Wikipedia:Reference desk/Archives/Science/2012 March 21

= March 21 =

who is closer phylo-genetically to the human?
Chimps or Bonobos?, and why it's in dispute?, thanks. 109.64.44.20 (talk) 02:35, 21 March 2012 (UTC)


 * The two Chimpanzee species last share a common ancestor with Homo sapiens around four to six milion years ago, while the last common ancestor of Pan troglodytes and Pan paniscus lived less than a million years ago - they are equally closely related to us. AndyTheGrump (talk) 02:45, 21 March 2012 (UTC)


 * While AndyTheGrump is scrupulously correct, from a scientific perspective, I can see where the question itself seems important from a psychological/philosophical standpoint. Bonobos like to fuck and cuddle, Chimps like to fight and kill, and as a human, it seems important to us to know what it is in our "nature" to be more like: The loving Bonobo or the violent Chimp.  Andy is, of course, correct that we are equally related to both of them; but it doesn't obviate the question, philosophically, which aspect of human nature (the good or the evil, the loving or the hateful, if you will) is "closer" to our true nature.  Unfortunately, cladistically that question cannot be answered, for as Andy notes, they're both as related to us as the other.  Even if they weren't, however, it still wouldn't answer the question.  Closest common ancestor is not going to give any clear indication on behavioral patterns and evolution.  -- Jayron  32  03:21, 21 March 2012 (UTC)


 * Eh? I love to go on a bit of a rant as much as the next guy, but the question was specifically about Phylogenetics, not philosophy or human nature. I'm not sure where you inferred that from. As to the question, I think using the human family tree analogy is the best way to answer these kinds of questions (along with the kind of answer Andy gave) because we are all so familiar with those. So, you are essentially asking, if your uncle has two kids, your cousins. They are closely related to eachother (they are siblings), but which is more closely related to you? The answer is neither, they are both your 1st cousins. More accurately, the siblings in this example should probably be cousins, but that makes the tree one more level deep which gets very complicated to explain without a diagram. Vespine (talk) 04:49, 21 March 2012 (UTC)


 * Well, I think Jayron is getting at the issue of what is the ancestral behavior of our ancestors. If you go back to that last common ancestor 4-6 million years ago, did that ancestor act more like a Chimp, a bonobo, a human, or something else entirely? Which animal are we more like, and which animal should we be more like, based on our genetics? With comparative genomics you can infer which mutations occurred in which ancestors and when, but we're left in the dark when it comes to complex animal behaviors, unless they can be causally linked to genes. Someguy1221 (talk) 05:04, 21 March 2012 (UTC)


 * I think we'd need to inject a little scepticism into Jayron's 'fuck and cuddle' vs 'fight and kill' characterisations of the two species too. While there seems to be evidence to bear this out to some extent, it needs to be pointed out just how limited the data is - and how we may be making invalid comparisons based on differing environments, rather than on any objective difference between the two species. As much as I admire Jane Goodall's seminal work amongst chimpanzees for example, I have to agree with her critics: much of the violence she observed may have been down to the unusual context - artificial provisioning, and a great deal of stress from human encroachment into the ecosystem. Given their obvious intelligence, their intensely social behaviour, and the degree to which chimps pass on knowledge through learning, I see no particular reason to assume that there is any great difference in an 'inherent propensity to violence' between the two species at all - a chimp, like a human being, is born with the potential to become all sorts of individual - and just what sort of individual it becomes will depend very much on the environment it is brought up in - both ecological and social. AndyTheGrump (talk) 05:17, 21 March 2012 (UTC)


 * I think we could potentially do better than just saying we are each equally related. Just like a person will share more chromosomes with some first cousins than others, either chimps or bonobos presumably share more active genes with us.  That is, since we split with our last common ancestor, humans have had some genes changes, and chimps/bonobos have has some genes change.  The genes which changed before those two split will presumably not distinguish between the two (unless some of those genes changed back after the final split).  However, genes which changed after the final split, but which humans still possess, should exist.  I'm not sure if anyone has ever done such a study, though.  Since there seems to be very little different between chimps and bonobos other than their behavior, I suspect that genes which control behavior are where the largest changes occurred.  Of course, some behavior might also be cultural/learned.


