Talk:random number

Deletion discussion

The following information passed a request for deletion.

This discussion is no longer live and is left here as an archive. Please do not modify this conversation, but feel free to discuss its conclusions.

---

random number

Non-idiomatic sum of parts: random (adjective senses 2 and 3) + number. — Ungoliant ^(falai) 23:56, 20 July 2014 (UTC)

• Keep, since this usually can be found in dictionaries. bd2412 T 04:01, 21 July 2014 (UTC)
• Keep, it has a rather specific meaning in statistics. Ƿidsiþ 07:15, 21 July 2014 (UTC)
  • No, it does not. If there is a specific mathematical meaning, it is contained in random, but truth is, even mathematicians usually use random essentially like laymen, because there is simply no good definition of randomness. I am curious to know what User:Msh210 has to say about this; though as far as I am concerned, this is SOP. Delete. — Keφr 18:20, 23 July 2014 (UTC)
    • Thanks for the ping. I'm no expert, having hardly studied randomness. (I have studied random variables some, and I'd think off the cuff that random variable is not SOP. I haven't thought it through, though.)​—msh210℠ (talk) 18:34, 23 July 2014 (UTC)
      • Oh, there is no question whether random variable is idiomatic; the term refers to quite specific conceptualisation of an unpredictable outcome of some process as a measurable function on a probability space — and none of this follows from either word. WT:FRIED, in essence. — Keφr 20:13, 23 July 2014 (UTC)
• Keep as a mathematical term and per Lemming principle. --Anatoli T. ^{(обсудить}/^вклад) 07:25, 21 July 2014 (UTC)
• Delete. The "specific meaning" is already covered in the entry for random. The word "number" can be replaced with anything: random integer, random card, random person, random portion, random distribution. --WikiTiki89 18:29, 21 July 2014 (UTC)

Compare Talk:prime number, which passed. I would want to delete both. Equinox ◑ 19:17, 21 July 2014 (UTC)

The difference is that "prime number" is much more restricted. It is much rarer to find "prime" used attributively with something other than "number". Out of the things I listed above for "random", "prime" can only be used with "integer" and even that is rare, even if this is simply due to the fact that all primes are integers and thus specifying "integer" is redundant. On the other hand, "prime" is regularly used predicatively with words like "integer" ("this integer is prime"). --WikiTiki89 19:28, 21 July 2014 (UTC)

prime number is, quite undisputably, a term of the English language. random number is much more disputable, you are right, unless it's a "fixed term of art", as mentioned below. Lmaltier (talk) 21:31, 22 July 2014 (UTC)

• Keep: per Widsith. Purplebackpack89 20:27, 21 July 2014 (UTC)

Delete. The keepers aren't making a good argument here. Other dictionaries have it, has it ever been Wiktionary's goal to copy as much from other dictionaries as possible. As for a specific meaning in statistics, does it? What it is? The entry just says it's a number chosen at random. Renard Migrant (talk) 09:44, 22 July 2014 (UTC)

I suppose none of your would argue that this actually meet CFI, just that we shouldn't apply CFI to this term, right? Renard Migrant (talk) 09:46, 22 July 2014 (UTC)

Right, this is RFD, not RFV. Attestable but not inclusion-worthy, like the example of "brown leaf" at top of page. Equinox ◑ 11:11, 22 July 2014 (UTC)

In that case, why don't we, instead of having an entry, merely have a page listing other dictionaries where the reader can actually find a definition of this term. Something like, "We're sorry, Wiktionary doesn't have an entry on this term, but you can find it in Meriam-Websters, Collins, Oxford, etc." bd2412 T 13:12, 22 July 2014 (UTC)

A truly inspired idea: we could combine {{no entry}} and {{R:OneLook}}. DCDuring TALK 13:40, 22 July 2014 (UTC)

@BD2412 They can already find the definition on Wiktionary at [[random]] and [[number]]. --WikiTiki89 13:47, 22 July 2014 (UTC)

