hvd

"the ORed result" -- Your answer is a good warning if your assumption here is correct, but I strongly suspect it is used as a concatenation operator here. | and || are fairly common ways of expressing that.

@MSalters You didn't just write about the right to an independent arbiter though, you wrote that you have the right to an independent judge. I interpreted "judge" as the profession, did you just mean it as the ordinary English word? If you did, then I do not see from the SC decision why SE's pick would necessarily not be considered independent.

@Adriaan As would I, but that doesn't matter if the court reads it differently. Quoting from the court ruling: "Hoewel het bepaalde in art. 6:236, aanhef en onder n, BW erop wijst dat de Nederlandse wetgever het arbitragebeding in een consumentenovereenkomst niet als een oneerlijk beding in de zin van de Richtlijn lijkt te hebben aangemerkt, [...]" (emphasis mine)

I believe this answer seriously misrepresents the case it references. Read it online (Dutch): ECLI:NL:HR:2012:BW6135, voorheen LJN BW6135, Hoge Raad, 11/04598 EU Council Directive 93/13/EEC suggests disallowing mandatory arbitration. Dutch law (art. 6:236 BW) would mostly allow it. The SC ruling (3.4 & 3.5) says a hard yes/no is wrong, each case must be determined individually.

@Ewan Exactly the same way this answer suggests checking beforehand can be done afterwards. That can either be done by manually writing out checks for all relations, or using EF's model and looking for all navigation properties.

"A. What if there were two FK constraints. Would you advocate parsing the exception message to work out which object was missing?" -- No, I would check the error code. If the error indicates a foreign key error, and I can actually handle foreign key errors (even if by highlighting the input field for the benefit of the user), and there is no official API for retrieving information on which foreign key was violated, then I would perform the checks you are performing beforehand.

@Darkwing "Not to mention that the situation - many men, few women, the women being addressed," -- That is one of the facts under dispute though. From the article: "He said it was a man, not a woman, who asked for the floors [...]." (And this doesn't mean either of them is lying: it's possible he honestly but wrongly thinks it was a man who asked the question.)

@supercat I'm sure there are better approaches, but one sort-of obvious way would be to add a unique constraint on x, and a separate unique constraint on a computed column with a value of CASE WHEN x IS NULL THEN 1 END.

@supercat That's begging the question. Yes, if you start from the position that any row in a key column must have a value that is equal to itself, then of course you will conclude that for that constraint, NULL should compare equal to NULL. That's just a re-phrasing of your starting position. You've redefined "unique" to something that not everyone will agree is a good definition.

@supercat But that's just it... if you have NULL, NULL, 1 and 2, you don't have any value which compares equal to any other value, so there is some logic in saying that all values are unique.

@supercat If you have two values that compare equal, then one is not necessarily a copy of the other. Whether a value is unique and whether it is a copy are two completely separate issues. "Duplicate" is confusing because it is used in more than one sense. It's sometimes used to refer to a copy, and it's sometimes used to refer to a value that happens to compare equal to another existing value. In the first sense, duplicates don't prevent uniqueness if copies don't compare equal. In the second sense, they do, but in that second sense, you don't have any duplicates if you copy nulls.

@2501 It's never been made explicit, confusingly. The logic used is that the standard describes how to build correct programs. One of the ways it describes leads to #define f() / f(), where f() has no arguments. Another of the ways it describes leads to #define f(...) / f(), where f() has one argument. The fact that there is also an invalid build leading to #define f(...) / f() is not the programmer's problem, but the compiler's: if there's a parse that leads to it being valid, and one that leads to it being invalid, it must choose the one that makes it valid.

@2501 That other case I mentioned, #define f(a, ...) and then f(a), that's a common extension that people sometimes use without realising it's an extension, which can then trigger diagnostics on other compilers, or even on the same compiler with the right command-line options. I suspect that's what you remember seeing. :)

@2501 See the C99 rationale, page 102, by the way: it's more explicit about this than the actual standard. It reads "A function-like macro invocation f() has the form of either a call with no arguments or a call with one empty argument. Which form it actually takes is determined by the definition of f, which indicates the expected number of arguments."

@2501 6.10.3p4 says the macro f may not be invoked with 0 arguments. But that's what I'm saying: it's invoked with an argument. The fact that that one argument is empty is irrelevant. The point of 6.10.3p4 is to reject #define f(a, ...) / f(a). In that case, there is only one argument, but the macro requires at least two.

@2501 Yes, it is. That's one of the new features of C99: in C90, empty macro arguments were undefined by omission. C99 explicitly supports them.