 * Also note that a comparison of chimp and bonobo genetics might give us great insight into how to predict, and even control, human behavior, so could be quite valuable for our survival as a species. StuRat (talk) 06:30, 21 March 2012 (UTC)


 * As has been explained on the RD before, including I think to you, even closely related people often don't share any chromosomes due to crossing over. So it doesn't make sense to talk about how many chromosomes first cousins share. And when it comes down to it, it seems likely it's theoretically possible two people could share a chromosome but have lower genetic similarity then two people who doesn't share any chromosomes. Nil Einne (talk) 17:56, 21 March 2012 (UTC)


 * Those effects are going to be pretty much irrelevant when looking at first cousins. On average (ignoring cross-over and mutations), they should share 1/4 of their chromosomes.  I can't imagine cross-over or mutation having any effect anywhere comparable to changing even one chromosomes worth of genes in a significant portion of such closely related people, so it can be safely ignored. StuRat (talk) 23:02, 21 March 2012 (UTC)


 * While it is definitely true that the populations are equally closely related, it is at least possible that one species may be "more related" than the other in the sense of lineage sorting. In other words, suppose the Ancestor of All Apes lived in a population with 50% soulgood allele and 50% soulBad allele.  Well, maybe over the course of evolution the bonobos lost all the soulBad alleles, or most of them, and likewise the chimps lost their soulgood alleles to some different kind of selection.  And the humans, having presumably somewhat less than 50% soulgood alleles, are then "more related" to the chimps.  If you did a typical 1970s kind of experiment and cloned out the genes from "the" chimp, "the" bonobo, and "the" human (by which I mean one specimen of each that blundered into reach of a grad student), and ran them on a zoo blot with each others' probes, then you'd see that the human was "more related" to the chimp.  (At that locus, but such niceties were not stressed so much in those days)  Now of course I suspect that personality is coded in more than one gene, and it's possible that lineage sorting occurred in many of these genes, so that you could come up with some kind of an answer.  But defining which genes those are, which changes are important, that's a problem.  Among other practical difficulties, many more random variations exist with no significance whatever which have in all likelihood been fixed in different species completely at random.  So the question is not complete nonsense but conceivably answerable, if you define what you're asking very, very carefully; but it is not presently answerable.  And even if answerable, it will be most likely a quite small effect. Wnt (talk) 18:23, 21 March 2012 (UTC)

USSR/Russia interplanetary probes
The USSR was an early pioneer in sending probes to the Moon and to Venus. They also sent some to Mars. Why hasn't the USSR/Russia sent probes beyond Venus and Mars? Bubba73 You talkin' to me? 05:23, 21 March 2012 (UTC)
 * It likely has a lot to do with cost. Cassini-Huygens cost about US$3.2B and according to Space industry of Russia, their budget as of 2011 is US$3.8B.  With the cost of sending people and cargo to the ISS, there isn't much left for probes to the outer planets.   Dismas |(talk) 05:58, 21 March 2012 (UTC)
 * Maybe if America hadn't been spending so much on interplanetary exploration, we might still have a Shuttle program. ←Baseball Bugs What's up, Doc? carrots→ 06:05, 21 March 2012 (UTC)
 * Well, there there is money, but it's spent on other things:
 * 2010 NASA budget = $18.724 billion (http://en.wikipedia.org/wiki/NASA_Budget)
 * 2010 Military Budget = $1.03 trillion (http://en.wikipedia.org/wiki/Military_budget_of_the_United_States)
 * Just saying... 217.158.236.14 (talk) 09:59, 21 March 2012 (UTC)


 * This is true, but there's no know-how left to send even a smaller probe anywhere. New Horizons cost $650M over 15 years. This would have been within the budget if Russian engineers knew how to build probes. Which they don't, because the last successful Russian/Soviet mission beyond the low Earth orbit took place (IIRC) in 1986.--Itinerant1 (talk) 06:17, 21 March 2012 (UTC)
 * Correction: I should've said "beyond the geosynchronous orbit". --Itinerant1 (talk) 09:27, 21 March 2012 (UTC)
 * Because the USSR collapsed in 1991 and very nearly took the entire Soviet space industry under as well. In the 90's, government financing of science was minimal, scientists and engineers were being paid starvation wages, existing technologies were becoming obsolete, and there was no money to develop new ones. Things got a little better in the 00's, but, as of today, starting salary of an engineer or a programmer at NPO Lavochkin is the equivalent of $700/month (this is near Moscow, which is one of the most expensive cities in the world, on par with New York or London.) Russian space industry lost an entire generation of employees because STEM students were going into business or emigrating from the country. The average age of an employee is in the 50's. It's a wonder that Russia can still launch anything at all, let alone send probes beyond Venus and Mars.--Itinerant1 (talk) 06:12, 21 March 2012 (UTC)


