Miscellaneous desk
< November 21	<< Oct \| November \| Dec >>	November 23 >

Welcome to the Wikipedia Miscellaneous Reference Desk Archives
The page you are currently viewing is an archive page. While you can leave answers for any questions shown below, please ask new questions on one of the current reference desk pages.

November 22

USS Indianapolis CA-35

How many Atomic bombs did the USS Indainapolis pick up at Mare Island, San Francisco to the Island of Tinian on July 26, 1945 ? —Preceding unsigned comment added by Kinkyotter (talk • contribs) 00:47, 22 November 2009 (UTC)[reply]

The Indianapolis only transported (most of) Little Boy - the uranium projectile rings and the bomb casing. All of Fat Man and the remaining parts of Little Boy - the uranium target rings - were delivered to Tinian by air. BTW, according to our articles, the departure point was Hunters Point, not Mare Island. FiggyBee (talk) 01:10, 22 November 2009 (UTC)[reply]

Semi-pedantic quibble: While Mare Island is in San Francisco Bay (or rather San Pablo Bay), it is in Solano County, California off Vallejo, rather than San Francisco, although the Naval District or other general designation encompassing Mare Island Naval Shipyard might have been "San Francisco". However, Hunters Point and the San Francisco Naval Shipyard at Hunters Point are definitely within the City and County of San Francisco. —— Shakescene (talk) 05:46, 25 November 2009 (UTC)[reply]

Firsts of Wikipedia

What was the first Wikipedia on Wikipedia? What was the first task force? and when were the founded? —Preceding unsigned comment added by 76.211.107.188 (talk) 01:10, 22 November 2009 (UTC)[reply]

What do you mean by "the first Wikipedia on Wikipedia"? -- JackofOz (talk) 01:17, 22 November 2009 (UTC)[reply]

Do you means what was the first language to have a Wikipedia? If so, English. If not, I have no idea what you mean. --Tango (talk) 01:21, 22 November 2009 (UTC)[reply]

This was the first Wikipedia on Wikipedia, I'd say. History of Wikipedia has the founding date. Vimescarrot (talk) 02:28, 22 November 2009 (UTC)[reply]

Wikiproject, sorry. My typing skills are poor :) 76.211.107.188 (talk) 04:30, 22 November 2009 (UTC)[reply]

WikiProjects were proposed by ManningBartlett in mid-2001. You can see the original proposal at [1]. As he states in the proposal, what became WP:WikiProject Tree of life was already in existence, although it didn't become known as a WikiProject until mid-2002. Warofdreams talk 17:27, 22 November 2009 (UTC)[reply]

Random numbers?

I read something that if I tell a computer to generate random numbers, the result won't really be random. WHy not/ THX in advance --Jake —Preceding unsigned comment added by 76.228.196.191 (talk) 03:34, 22 November 2009 (UTC)[reply]

Traditional random number generators, at least, start with a "seed" which is based on the timestamp. Ask yourself how you would create a set of random numbers. For example, by putting 10 numbered objects into a bag and withdrawing them one at a time. The trouble is, a computer can't quite do that. ←Baseball Bugs ^{What's up, Doc?} carrots→ 03:36, 22 November 2009 (UTC)[reply]

(ec) A software pseudorandom number generator can only perform arithmetic on a "seed" value, a number that is given to it at the start. They can't conjure up numbers out of nowhere (for that, you need a hardware random number generator). Also, if the algorithm used isn't very good (RANDU for example), the spread of "random" numbers it produces may be less than ideal. FiggyBee (talk) 03:45, 22 November 2009 (UTC)[reply]

Technically, it can—that is, if you can program a computer (or robot) to bang on its own keypad for a while. Seriously though: by definition, any number generated by an algorithm is is not random; therefore, for instance, if I take the example above and bang on my laptop's keypad, I would have a truly random string of numbers (and possibly a broken laptop). However, since computers can only operate by algorithms using seed values (see Random number generation), they cannot generate true random numbers. Intelligentsium 03:53, 22 November 2009 (UTC)[reply]

