Timeline for Dealing with "Find out who's going to buy the croissants"
Current License: CC BY-SA 3.0
27 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 26, 2013 at 0:18 | comment | added | Maks | @Sklivvz I'm sorry but if I had enough rep here on meta, I'd give you a down vote for making the assertion that its a simple question with the only evidence being a hand-wavy answer about "just pick a person and redo if unacceptable" and then following up with "I honestly think". Theres clearly enough contra views and evidence to show that its not. | |
| Jul 25, 2013 at 8:03 | comment | added | Sklivvz | Still, there's no complexity hidden in my answer. I honestly think that the question is simple, the correct answer is simple and it uses basic probability theory. Now, I'll agree that some people don't understand the basics of that (gambler's fallacy), but, there's nothing really to it. | |
| Jul 25, 2013 at 7:50 | comment | added | Rex Kerr | Wellll I admit it was chosen for dramatic effect, but it's an apt analogy. Knowing who you want and who you don't is kind of the whole question. How about the quicksort analogy: it's simply "pick a good pivot, then sort greater and lesser on either side, and recurse." Um, yes, but how do you pick a good pivot? | |
| Jul 25, 2013 at 7:42 | comment | added | Sklivvz | @RexKerr that's an unfair statement, come on :-) | |
| Jul 24, 2013 at 20:06 | comment | added | Rex Kerr | @Sklivvz - And the key to a good life is "don't do things that lead to a bad life". It's extremely simple. Or maybe having a magic filter is not an actual solution. (Don't forget e.g.--"Generally speaking, everyone should buy croissants as many times as the others".) | |
| Jul 24, 2013 at 19:54 | comment | added | Sklivvz | @RexKerr the answer is simply "pick one at random, until it's not a one you don't want". It's extremely simple. If people complicate the issue it's because it's probably too simple to see... which again doesn't make it good, or complex, or a viable question. | |
| Jul 24, 2013 at 19:42 | comment | added | Rex Kerr | @Sklivvz - It doesn't follow in every case, but in this case there were a bunch of answers that failed to do what the OP asked for, probably because the underlying problem--of having both randomness and fairness is not actually straightforward. HevyLight's 6th top rated answer is the first that answers the question, and he admits that it has a sizable defect--that you can't control the degree of randomness--which to fix is "complicated" and best attempted by someone "well-versed in probability theory". I think the issue is that the problem is too cutesy for people to use their heads! | |
| Jul 24, 2013 at 19:38 | comment | added | Shog9 | @Asad: I'm still struggling to understand what rule you're applying here. It sounds like you're settling on a standard that allows questions on algorithms only if those algorithms are 1) already commonly-known (so, if it's in a CS textbook I can ask about it) or 2) clearly intended for use in software but not application software contexts (comparing sets of abstract data types is good, comparing contact lists is bad). This strikes me as a brutally difficult rule to put into practice - and one that prioritizes homework questions over real problems people face. | |
| Jul 24, 2013 at 19:10 | comment | added | Sklivvz | @RexKerr I am not sure it follows. If I ask a hipothetically trivial question and I get a bunch of wrong answers does it follow that my question is non-trivial? I think it follows that the question is badly posed, maybe, but not necessarily non-trivial. A complex question probably will have few answers, of non-trivial nature. | |
| Jul 24, 2013 at 19:04 | comment | added | Rex Kerr | The "super simple" question has very few correct answers. Almost all fail to meet the requirements without obvious ways to fix them (that don't create yet more edge cases). The top five answers have essentially no randomness beyond perhaps putting names on in the list unsorted. You can't claim it's trivial if most everyone gets it wrong! (It may be a bad question, but not because it's easy or obvious.) | |
| Jul 24, 2013 at 17:38 | comment | added | user200500 | @Shog9 I'm not sure what you're saying here. The binary search algorithm is another algorithm most commonly seen in computing contexts. As such, a question about how the binary search algorithm applies to leveling sensitive equipment by hand would potentially be useful on a site about programming, similarly to a question about using TS to map routes to cities. I'm saying these types of algorithms (by virtue of being a frequently used programming tool in and of themselves) are an exception: it doesn't matter whether they're operating on elements of software or day to day objects. | |