 * Russia has a bit of a curse when it comes to Mars probes. So while it's true that they "sent" a whole bunch of probes to Mars, I don't think that any of them actually got there in one piece and then functioned properly. (Their Venus probes were pretty awesome though. Go figure.) APL (talk) 10:36, 21 March 2012 (UTC)


 * Ok, I see about the economic problems since 1991. But in the 1970s and 80s we were sending probes to Mercury and the outer planets.  Why didn't they do it too?  Bubba73 You talkin' to me? 15:01, 21 March 2012 (UTC)


 * Soviets put all their interplanetary efforts into Mars and Venus. Their Mars program was notoriously unlucky (as APL says), they sent close to ten probes but the best they could achieve was to land one probe on the surface and to have it transmit garbled data for 3 minutes. They had more luck with Venus, with over ten successful probes of varying complexity (Venus has a number of unique problems, with its dense atmosphere and 460 °C surface temperature, and Americans pretty much skipped it.)
 * In the late 70's Soviets decided that they had to have their own Space Shuttle, and resources were diverted to that program from interplanetary exploration.
 * There were supposedly some plans to send probes to Saturn and to the asteroid belt in the late 80's, but those never progressed beyond the planning stage by 1991. --Itinerant1 (talk) 16:47, 21 March 2012 (UTC)


 * Was it really just bad luck, or more a matter of not devoting sufficient resources to adequately test designs and provide backup systems ? StuRat (talk) 22:56, 21 March 2012 (UTC)

Solid fluorine
What is the appearance of solid fluorine? Double sharp (talk) 13:39, 21 March 2012 (UTC)
 * This paper should have some information for you. -- Jayron  32  16:29, 21 March 2012 (UTC)
 * Although it has no pictures, our article Fluorine describes both the alpha and beta phases of solid fluorine.  Spinning Spark  16:34, 21 March 2012 (UTC)

Application for vocational studies in Sweden.
Please sir/madam, kindly direct me on how and when to apply  for vocational studies under science department in Sweden. Thanks, am Cosmas from Nigeria. — Preceding unsigned comment added by 95.209.176.236 (talk) 15:00, 21 March 2012 (UTC)
 * I'm not sure if this is what you're looking for, but Myndigheten för yrkeshögskolan, the Swedish National Agency for Higher Vocational Education has parts of its web site in English. Sjö (talk) 19:09, 21 March 2012 (UTC)

Astatine alpha-decay
I hope you can help me. I am writing astatine and I have a question about its alpha decay characteristics. At-211 has 126 neurtons, which is a magic number. Why then energy of alpha decay of At-211 exceeds that of At-210? Also notable that At-213 (N=128=126+2) has two more protons than 126. Its alpha decay half-life is in accordance: the shortest of all astatine isotopes. IS there a reason why At-211 is not the longest-lived astatine isotope? Thanks--R8R Gtrs (talk) 15:19, 21 March 2012 (UTC)


 * It's fairly complicated - there are multiple mechanisms at work.
 * This is the table of binding energies per nucleon for isotopes of astatine from 204 to 214, in keV:


 * 204	7803.50
 * 205	7810.38
 * 206	7809.10
 * 207	7814.07
 * 208	7811.69
 * 209	7814.83
 * 210	7811.73
 * 211	7811.42
 * 212	7798.35
 * 213	7790.07
 * 214	7776.43


 * Source? This could be useful for either the Astatine article or the isotopes of astatine article. Thanks! Allens (talk &#124; contribs) 14:29, 22 March 2012 (UTC)
 * Source--Itinerant1 (talk) 19:44, 22 March 2012 (UTC)
 * First of all, if you forget for a second about magic numbers and other discrete effects, the isotope of At that maximizes binding energy per nucleon would be At-208.
 * I'm also working on Astatine (copy-editing/clarifying). Why would that be? Thanks! Allens (talk &#124; contribs) 14:29, 22 March 2012 (UTC)