Wait a minute, DCDuring. Don't get the President yet. Maybe we should consider this for a second. Maybe, and I'm just spit balling here, maybe, we have a responsibility as dictionary writers to define phrases with meanings set well enough that other dictionaries have seen fit to define them. Maybe we as dictionary writers have a responsibility to this project to see to that it reflects the determinations of trained professionals. Yes, I'm certain that I read that somewhere once. And now I'm thinking, DCDuring, that your suggestion of deleting this entry, while expeditious and certainly painless, might not be, in a matter of speaking, the American way. Random number stays where it is. We're gonna define the phrase. bd2412 T 15:09, 22 July 2014 (UTC)

Nope. Renard Migrant (talk) 18:15, 22 July 2014 (UTC)

I take it that's a rhetorical "nope", since I don't see any practical basis for it. bd2412 T 01:43, 23 July 2014 (UTC)

I was just admiring your idea. I didn't vote delete. I'm inclined to keep an entry with one or more well-written definitions. The only thing that the existing definition has going for it is that it has been in Wiktionary for nearly 10.5 years. DCDuring TALK 17:36, 22 July 2014 (UTC)

• Keep: The "brown leaf" test is irrelevant because "random number" is a fixed term of art. If you are interested in a translation, you must use the specific phrase-based translation that already exists in the target language. An SOP-inspired translation based on "random" + "number" is going to be dicey. The existing translation table is quite handy. To rely on WP foreign language links to Random number is a poor substitute. For example, the table has French and Russian, but apparently no one has written the WP article yet in those languages. (Compare with nonsteroidal anti-inflammatory drug, which was kept as a fixed term of art.) Choor monster (talk) 12:30, 22 July 2014 (UTC)

  Brown leaf is a set term. Renard Migrant (talk) 18:15, 22 July 2014 (UTC)

  For what? Perhaps we should have it. Is it some synonym for tobacco? Choor monster (talk) 18:51, 22 July 2014 (UTC)

• I have a strong suspicion that random number is used most commonly to mean something very specific that does not follow from random and even number (at least by any of our definitions. In the normal speech of those normal humans that might use this, say, in talking about a lottery, or come across it in a magazine, number means a finite, context-determined subset of natural numbers and random refers implicitly to a uniform distribution over that subset. I think almost any other definition would need to be stated explicitly to such normal humans. Even among statisticians and applied mathematicians, random number by default refers to a number selected by a process that corresponds to a uniform distribution over a finite set of natural numbers. Thus I think we need the entry, though I can't say as much for the specific definition. DCDuring TALK 14:21, 22 July 2014 (UTC)

  "Random" does not necessarily imply a uniform distribution. You can have randomness with a normal distribution, or gaussian distribution, or really any distribution you want. "Number" is unspecific and is usually specified by the context; for example, in the context of the Mega Millions lottery, "number" means a natural number from 1-75. --WikiTiki89 14:42, 22 July 2014 (UTC)

  I know it doesn't necessarily mean that to sophisticates such as you, I think it does mean that to those normal people who come across the term and those who write for them. DCDuring TALK 14:52, 22 July 2014 (UTC)

  Can you give a specific example? --WikiTiki89 15:01, 22 July 2014 (UTC)

  Very possibly "large number" typically means different things to mathematicians (Graham's number) than to others (anything bigger than a few hundred?). I would hate to see an entry for it on those grounds. Equinox ◑ 14:57, 22 July 2014 (UTC)

  @Equinox What do our hates and loves have to do with the meanings of words as used in the language? DCDuring TALK 18:25, 22 July 2014 (UTC)

  Ignore the second sentence, focus on the first one, and drop the tiresome equivocation once in a while. Equinox ◑ 18:27, 22 July 2014 (UTC)

  I think equivocation is inevitable in trying to reach a conclusion in applying principles to concrete cases. DCDuring TALK 19:44, 22 July 2014 (UTC)

  It doesn't though. ‘Large number’ has no specific meaning in mathematics. ‘Random number’ does, it doesn't just meaning any number I choose at random, it means a number generated from a given set by a truly random process, such that – in the OED's words – ‘all the numbers in the set have the same chance of selection’. Ƿidsiþ 15:02, 22 July 2014 (UTC)

  A random element of a set does not have to be a number. You can have a set of elephants, for example, and choose one using a truly random process. There is nothing special about the word "number". --WikiTiki89 15:11, 22 July 2014 (UTC)

  Note that, for example, "Large cardinal" and "Geometry in the large" do have meaning within mathematics. In each case the ordinary sense of large is merged into something more domain-specific. In these two examples, the meaning is actually non-precise, with an aspect of "I know it when I see it". In contrast "Huge cardinal", has a very precise meaning.