| Jul 24, 2013 at 17:25 | comment | added | Shog9 | My original example was the common binary search, @Asad. | |
| Jul 24, 2013 at 17:22 | comment | added | user200500 | @Shog9 Yes, because most non programming algorithms cannot be as easily repurposed for non trivial programming problems. It isn't really a fine hair to split unless you're deliberately looking at algorithms from CS theory (such as traveling salesman), which originate in problems dealing with computing and have merely been analogised for ease of understanding. | |
| Jul 24, 2013 at 17:12 | comment | added | Shog9 | That seems like a really fine hair to split, @Asad. So the travelling salesman problem is OT if I'm actually mapping routes to actual cities, but on-topic if I'm working with an object graph? | |
| Jul 24, 2013 at 16:30 | history | edited | Sklivvz | CC BY-SA 3.0 |
added 762 characters in body
|
| Jul 24, 2013 at 16:16 | comment | added | user200500 | @Shog9 The reason I've been saying this isn't a software algorithm is because to my mind, a software algorithm needs to operate upon elements of software, instead of simply being capable of demonstration/implementation as a software package. An algorithm for calculating optimal moves in a chess game, or an algo for stock prediction number crunching is not a software algorithm, regardless of being an algorithm that can be implemented as software. | |
| Jul 24, 2013 at 16:13 | comment | added | Sklivvz | Also, it's a "As a programmer, how can I choose the next guy who buy croissants" question. Boat programming | |
| Jul 24, 2013 at 16:10 | comment | added | Sklivvz | To be fair I never argued that any of the points I make are enough to close by themselves. I do believe that altogether they don't paint a good enough picture to keep this question though. | |
| Jul 24, 2013 at 15:59 | comment | added | Shog9 | Let's focus on that then, @Asad. Much better argument than "lack of effort", IMHO - or the slipperly-slope of "if it doesn't require a computer to execute then it's not a software algorithm" (who here has never done a binary search by hand?) | |
| Jul 24, 2013 at 15:56 | comment | added | user200500 | @Shog9 Yes. The requirements are poorly specified (the algo must be random, but also "fair", for varying interpretations of fair). The primary discussion on an SO question shouldn't be about ethics. If it is, IMO that is a symptom of something being wrong. | |
| Jul 24, 2013 at 15:46 | comment | added | Shog9 | Multiple correct answers are neither the hallmark of a good question nor a sure indicator of a bad one, @Robert. A lack of criteria for incorrect answers comes closer to the real test there. Is this question too broad? Does it lean too heavily on subjective opinion rather than demonstrable fact? | |
| Jul 24, 2013 at 15:44 | comment | added | Sklivvz | 2) we have dropped too localised but this doesn't make this question valuable or general enough. 3) then close the other questions too as too basic. SO is for programmers not the general public. 4) see 3. 5) clearly you haven't read the answers ;-) 6) see 5. | |
| Jul 24, 2013 at 15:36 | comment | added | user102937 | @casperOne: 4) Wait, what? Many correct and equivalent answers is the hallmark of a good question now? I smell a new definitive book list question in the works. | |
| Jul 24, 2013 at 15:28 | comment | added | casperOne | 1) agreed, reason to close, not delete or migrate. 2) disagree - there's no "too localized anymore" 3) disagree - you don't need programming expertise in many algorithm questions, but we need algorithms in programming 4) disagree - This is the nature of many of the questions and answers we consider good on Stack Overflow 5) disagree - It's an algorithm question, code isn't required, but can be translated into code. 6) disagree - You don't substantiate this singular statement in any way and it's subjective at best. | |
| Jul 24, 2013 at 15:06 | comment | added | Sklivvz | I mean, when you get answers starting with "This is a much more interesting problem when considered as an economics problem as well as a programming one." you know there's something very wrong. | |
| Jul 24, 2013 at 14:59 | history | edited | Sklivvz | CC BY-SA 3.0 |
deleted 41 characters in body
|
| Jul 24, 2013 at 14:51 | history | answered | Sklivvz | CC BY-SA 3.0 |