 * On top of that, we have a discrete parity term (see Semi-empirical mass formula), which lowers the binding energy of 206, 208, 210, etc. So, in practice, the isotope with the highest binding energy per nucleon is 209.
 * Magic number of neutrons seems to boost the binding energy of the 211, but not enough to make it higher than 209 or 210.
 * Then it turns out that having the highest binding energy is not enough to be the most stable isotope. We also have to look at binding energies their prospective decay products. In case of At-209 vs 210 vs 211, numbers work out in such a way that At-211 has just enough excess energy over Bi-207 that it will commonly alpha-decay, but 209 and 210 have a little less and they almost always decay through electron capture.
 * If the binding energy of At-211 were just a little higher (say, 7812.5 keV/nucleon), it wouldn't commonly decay through alpha-decay either, and it would become the most stable isotope.--Itinerant1 (talk) 19:22, 21 March 2012 (UTC)

Version:1.0 StartHTML:0000000167 EndHTML:0000001140 StartFragment:0000000484 EndFragment:0000001124

Again, I don't know what your looking for but I found Mathematics for the Million pulled together all the abstract stuff I learnt in skool into something where I could see how I could apply it in the real world (inc. probability and statistics ). It also filled in many, many gaps  of what I wasn't taught. I think it is well worth a read – several times over.

Recommended reading
I graduate from my computer science bachelor's this year. I've enjoyed the probability and statistics elements of the subject, both taught and extracurricular. These taught courses have been fairly informal (e.g. no precise description of space) - I'd like to improve on this enough that descriptions elsewhere are more easily understandable, but I have also made the deliberate choice to post on this desk rather than the mathematics one. I've enjoyed that much of what I've learned was immediately useful, so I'm particularly interested in algorithmic processes which address practicalities of real-world problems. That's quite general, so I mean things like PCA, Metropolis-Hastings and Levenberg–Marquardt. With this in mind, if wonder if the fine folks of the reference desk can recommend any moderately comprehensive introductory texts? Thanks very much. 77.97.198.48 (talk) 15:31, 21 March 2012 (UTC)
 * I've no idea if this book bears any resemblance to what you're asking, but if you enjoy the mathematics of real-world problems, you should read The Road to Reality by Roger Penrose. --TammyMoet (talk) 16:09, 21 March 2012 (UTC)


 * Again, I don't know what if this is what your looking for but I found Mathematics for the Million pulled together all the abstract stuff I learnt in skool into something where I could see how I could apply it in the  real world (inc. probability and statistics ). It also  filled in many, many  gaps  of what I wasn't taught. I think it is well worth a read – several times over. --Aspro (talk) 18:56, 22 March 2012 (UTC)
 * What you are looking for is a textbook of numerical methods. There are numerous ones and it takes time and effort to understand them in depth. Start with small things like the Newton-Raphson method, implement it, ponder it, understand the Taylor-Maclaurin series and enjoy the path. Shyamal (talk) 10:46, 23 March 2012 (UTC)

15 amu and an atomic number of 7(reply FAST)
If a neutral atom has a mass of 15 amu and an atomic number of 7, determine the number of protons, neutrons, and electrons in the atom.---74.178.186.35 (talk) 19:45, 21 March 2012 (UTC)


 * Ok, here's a fast reply - do your own homework. AndyTheGrump (talk) 19:47, 21 March 2012 (UTC)


 * Here is my RE,back off.74.178.186.35 (talk) 19:50, 21 March 2012 (UTC)


 * It's not my homework my little bro said I might no know this one,he gave me time.74.178.186.35 (talk) 19:53, 21 March 2012 (UTC)
 * Tell him you don't know but offer to help him find the answer and then see if you can do it without asking for help which should be possible as it's a fairly easy question. If he already knows and this is some sort of game, then I would suggest by having to ask here you've already failed. You may also want to look for another place for timely requests (which you always seem to have) as the RD isn't really suitable for that sort of thing (you may find you'll need to pay for a service which can guarantee the sort of response you always seem to need). Also as I believe you may have been told before, don't ask the same question on multiple desks. Nil Einne (talk) 20:19, 21 March 2012 (UTC)


 * See Atomic number and Atomic mass unit. 70.59.28.93 (talk) 20:07, 21 March 2012 (UTC)
 * Hm thanks for the help,I'll tell that sucker straight.74.178.186.35 (talk) 20:23, 21 March 2012 (UTC)
 * Fellows, let's be reasonable, huh? This is not the time or the place to perform some kind of a half-assed autopsy on a fish... And I'm not going to stand here and see that thing cut open and see that little Kintner boy spill out all over the dock.Anthony J Pintglass (talk) 21:39, 21 March 2012 (UTC)

Nitrogen-15. Whoop whoop pull up Bitching Betty 22:00, 21 March 2012 (UTC)


 * In case anybody hits on this in the archive in the future, let's be clear:


 * PROTONS = atomic number
 * NEUTRONS = amu - atomic number (i.e. amu = protons + neutrons)
 * ELECTRONS = protons unless the atom is ionized. The number of + or - indicates the number of excess protons or electrons, respectively.  (Ca++ has two fewer electrons than the protons)  Ionization changes the number of electrons, not the number of protons.