Long story short -- any process you think of that can generate random numbers is inherently pattered. There simply aren't any truly random (i.e. bizarrely non-predictable) phenomenon out there that can be harnessed for this purpose. To paraphrase Einstein, God doesn't play dice. Vranak (talk) 04:07, 22 November 2009 (UTC)[reply]

Einstein was wrong, God does play dice. There are truly random quantum processes and they can be used to generate random numbers (the number of clicks on a Geiger counter in a given second, say). --Tango (talk) 05:11, 22 November 2009 (UTC)[reply]

That's debatable. For example see quantum suicide. Rckrone (talk) 05:44, 22 November 2009 (UTC)[reply]

There is plenty of randomness in nature. Computers aren't quite "in nature", though. ←Baseball Bugs ^{What's up, Doc?} carrots→ 05:46, 22 November 2009 (UTC)[reply]

Quantum suicide only affects things which can result in the observer ceasing to exist. It says nothing about randomness more generally. --Tango (talk) 06:31, 22 November 2009 (UTC)[reply]

I suggest you lads actually read the article on random number generation before contradicting myself or Albert. Vranak (talk) 13:28, 22 November 2009 (UTC)[reply]

There's certainly no question that computer random number generation is typically not random, but Tango is right that actual randomness exists in the form of quantum measurements. The assertion that God doesn't play dice is wrong. There's no reason that you couldn't build some hardware to take advantage of that, as Trovatore discusses below. Rckrone (talk) 18:01, 22 November 2009 (UTC)[reply]

Until such time as such a device exists, God does not play dice! Vranak (talk) 19:03, 22 November 2009 (UTC)[reply]

Hmm? It exists now. It's even a "device" in the Linux sense -- /dev/random, as I said before.

Of course the entropy estimate for /dev/random is meant to defeat some vaguely realistic opponent, not one who knows the entire wavefunction of the universe. If you really want all the entropy to come from the collapse of the wavefunction, you probably have to take the bits out slower. I don't know how much slower. Just the same, in principle, /dev/random is producing truly random bits, just maybe not as fast as it appears to be. --Trovatore (talk) 20:13, 22 November 2009 (UTC)[reply]

Alright I see I am over my head. This is why I quit Computer Science after two years! Vranak (talk) 20:27, 22 November 2009 (UTC)[reply]

You have to distinguish between computers, which are physical objects, and algorithms, which are mathematical abstraction. It is certainly possible to get a physical computer to produce (truly) random numbers, if you provide it with the needed hardware. If you need a lot of them, you might want to hook up some sort of thermal noise generator, like a diode of some sort, and feed the output into a hash function that acts repeatedly on the same data in ciphertext feedback mode. This can concentrate the entropy (information theory) in such a way that you can then pull out bits that are truly random to a very good approximation.

If you're happy with a slower rate, you don't need the diode; it's enough that the computer have access to ordinary peripherals such as hard drives and keyboards. A daemon will observe things like the access times to these and feed it into the entropy pool as above. On Unix/Linux machines this entropy pool may be accessed through /dev/random. Update: I just looked at the article, and apparently it's true-random on Linux, but not necessarily so on some other Unices.

The thing is, for most non-cryptographic applications, we actually don't want truly random data. We want values that can be reproduced. For that purpose the PRNGs are ideal (and also much faster). --Trovatore (talk) 06:42, 22 November 2009 (UTC)[reply]

There is also only a limited number of random numbers available. For example, if the algorithm outputs to a single byte there will only be 256 possible values. The seed will set the start point and, if the algorithm has perfect coverage, each number will come out just once before the sequence repeats itself. During testing, the seed is usually set to a known value like zero, in order to produce a predictable sequence. When shipped, the seed is often set to a value based on the current timestamp to give a seemingly random start point in the sequence. Astronaut (talk) 16:23, 22 November 2009 (UTC)[reply]

A "random" number generator that produced each possible value just once before repeating would be very obviously non-random, and would probably not be random enough for most applications. Algebraist 18:55, 22 November 2009 (UTC)[reply]