  As to "random elephants": that's something of a red herring. When people study "random surfaces" or "random graphs" or "random flows", they are not taking any random meaning of random and applying it to the specific concept, the way you are with "random elephants". They are referring, in these cases, to a precise sense of random that leads to interesting notions worth studying in each case. Someone who comes late to the game and wants to use a different notion of random will typically use a different word. We have things like "Markov fields" and "stochastic processes". And some instances are even more extreme, like "random reals", where the sense of random is 100% technical, not covered under anybody's dictionary or encyclopedia definition, yet somehow aptly named. Choor monster (talk) 16:24, 22 July 2014 (UTC)

  But this isn't about "random surfaces" or "random graphs" or "random flows" (which may be worth including), this is about "random numbers" and you have failed to show how "random number" is any more specific than "random" + "number". --WikiTiki89 16:47, 22 July 2014 (UTC)

  That's a different issue. Objecting to a lousy definition is not grounds for deletion. But I've made a stab at fixing the definition up. Choor monster (talk) 18:51, 22 July 2014 (UTC)

  At RFD we discuss the definition that is there. We are not mind readers. --WikiTiki89 19:03, 22 July 2014 (UTC)

  You could do that, and maybe if people did this a lot then ‘random elephant’ would be a set phrase as well. But they don't and it isn't. Ƿidsiþ 17:38, 22 July 2014 (UTC)

  I think you're missing the point that in mathematics this is done with many different things other than just numbers, even if not with elephants. --WikiTiki89 17:42, 22 July 2014 (UTC)

  The point is this isn't really about mathematics here, although we mathematicians like to think we own the term. "Random numbers" have penetrated outside mathematics—the worlds of gaming and computer security—unlike "random matrices" or "random walks" or any of the other examples I mentioned above. These other notions are subject to user-redefinition as needed in technical context, "random number" less so. If you're doing Gaussian, you say Gaussian. See [1] for an example. Choor monster (talk) 18:51, 22 July 2014 (UTC)

I remember watching a documentary in the UK simulating a trial (this is because filming and rebroadcasting a legal trial in the UK would be illegal). The defence lawyer I remember said that his defendant had no defence in law but he could still try and get the defendant off because of the human nature of the jury. This is what this debate reminds me of. There's no plausible defence for this entry but if it just a matter of blind voting it might survive. Fine, but at least have the guts to admit you want to keep this for reasons other than CFI. Renard Migrant (talk) 18:12, 22 July 2014 (UTC)

This is unbecoming trash talk. Choor monster (talk) 18:51, 22 July 2014 (UTC)

I wouldn't call it "trash talk", but I agree that these kinds of statements are unhelpful. --WikiTiki89 19:03, 22 July 2014 (UTC)

• @Wikitiki89 So therefore among probability theorists and applied mathematicians who work with stochastic processes, probability distributions, etc, random number is not a set phrase or idiom. We probably need a usage note for that or perhaps a definition line as follows:

1. (mathematics) Used other than figuratively or idiomatically: see random,‎ number.

For others random number seems a set phrase with various possible definitions, including "a number provided by a random number generator (in turn defined as "usually an algorithm that generates pseudo-random numbers")". DCDuring TALK 18:25, 22 July 2014 (UTC)

So what will the supposed non-idiomatic definition be? --WikiTiki89 18:29, 22 July 2014 (UTC)

@Wikitiki89 I've provided two so far, as well as the &lit line that follows from your helpful drawing of our attention to the range of collocations in mathematics. The definitions that other dictionaries have are satisfactory to me, though they make it harder to make a clean case for inclusion in this kind of free-for-all. DCDuring TALK 19:29, 22 July 2014 (UTC)

@Renard, the CFI is there to codify our own instincts about what terms should be included. Not the other way around. The CFI have been, and may still be, reworked to fit our collective judgement. In my opinion, ‘random number’ is a set term, and I am far from alone in thinking this. In fact it seems to me that the editors of almost every major dictionary feel the same way. It is used by statisticians as a single entity, and I think of it as being ‘one’ term, not two stuck together. This is not very scientific, but it perhaps helps explain to you where I am coming from. What I don't understand is how – faced with many people who see the value in this – you think Wiktionary would be in any way improved without it. If you don't find it useful, you don't have to look it up. Ƿidsiþ 18:56, 22 July 2014 (UTC)