I don't see why it's a problem that int b = f(5)(3)(2)(0); failed. f(5)(3)(2)(0) was a valid expression which produced the correct value. It's just that it was used in a context that didn't allow arbitrary expressions (which, admittedly, may be most contexts).

@2501 True but misleading. Given #define f(...) whatever, the macro invocation f() is just as valid as f(1) in standard C99/C11. Both invoke f with one macro argument, which in one of the two cases happens to be empty.

Ah, I missed that, but that's trivially avoidable: don't use the function result. Change | to , to achieve that. :)

There's no need to set i to 0 at the end of a function if it's function-local. 86 bytes if you get rid of both.

@Andrew Yes, I understand that. "All men" or "All people" as I suggested are plural forms, which avoids the problem, because then "their" becomes unarguably grammatically correct.

The gender neutral version can be turned into a plural form ("All men/people should love their wives/spouses.") to avoid the grammatical problem. It doesn't avoid the problem (same with "Everyone") that it explicitly addresses all people regardless of marital status though. The OP's "One should love one's wife" avoids this by implicitly addressing only people who have wives.

@JoeBlow I'm not sure that's right. At least equally believable would be that the question's quote and any particular book use "hack" in the same sense with the same origin.

I don't see that as a meaningful question. Were you expecting a response to that other than "huh?" ?

@Deusovi No, I think "a member of the set of things that "are" 0/0" was a fair way of describing it. Which does indeed imply all numbers "are" 0/0.

Now there's no need for that.

I think that's as good a way of describing it as any, @immibis. I think I see your point of view, and while it contradicts what I've been taught, it does make perfect sense. It's possible that you've been taught differently, and I'll even readily admit it's possible that what you've been taught is the norm...

@Deusovi That's the English "is", which does not imply equivalency and which I had not intended to imply equivalency. I thought we already covered that.

@Deusovi When did I say that?

@immibis So you're saying, correct me if I'm wrong, since we're writing 0/0, the fact that this could have arised as a result of a limit should be ignored. If it's a limit that would resolve to 0/0, and it's written in a form that no longer conveys that limit, it's written wrong?

@Deusovi I assure you, I'm well aware of that.

@immibis I can't disagree with that.

@Deusovi And to adjust it to what I wrote, how about lim(x->0) (1 if x=2x, 0 otherwise)? I don't know how to format it properly, but I think you get the idea.

And I suppose that's where you say limits have to be specified explicitly: x=2x does not hold as x approaches zero, it only holds if x equals exactly zero.

@immibis I would say no. I do see where you're going with this. I'm having trouble seeing at precisely which point it's going to go wrong.

@immibis I'm not 100% sure, but I think I probably would say x/x = x/x at x=0.

@Deusovi A cat is an animal. A dog is an animal. A cat is not a dog. I didn't claim 1=0/0. I used "is" there, so when I later wrote "I wouldn't dare claim 1=0/0", that's not a contradiction.

@immibis Sorry, it takes time to think and write a comment. Please give me enough time. :)

@Deusovi "is" in the English sense does not imply equivalence.

@immibis I wouldn't dare claim 1=0/0, 2=0/0, therefore 1=2 :)

@immibis It depends on where the expression 0/0 comes from. Which typically means turning it back into a limit.

@immibis Not enough information is available to answer that question.

@immibis Yes, that indeterminate form and undefined are two different things.

While the links you sent don't contradict your point of view (and I'm not precisely sure what that is), they don't contradict mine either.

0/0 can come up in ordinary mathematics, where it doesn't have any specific value. Limits can be used to determine what value it should have in that particular case, or some other calculation that avoids the division can be used instead.

I think we're mostly in agreement about indeterminate forms, except for a slightly different starting point in looking at it.

Undefined in the mathematical sense pretty much means it's never valid is what I've been taught, and the only one of the links you've given that contradicts that is the Wikipedia page.

@Deusovi "It's an indeterminate form" means something different from undefined. mathforum.org/library/drmath/view/57322.html Are you using "undefined" in its ordinary English meaning? If so, you're not wrong, but the argument is somewhat pointless. I'm willing to concede that there are dictionaries out there that provide definitions of "undefined" that agree with you. But in mathematics, undefined and indeterminate are different.

@Deusovi If you're going to ignore a request for clarification and an external reference to just repeat your previous assertion, there's no point in continuing this.

@Deusovi No, it's not. n/0 is undefined for n!=0, but 0/0 is indeterminate. Undefined means it never gives a number. Indeterminate means that it can give different numbers in different contexts. For instance, the limit of n/n as n approaches zero is 1, and also 0/0. The limit of 2n/n as n approaches zero is 2, and also 0/0.