Computers are intended to be deterministic machines - all randomness is carefully removed from their design. It is possible to generate long sequences of numbers that appear random - and perhaps stand up rather well to statistical measures of randomness - however, they are always, without fail, deterministic. We call these sequences "pseudo-random". In some situations (perhaps in a computerized game of Poker) it's important that the game does not produce the same sequence of numbers every time it's run - but apart from that - so long as the numbers it produces are statistically unpredictable - pseudo-random numbers work just fine. So some source of non-deterministic data is required to pick the first random number to be generated - and everything after that is done deterministically. One commonly used possibility is to time some event outside of the computer - the time between starting the program and the user first clicking on the mouse or keyboard, for example. You can measure that accurate to (say) a millionth of a second and (for example) use the last few digits of that number that to set up the 'seed' for the random number generator. After that point, the numbers are pretty much perfectly random as far as any outside observer can tell. Sometimes, the position of the disk platter in the hard drive is used, sometimes the latency in a network packet sent to some remote server. Sometimes all of those things added together along with...oh I dunno...the number of microseconds until easter plus the phase of the moon...or something equally crazy. It doesn't really matter what the seed is - so long as it's not always the same and so long as it's not possible for an end-user to force it to come out to a known value.

In some situations, it's desirable that the random numbers DON'T start off from a random point each time. I write computer games for a living - and for us, it's highly desirable that the exact same sequence of numbers are generated every time while we are testing the game. So when one of our game testers reports a bug, we can reproduce it perfectly by replaying the game from the start using the exact order and timings of his user-inputs (which we carefully record during play). If the random numbers come out the same when we reproduce it - we can find and fix the problem. When the game is played 'for real', we initialise the random number generator with some time-based thing - and the numbers are really, truly different every game for 'real' game players.

A few computers have been built with truly random numbers built in at the hardware level - the idea of using Schottky noise to produce truly random numbers from quantum processes can be exploited fairly easily using custom electronic hardware. But pseudo-random numbers are good enough for almost all applications - so this has been limited to a very few, super-specialized applications.

SteveBaker (talk) 17:23, 22 November 2009 (UTC)[reply]

Steve, you might have taken a glance at what I wrote, which covered the same ground, but (for once) more comprehensively. In particular it is not just the seed that is supplied nondeterministically, for extremely widely used schemes like /dev/random. Entropy is supplied nondeterministically to the entropy pool on an ongoing basis, and random values are extracted only as fast as entropy goes in.

Responding to Astronaut: You seem to be thinking of very old PRNG schemes, used in the days when memory was enormously more expensive than it is now, where the entire state of the RNG can be known from a single output. That is not even close to true for modern techniques — a much larger state is kept, and updated each time a random value is needed. For example /dev/random has a maximum entropy content of 4096 bits on my machine, so we can deduce that the RNG state is at least 4096 bits (it may well be much larger than that). Even if you stopped supplying any new entropy, but kept taking values out, it would not need to repeat until 2^⁴⁰⁹⁶ iterations. --Trovatore (talk) 21:54, 22 November 2009 (UTC)[reply]

Sorry, I wasn't thinking of particularly old random number generators and I am aware the "perfect coverage" I described is actually far from ideal. My intention was to simplify my example for the benefit of the OP. Astronaut (talk) 00:43, 23 November 2009 (UTC)[reply]

Oh, by the way, it's not only "super-specialized apps" that use true-random numbers. You need them every time you check your bank statement.

What happens, at least conceptually (in practice I imagine there are other complications I don't know about), is the following: Your browser looks up the bank's public key somewhere (exactly how this is validated I'm not sure; it's a key point of attack so I'm sure much skull sweat has gone into making this hard to hack). Then your computer generates a random session key and encrypts it with the bank's public key, and sends it to the bank. The bank decrypts the session key, and now the two of you can use an ordinary symmetric cypher to send data back and forth securely.

But if an opponent could predict your session key, he wouldn't need to break the public key cryptosystem used to send it to the bank — he could just steal your data, and eventually your money, by using the session key he knew you'd pick.

So in Linux we have /dev/random which nicely deals with this sort of situation (I don't know for sure whether browsers working in Linux specifically access /dev/random, but if not, they probably do something similar). I expect the Windows and Mac kernels also have some such functionality.