• See my revised definition. The point is that in "common" usage, random here means "discrete uniform", which fact is definitely not SOP. Choor monster (talk) 18:51, 22 July 2014 (UTC)

  If what you say that "in 'common' usage, random means 'discrete uniform'" is right (which I still disagree with), then this "common" definition should be added to [[random]]. And either way I don't see what this has to do with the word "number". Anyway, I think the common usage definition of "random" is more like "unpredictable" and has nothing do with any type of distribution. --WikiTiki89 19:03, 22 July 2014 (UTC)

  What some of us who want to keep this are saying is that it is not possible to write definitions in a given usage context and register of at least one of random or number (let's say random) that quite covers the actual usage of the words in random number without requiring that the definition of random be more or less restricted to use with number and its synonyms and hyponyms. DCDuring TALK 19:44, 22 July 2014 (UTC)

  I get that, but I don't think that is true. --WikiTiki89 19:46, 22 July 2014 (UTC)

  Well, what would the definition of random look like? DCDuring TALK 20:00, 22 July 2014 (UTC)

  For any definition of "random number" you give me, I'll give you back a suitable definition of "random" and "number". All I need is for you to specify which definition you want me to split. --WikiTiki89 20:49, 22 July 2014 (UTC)

  There are now three in the entry. Stop dithering and start defining. DCDuring TALK 21:32, 22 July 2014 (UTC)

  Those are all straightforward, I thought you had a different definition in mind. From all three of those, you remove the word "(a) number" from the definition and you have your definition of random. Note that I also dispute the accuracy of the "uniform distribution" definition, but that's irrelevant to this point. --WikiTiki89 01:59, 23 July 2014 (UTC)

  Regarding your request several indents back/comment here, regarding "uniform distribution" sense to random: it's been there since 2005, when it was added as #1, and it is still there, part of #1 among Adjective meanings. Choor monster (talk) 15:03, 23 July 2014 (UTC)

  In that case how is your definition of "random number" not SOP? --WikiTiki89 15:07, 23 July 2014 (UTC)

  As I've explained: the "uniform distribution" (sub)sense is not a priori obvious, especially to those with more than a little education. (The fact that it is typically the smart choice, a point you made earlier, is irrelevant.) Choor monster (talk) 11:56, 24 July 2014 (UTC)

  Which is why it is defined at [[random]]. --WikiTiki89 13:56, 24 July 2014 (UTC)

  Yes. Which isn't all that helpful when identifying which uses of random are part of set phrases and which are not. Choor monster (talk) 14:02, 25 July 2014 (UTC)

• I have restored the original definition and marked it with {{rfd-sense}}. Could we either

1. close this out by deleting the original definition and sending the entry to cleanup or RfV OR
2. have separate RfD discussions for each suggested definition, freezing each definition until discussion thereof is concluded?

Definitions are not really independent of one another, but we can probably have more focused discussions if we treat them as completely independent until a consensus on keeping or a specific change has emerged. DCDuring TALK 20:00, 22 July 2014 (UTC)

• I propose to restore the entry to this revision. The changes made after that seem pretty ridiculous to me, almost certainly not based on evidence. --Dan Polansky (talk) 20:43, 22 July 2014 (UTC)

  I agree with Dan that if we are to keep this, it would have to be the definition at this revision, possibly with some minor changes and with the addition of an {{&lit}} sense. --WikiTiki89 20:49, 22 July 2014 (UTC)

  Feel free to RfD or RfV each sense. Feel free to improve the definitions at random#Adjective and, if necessary, number#Noun. Or stand with the current definitions at those entries, or find definitions at other dictionaries that suit your purposes. DCDuring TALK 21:32, 22 July 2014 (UTC)

• I've included three citations of random number meaning one selected based on a discrete uniform distribution. I've also altered the definition to refer to "continuous", there are numerous citations for that too that I haven't bothered to enter. I voted "Keep", but I wish to clarify this refers to the two new definitions only. The original was borderline SOP, but the &lit handles that. Choor monster (talk) 12:09, 23 July 2014 (UTC)

