like 'int j = ++i;", then you are modifying the values of two variables, and the Standard has nothing to say about the order of operations except that everything's sorted out at the next sequence point.

That's a very standard case of undefined behavior, and anybody with any real knowledge of C or C++ would know that this doesn't work.

Theoretically, the reduced test case is where you hand the bug over to whoever makes the compiler. Of course, there was the time when the response was, "Yes, that's a bug." Even a clue as to how to work around it would have been more useful.

One time, I had the reverse problem. Everything worked on my computer, but not on the users' computers, which were pretty close to identical to mine. That's even more frustrating.

You may not have noticed, but there are software systems that have been available for a few years that can manage different versions of a file that will have the same name. You can even revert to an earlier version when you like.

You might want to brush up on your history. By the time the feudal period started, the Visigoths were the establishment defending against raiders.

Becoming a company isn't that hard. You want to be a company, look up incorporation options in your jurisdiction.

So, once this has happened, the homeowner has the legal right to steal the intruder's stuff and rape the intruder's wife? Or that of somebody who looks like the intruder?

You do realize that poor people are going to suffer more than the affluent from climate change effects, don't you? Barring some sort of intervention, they're screwed either way.

As our tech gets better and better, we'll be able to find new ways to reduce CO2 emissions, and if it's increasing slowly enough we can use other means (such as trees) to reduce CO2 in the air. Therefore, the rate at which things get worse matters.

As a cynical liberal, you obviously don't know what you're talking about.

Businesses have only a very limited ability to pass costs to their customers. The reason is that the cost, in fields where there is competition, is pretty well set by a fairly standard supply-demand curve. If it becomes more expensive to produce something, the supply curve changes, but there's no bloody way a business can pass along costs and add their profit margin to it (unless you're talking about a contract specifically written that way).

If a business could just pass its costs onto its customers, then it could just raise its prices now and make more money. Do you think businesses in general keep their prices low and their profits low, for some reason? My observation is that they like profits, and set prices to their best guess on what to charge to make maximum profit.

Besides, if there's different ways to produce something, that produce varying amounts of carbon dioxide, the tax shifts production towards the methods not using as much carbon dioxide, reducing the amount of tax paid anyway.

We can estimate a 9/11 plane crash as killing about fifteen hundred people. There is one nuclear power disaster that has caused deaths on that order, which is Chernobyl. Chernobyl is, very simply, not happening again. (Who ever claimed that Chernobyl was absolutely safe? Do you have some sort of cite?)

Your other examples don't support your claims. Fukushima may or may not have killed someone, and few people not immediately involved in the cleanup have any significant harm. The exclusion zone is much less than a state, and it's real hard to poison something for a thousand years when you're working with a dangerous isotope with a half-life of about forty years. Three Mile Island was even more innocuous.

There are two reasonable definitions of the value of somebody's work. One is the amount you need to pay to get somebody to do it, and one is the amount it contributes to the employer's bottom line. There is a class of jobs that pretty much anybody can do, and unless we're at full employment those jobs are going to pay minimum. These jobs are often worth considerably more to the employer than the employer is paying. The immediate effect of raising minimum wage is that a few jobs will go away and the rest will pay better. The longer-range effects are much less predictable. For example, if an employer is paying more for an employee the employer might find it worth investing in making that employee more productive. This could spur the economy.

Do you have any evidence that the cost of living rises quickly? Most prices are tied to labor costs, to some extent, but in most fields the bulk of the labor costs are from employees making more than minimum wage anyway.