What wouldn't surprise me is if the Windows version were a little less paranoid. /dev/random is designed to give only as much output as there is entropy fed in. That means that an attacker who could intercept all previous outputs of /dev/random could still not do any better than random in guessing the next output. That's actually going a bit overboard -- say you visit 32 secure websites, generating a 128-bit session key for each, out of a 4096-bit entropy pool. In theory, if an opponent could listen in to all of those transactions, that could be enough to reconstruct the PRNG state and predict your next session key. In practice I suspect that that problem would be considerably more difficult than breaking RSA, which would be a more direct way of stealing your money. --Trovatore (talk) 00:16, 23 November 2009 (UTC)[reply]

Windows appears to use CryptGenRandom for these purposes. Algebraist 00:49, 23 November 2009 (UTC)[reply]

Oblig xkcd --121.127.200.51 (talk) 10:39, 23 November 2009 (UTC)[reply]

Does this kind of paper have a name?

The gray kind of paper usually used in coloring books and crossword/sudoku penny press-type publications that you see at the supermarket. It has a different feel to me than newspaper, so I don't think it's that. 71.161.45.144 (talk) 17:44, 22 November 2009 (UTC)[reply]

Recycled paper / unbleached paper (Kraft process)? --Cookatoo.ergo.ZooM (talk) 18:07, 22 November 2009 (UTC)[reply]

Newsprint? Newsprint can come in versions that are thicker than that used in newspapers.

Art supply stores sell newsprint drawing pads. It is relatively inexpensive; the paper is a little more stout than that found in newspapers. It has a good "tooth." That means it takes charcoal and soft pencil (graphite) well. On the downside it is not archival. That means it won't likely last very long. In a few years it may become brittle. Its color may change too. It is made of wood pulp, instead of for instance cotton, and bleaching agents are used, which linger in the finished product, and cause the likely eventual embrittlement and color change.

Of course, I am not sure if this is the type of paper you are referring to. Well-stocked art supply stores (or even online) can help you to find out a lot about paper. Bus stop (talk) 18:14, 22 November 2009 (UTC)[reply]

Newspapers from a century ago used thicker paper than they do now. ←Baseball Bugs ^{What's up, Doc?} carrots→ 04:19, 23 November 2009 (UTC)[reply]

One thing that determines how glossy the surface of paper is is the amount of calendering done. The high-gloss paper used in many magazines is supercalendered, so fine details print well, but this paper doesn't take pencil or felt-tip pen marks well. The paper you're asking about is unsupercalendered, but I don't know if they go to the other extreme and use uncalendered paper or what. --Anonymous, 05:55 UTC, November 23, 2009.

SidewaysToilet Pan?

I've recently moved into a new place and the bathroom is quite small so a regular toilet (like the one in this picture http://www.wickes.co.uk/Toilet-To-Go/invt/190476&temp=largeimage&layout=popups) won't fit properly. In my old place I had a toilet pan that had a waste/refuse/toilet pipe? (don't know what the actual porcelain part on the pan itself that the metal pipes attach to is called) which went sideways and it is exactly what I need. However I don't the term for those toilet pans and hence I can't search for them either - whether it is online or in retailers, the people have no idea where to get one or even what I'm talking about.

Any help with finding a manufacturer or store that deal with these types would be greatly appreciated. Failing that, a specific term for these toilet pans that gives results in google would be of help too. —Preceding unsigned comment added by 77.98.27.27 (talk) 19:17, 22 November 2009 (UTC)[reply]

The only thing that comes to mind is squat toilet. Bus stop (talk) 19:48, 22 November 2009 (UTC)[reply]

I think its called the toilet bowl in UK English. Gosh, your bathroom must be tiny. I would have thought it would be against Building regs to build them that small. Bring back Parker Morris Standards. 92.29.150.5 (talk) 20:15, 22 November 2009 (UTC)[reply]

There are toilet bowls which are affixed to the wall (and not the floor). They have a P-trap (instead of an S-trap) which connects to the outflow pipe in the wall (probably 30 cm up from the floor). You can save a bit of space by using one of those as there is no "gap" between the back of the bowl and the wall. On the downside, they generally require the flushing cistern to be built flush into the wall. --Cookatoo.ergo.ZooM (talk) 22:43, 22 November 2009 (UTC)[reply]