• I've seen two users state they don't like my revision. All I can say is you two simply do not know what you are talking about. People instinctively think "random", when not being used to mean "arbitrary", refers to a uniform distribution, and this has been very well documented, most famously regarding the Monty Hall problem. The most interesting example I've come across was when I once consulted for NASA to help check and doublecheck simulation code, and at some point came across a mildly complicated probability calculation where the engineer had made exactly this mistake. Fixing it was easy. The hard part was figuring out a way to soundly convince the other engineers I was correct (which I did before I told them, precisely because this mistake is so notorious). Choor monster (talk) 14:02, 23 July 2014 (UTC)

  Would people think that the sum of two or more dice rolls (which is used in many board games and does not have a uniform distribution) is not random? And as for the Monty Hall problem, notice how the wording of the problem ("Suppose you're on a game show, and you're given the choice of three doors: Behind one door is a car; behind the others, goats. You pick a door, say No. 1, and the host, who knows what's behind the doors, opens another door, say No. 3, which has a goat. He then says to you, "Do you want to pick door No. 2?" Is it to your advantage to switch your choice?") does not use the word "random" at all and yet people still assume a uniform distribution. This is because if you do not know how the door was selected, even if you know it was not random, you still have to assume a uniform distribution. --WikiTiki89 14:05, 23 July 2014 (UTC)

• I deliberately said "based on", not "is", a uniform distribution. That includes dice, as mentioned in my Usage Note.

  Regarding the MHP: HUH??? If one particular wording does not use the word "random", so what? My point is that people typically interpret everything as following a uniform distribution, for example, when asked to choose which of two doors hides the car and which hides the goat. I'm mentioning this as a FYI, not as a potential citation per se. Choor monster (talk) 14:40, 23 July 2014 (UTC)

  Any probability distribution can be "based on" the uniform distribution. If "people typically interpret everything as following a uniform distribution" then it has nothing to do with the word "random" and even less so with the phrase "random number". --WikiTiki89 14:52, 23 July 2014 (UTC)

  This isn't mathematics.

  But there's no easy way to indicate at random which combination terms with it are going to implicitly assume uniformity, it's best to make the indication at such terms. That's why it's not SOP. In other words, we don't define "random X" based on people's misconceptions about randomness, whether or not the term accords with or contradicts them Choor monster (talk) 15:19, 23 July 2014 (UTC)

  You seem to be contradicting yourself. Do "people typically interpret everything as following a uniform distribution", or only for certain "combination terms"? --WikiTiki89 15:43, 23 July 2014 (UTC)

  You are mixing apples and oranges. The everything remark was about how people get it wrong, the combination refers to what people actually say, right or wrong.

  Speaking in exaggerated caricature, the mathematician would prefer to keep control of mathematical vocabulary. Listening to laymen getting it all wrong every day, when there's no point of a correction even getting through, breaks my heart. And no dictionary cares, either.

  In this case, popular misconception has led to "random number" not being the mathematically correct SOP, but the bastardized "uniformly distributed random number" special case. So it goes, and our remit here is to document this. It has not led to a bastardized meaning for other "random X" combinations that I know of. "random walk", for example, doesn't seem to have suffered out in the wild.

  So long as the highly technical mathematical term "real mouse" never gets misused, I'm happy. Choor monster (talk) 16:52, 23 July 2014 (UTC)

  What exactly do people get wrong? Without any prior knowledge of a given situation, uniform distributions are the best assumption. Anyway, the layman's definition of "random" is nothing other than "unpredictable", and has nothing to do with any sort of mathematical notions such as "uniform distributions". --WikiTiki89 17:19, 23 July 2014 (UTC)

  What they get wrong is that they equate "random" with "follows (or is based on) a uniform distribution". If loaded dice are used in a game and is generating numbers according to a non-uniform distribution, the results are still "random" in a mathematical sense, but common usage will describe the result as non-random. Here are three from the first page of Google hits for "non-random dice":