I suppose I'm not very good at desciptions - it's a regular flush toilet alright, its just that the layout of the room doesn't allow the door open fully against the wall since it hits the current toilet which is literally jutting out :p. What I'm talking about the Back of the toilet where you attach the metal pipes. They're usually straight or going down to the floor depending on the toilet design, right? Well the one I'm on about is just like those but instead of going straight or down, the back part is at a 90 degree angle either to the right or to the left. —Preceding unsigned comment added by 77.98.27.27 (talk) 22:49, 22 November 2009 (UTC)[reply]

Are you thinking of Saniflo masher toilets? Perhaps have a look at their products and see if any look familiar [2]. I don't know who else makes this sort of thing. 86.144.149.168 (talk) 22:59, 22 November 2009 (UTC)[reply]

Well, a picture's worth a 1000 words so I got hold of these pictures of the design I'm on about: the first - [3] the small pipe on top is the cistern, with the lower one being the one that takes the contents of the toilet out to the main sewage pipe. MY MAIN Point is the fact that the toilet's "nozzle" for lack of a better word is facing at 90 Degrees to the toilet itself making it fit better into the space! I've provided a second picture of the other side in case you can't see there's a bend (well tried to anyway most of the picture is taken up by the dirty pipe) - [4] the pipe is perpendicular to the toilet so that means the Nozzle isn't straight - it's on the side. I hope I've finally managed to communicate across what I'm looking for exactly. —Preceding unsigned comment added by 77.98.27.27 (talk) 23:31, 22 November 2009 (UTC)[reply]

Metal pipes? There would be a small pipe for the cold water into the cistern, usually a small overflow pipe, and then the sewage pipe itself. Is that what you mean please? You could try asking plumber's merchants like Plumb Centre if they have any. Edit: I have realised that your first picture is rotated 90 degrees from the vertical. Still a nauseating image anyway. Perhaps you need an old-fashioned style of high-level cistern as well. 84.13.162.136 (talk) 10:15, 23 November 2009 (UTC)[reply]

The picture of the one-piece sideways toilet you linked to looks very unusual to me - I've never seen one like that and so I'm not surprised that the people in the shops don't know about it. A real specialist part. However, you might be able to assemble something that will do the same job. Many toilet bowls these days end with a horizontal pipe that does not extend backwards very far (in many cases other parts of the toilet extend back further, so the pipe is not the furthest-back part of the design). See eg this picture where the pipe is visible coming out the back of the toilet. You can also get "pan connectors" that fit directly onto such a horizontal pipe, and make an immediate 90º bend. See this or, if you needed to cheat another inch or two, this turned horizontally. Putting such a toilet (probably not that exact one, as it's meant for a corner, but was the first one I found where the outlet was visible) and such a pan connector together, you'd have something that ought to fit in the space of your strange 90º toilet. 93.97.184.230 (talk) 19:25, 23 November 2009 (UTC)[reply]

Since you are going to need a plumber to fit a new toilet (would definately not advise trying to do it yourself, if that is what you are thinking), why not ring around a few plumbers to see if any of them understand the problem? Hopefully you would pay a trade price through them rather than a retail price if you bought it yourself. 84.13.162.136 (talk) 20:42, 23 November 2009 (UTC)[reply]

I fitted a new toilet in my bathroom a couple of months ago. It's not especially difficult. In fact I refitted the whole bathroom, and while it was hard work, no individual part of it was difficult as such. Except tiling; I got a pro in for that as it's a real hands-on skill I don't have time to learn.

Trade vs retail pricing for parts in the UK is a dying concept; the cheaper merchants are happy to sell to anyone, and the old-fashioned ones that maintain a protectionist "trade only" policy are generally more expensive hence not worth dealing with anyway; I know "trade" sparks and plumbers who've moved to Toolstation etc rather than the local merchant for this reason.93.97.184.230 (talk) 01:51, 24 November 2009 (UTC)[reply]

Random Number Psychology

Some time ago I heard of a research study that asked a large pool of participants to select a "random" number from 1 to 100. Because human beings aren't really random, then tended to choose some numbers more frequently than others (the point of the study was to look at these psychological biases). If I recall correctly, 37 was the most frequently chosen "random" number and 20 was the least frequent.