  • • "Folks who spend time at the craps table agree that eventually the total number of each number may well tend to equal out. They notice, however, for the smaller segment of the time they are there, runs and tides of number patterns seem to take over. Finding those currents of non-equal, non-random dice rolls is what the seasoned craps player is all about." [2]
    • "But some dice really do produce better results, since mass-produced dice never can be 100% truly random." [3]
    • "In fact, if you have dice at home, there's a good chance that they will also give you a non-random distribution. Dice often have the pips formed by a dimple in the surface. This means that the face with 1 pip weighs more than the face with 6 pips. This will skew your results in the long-term and that's why dice often have the pips printed rather than gouged out."[4]

  The "correct" meaning, of course, is that the dice are being physically controlled in some manner, not that they are biased away from a perfect uniform distribution. Similarly, you can find discussions on whether physical coin flips are really "random"—it's been "proven" that there is a teeny bias towards a caught flipped coin being the same side it was started at. (Yet another famous Persi Diaconis discovery!) Choor monster (talk) 17:56, 23 July 2014 (UTC)

  I think it can be captured by a sense of random synonymous with "unpredictable". Nothing about it is specific to numbers. (By the way, I think prime number is simply a WT:JIFFY.) — Keφr 20:13, 23 July 2014 (UTC)

  The whole point is that it is not clear which possible sense one ought to reply. For example, consider the following:

  "What is a flash mob? It’s when a random number of people start dancing a choreographed routine with it gradually involving more and more people from the unsuspecting crowd." [5]

  I've read somewhere that flashmobs follow a Lévy distribution, but the speaker above did not have that in mind. Choor monster (talk) 11:56, 24 July 2014 (UTC)

  If I say something's a bad idea, am I saying it's "Not suitable or fitting", "Evil; wicked" or "(slang) Fantastic"? Does that mean we should have an entry for bad idea? I wish we could retire "how can it be the sum of its parts if we aren't sure which parts" as an argument. After all, users are still going to have to make the same choice, either in the form of which combination to choose of the senses for each of the separate entries or in the form of which sense of the combined entry to choose. Chuck Entz (talk) 13:53, 24 July 2014 (UTC)

  You're just proving yourself wrong here by showing that any sense of "random" can be behind the meaning of "random number". --WikiTiki89 13:56, 24 July 2014 (UTC)

  Uh, no, I'm proving myself right. The phrase "random number" has two particular set phrase meanings (#3,#4), and one free-for-all meaning (#1). The free-for-all meaning can be rendered in all sorts of ways, in English or in translation, like "bad idea". The set phrase meanings can not. Choor monster (talk) 14:02, 25 July 2014 (UTC)

  For ease of understanding, I have created uniform distribution. Cheers! bd2412 T 16:11, 23 July 2014 (UTC)

• I added a context label. Mathematical terms often have different meanings depending on whether they're used in lay or mathematical contexts, so a mathematical definition should always be labeled as mathematical. Chuck Entz (talk) 14:18, 23 July 2014 (UTC)

  Your use of the label is backwards. I gave a mathematical definition for the lay usage, probably because I don't know any better. In a mathematical context, the distribution is spelled out or otherwise known from context. Choor monster (talk) 14:40, 23 July 2014 (UTC)

• IMO delete as SOP per Wikitiki (18:29, 21 July 2014 (UTC)).​—msh210℠ (talk) 18:34, 23 July 2014 (UTC)

Per WT:Lemming principle: definition of random number in Collins dictionary. (This is for the entry as a whole, not specific senses)--Anatoli T. ^{(обсудить}/^вклад) 00:31, 24 July 2014 (UTC)

I find it hard to defend the Collins definition: "any of a sequence of random numbers". As they do not define sequence of random numbers, it is circular. I don't know how it got past their editorial process. I don't think that the lemming principle requires that we discard all of our other principles. IOW, we shouldn't follow a rabid lemming. DCDuring TALK 01:11, 24 July 2014 (UTC)

Well, definitions don't have to be the same, in fact, some people strongly object copying definitions, I provided the link not for the purpose of fixing ours. The important fact is that a term exists, even if one can argue its definition. The lemming principle doesn't imply having the same info as the other dictionary but saving people on useless RFD discussions, when time is better spent on something more productive. --Anatoli T. ^{(обсудить}/^вклад) 01:21, 24 July 2014 (UTC)

Sometimes lemmings disappoint me. Not that this is not a difficult area for lexicographers. DCDuring TALK 02:26, 24 July 2014 (UTC)