Does anyone know about this study? I would like to find the original publication. Dragons flight (talk) 22:08, 22 November 2009 (UTC)[reply]

Makes sense to me. 37 is the first prime number that doesn't get regular usage. Everything up to 31 gets used as a day of the month. 32-36 are not primes! And 20 sticks out like a sore thumb when it comes to ordinary numbers. I think that solves that. Vranak (talk) 22:46, 22 November 2009 (UTC)[reply]

Except the OP wasn't asking for an explanation - they asked for information and references regarding this study. Vimescarrot (talk) 23:58, 22 November 2009 (UTC)[reply]

Well I am happy to provide one regardless, dear sir. Vranak (talk) 00:09, 23 November 2009 (UTC)[reply]

Don't know of any direct references, but I can remember being taught in intermediate psychology at university that 37 was the two digit number picked most frequently at random by experimental subjects. There seems to be a potential academic source in a reference at this site, though I haven't seen the original paper, so I don't know if it has any experimental evidence. Grutness...wha? 00:18, 23 November 2009 (UTC)[reply]

Casey Stengel said a lot of things that seemed random but were actually thoughtful on close examination. Perhaps it's not a coincidence, that his uniform number was 37. ←Baseball Bugs ^{What's up, Doc?} carrots→ 04:18, 23 November 2009 (UTC)[reply]

42. DOR (HK) (talk) 04:45, 24 November 2009 (UTC)[reply]

When asked to choose between 0 and 9, 7 is the most common result according to various studies (see page 373 here [5]). The authors - Thomas Griffiths and Joshua Tenenbaum - suggest that people are reluctant to pick numbers that have some obvious property that makes them stand out (or equivalently to pick numbers which are members of a specific subset), e.g. start/end of a range; multiples of 2, 3, 5; powers of two - they use this to produce a mathematical model which closely reflects the actual distribution. 7 comes into none of their categories (it is prime, but fewer people can spot a prime number than an even number). Although the authors don't consider numbers above 9, 37 is also a number with few obvious properties, while 20 has many (being a multiple of 2 and of 10). The paper has some references to other studies which you might check out. --Maltelauridsbrigge (talk) 14:47, 24 November 2009 (UTC)[reply]

I always choose one of the "end" numbers figuring no-one ever choses them..hotclaws 17:06, 25 November 2009 (UTC)[reply]

joseph greiner

Good morning, I am a WWII enthusiast, during my investigations I have come across a picture. I am unable to load it here as I am unsure of how this is done, th picture can be viewed on google if you search for the above mentioned name. The picture is of 2 native american gentleman standing next to a cross or what appears to be a grave the large cross that makes up the headstone reads joseph greiner and has a swastica on it. Now when one thinks about the ratlines, and nazi's escaping europe and fleeing to south america, it would appear this may be the grave, or something like that, of a nazi. Can anyone shed any light on who joseph greiner was? That is what I would like to know. who was joseph greiner. Thank you very much. —Preceding unsigned comment added by 86.18.35.133 (talk) 22:47, 22 November 2009 (UTC)[reply]

The person died in Brazil in 1936, which was 9 years prior to the time Nazis from the defunct 3rd Reich fled to South America. Joseph Greiner was a member of the German Jary expedition which was involved in a lunatic plan to occupy French, British and Dutch Guyana to establish a German colony. Some of the zoological samples taken by the expedition are still on display in Berlin. The whole plan was dumped by Himmler in 1937 (French Guyana was a colony of Vichy France and the whole scheme was crazy, anyway). Reference: Spiegel.online, [6]. If you read German: Jens Glüsing: "Das Guayana-Projekt. Ein deutsches Abenteuer am Amazonas". Chr. Links Verlag, Berlin 2008, 240 Seiten. 19,90 Euro. --Cookatoo.ergo.ZooM (talk) 23:26, 22 November 2009 (UTC)[reply]

A colony of Vichy France in 1937? That's quite a trick... FiggyBee (talk) 11:21, 23 November 2009 (UTC)[reply]

Those darn Nazis and their time travel... DJ Clayworth (talk) 16:51, 23 November 2009 (UTC)[reply]