I don't know if other dictionaries have an idiomacity rule; I trust that their authors, with the limited space they have to print, feel that the words they include are worth defining in a dictionary, perhaps even if they are idiomatic. The ultimate goal is not to be good rules-followers, but to help the reader. bd2412 T 03:32, 24 July 2014 (UTC)

MW Online seems to be the most exclusionary of multi-word expressions. I say seems because I thing their paywall may hide some idioms from OneLook. I need to test that. DCDuring TALK 03:58, 24 July 2014 (UTC)

Among the considerations that other dictionaries have is the one that says they need to either sell dictionaries or online subscriptions. That, in turn, gives them a reason to care about users, beyond altruism and outside of any theory of idiomaticity. Benefits of a simple lemming rule include that it would add a bias toward helping the reader and reduce the amount of repetitive gum-flapping on this page. DCDuring TALK 03:58, 24 July 2014 (UTC)

By way of contrast, there's an entry for Random-Number Exaggeration at UD. Choor monster (talk) 12:03, 24 July 2014 (UTC)

• We might consider the contents of w:Random number, a WP disambiguation page, offered here for convenience, with links to the WP articles:
  • A number generated for, or part of, a set exhibiting statistical randomness.
  • A random sequence obtained from a stochastic process.
  • An algorithmically random sequence in algorithmic information theory.
  • The output of a random number generator.

These omit explicit treatment of the lowest-common-denominator use in everyday language, some of which predates mathematical usage and some of which is a simplification or specialization of that usage. DCDuring TALK 02:36, 24 July 2014 (UTC)

I am constantly amazed at how vociferously people will argue to keep terms out of this dictionary. When many people clearly see value in having it, why should we exclude it? What possible gain is there? (Leaving aside the question of why editors here think they are better informed on this matter than the editors of the OED, Chambers, Merriam-Webster etc.…) Ƿidsiþ 11:52, 24 July 2014 (UTC)

Widsith because we're not trying to copy other dictionaries. We cannot do a better job of being the OED than the OED can, or a better job of being Chambers than Chambers can. It seems to be if you were able to understand that, you'd understand it already. I find you're clearly unable to understand what other editors are saying to you. I feel it's pointless to try and explain it to you because you just won't understand. Renard Migrant (talk) 12:29, 24 July 2014 (UTC)

He's not talking about mimicking other dictionaries (that's what "leaving aside" means), he's talking about keeping entries that could potentially help a reader understand something. How does it help Wiktionary to exclude a phrase that in fact exists and has meaning in the real world? Aside from being rude, it is bizarre that you are criticizing an editor as being "unable to understand" something when you were clearly unable to understand the point that he was making, and addressed your answer to something completely different. Why discuss the issue if you don't care enough to get it right? bd2412 T 12:51, 24 July 2014 (UTC)

I assume that we are not endorsing open-ended inclusionism of multi-word entries ("MEWs"). I have come to see that we need to include many expressions that are not clearcut idioms. A Lemming Rule, if we enacted one, would be a help.

But I suspect that many MWEs exist because, 1., they are missing, possibly shown as redlinks in other entries and, 2., it is easier for a contributor to come up with a definition for an MWE than to examine the often much lengthier entries for the constituent terms. This difficulty corresponds to the difficulty that users must have in decoding and MWE. To me that suggests that it would be valuable to come up with a user interface that would facilitate simultaneous user examination of the constituent-term entries, facilitating substitution of specific definitions and translations for them. Such an interface would also be useful in identifying missing senses for those willing to attempt such contributions.

One of the most constructive uses of RfDs of multi-word expressions, IMO, is to identify and assign some priority to weaknesses in our definitions of the constituent terms. I often find that I need to go to other dictionaries to find well-worded definitions for missing or deficient senses of RfD candidates. So many of our English entries have dated, misleading wording as well as plain, ordinary errors. The Sisyphusian or Herculean nature of addressing these deficiencies is a discouragement to addressing them globally. DCDuring TALK 13:44, 24 July 2014 (UTC)

• RFD kept as no consensus for deletion. For keeping: bd2412, Ƿidsiþ, Purplebackpack89, Choor monster, Anatoli T.; for deletion: Ungoliant, Equinox, Keφr, WikiTiki89, Renard Migrant, msh210. --Dan Polansky (talk) 08:30, 10 August 2014 (UTC)