Wnt (talk) 02:43, 22 March 2012 (UTC)


 * Actually it's the number before the "+" or "-" that indicates the number of excess protons or electrons respectively (it's Ca2+, not Ca++). Double sharp (talk) 12:50, 22 March 2012 (UTC)
 * Just to put a finer point on it, "++" appears to be pretty common (but not overwhelming majority) in older literature, and usually (but not always) superscripted. However, current IUPAC recommendation (IUPAC Red Book 2005) is "A charge placed on an atom of symbol A is indicated as An+ or An–", and similar instructions for charge on a polyatomic unit. DMacks (talk) 14:01, 22 March 2012 (UTC)
 * I think of these as simply synonymous; sorry for any confusion. Wnt (talk) 01:43, 24 March 2012 (UTC)


 * "Ionization changes the number of electrons, not the number of protons". Yup, which is why when there are more electrons we add... er hang on a minute... we subtract... um... Neutral means an equal number of electrons an protons, and a positive charge means that there are less electrons. I think that what we have here is further evidence that Sod's Law is superior to the Laws of Physics. Never trust a physicist to make a guess... ;-) AndyTheGrump (talk) 06:36, 23 March 2012 (UTC)
 * And what about beta decay, where an electron that didn' exist is ejected from the nucleus where electrons don't reside? This change of ionization by changing the number of protons and not the number of electrons is the basis for Technetium-99m production for medical applications. DMacks (talk) 06:54, 23 March 2012 (UTC)
 * Alpha decay removes two protons and two neutrons from the nucleus, and would also be a form of ionization. However, I think the usual meaning of "ionization" excludes radioactive decay and refers only two to electrons being gained or lost. Double sharp (talk) 08:55, 25 March 2012 (UTC)
 * I think you meant to not two Nil Einne (talk) 19:38, 25 March 2012 (UTC)
 * Oops, yes. (LOL!) Double sharp (talk) 14:28, 19 August 2012 (UTC)

accelerating expansion of the universe
What if any, is the effect of radiation pressure on the expansion rate of the universe?203.110.136.172 (talk) 22:34, 21 March 2012 (UTC)


 * Radiation (such as the CMBR) slows the expansion. -- BenRG (talk) 23:11, 21 March 2012 (UTC)
 * Due to the radiation's gravity? --Sean 20:28, 22 March 2012 (UTC)
 * Yes, I think so. See Friedmann–Lemaître–Robertson–Walker metric.  If my understanding is correct, radiation such as the CMBR contributes to both the pressure p and the mass density ρ, and both p and ρ contribute toward decreasing the expansion rate according to the Friedmann acceleration equation .  Red Act (talk) 04:56, 23 March 2012 (UTC)

Chemical weapons on the deep sea floor
Through random browsing I came came across the following two (seemingly unrelated) articles: Operation CHASE and Wreck of the RMS Titanic. Apparently before the Titanic wreck was discovered scientists thought that the anaerobic and near-freezing sea floor would preserve the ship almost perfectly. This is presumably the same mentality that green-lighted the sinking of thousands of tons of chemical weapons in Operation CHASE. Now that we know how quickly steel structures gets eaten away at the sea floor, has there been any new studies done on the Operation CHASE wrecks? Anonymous.translator (talk) 22:41, 21 March 2012 (UTC)


 * Anaerobic and near freezing temps don't preserve steel, they preserve certain organic objects, like wood (although I'm not sure conditions aboard the Titanic qualify). As for chemical weapons disposal on the sea floor, I believe the idea is that any which slowly leak out will be sufficiently diluted by all the water in the oceans to be rendered harmless.  They might also break down is sea water.  The concern I'd have is that plants and animal right near the leaking chemicals might absorb them while still concentrated, and, via biomagnification, some fish or other seafood up the food chain from them might end up on our dinner table. StuRat (talk) 22:49, 21 March 2012 (UTC)


 * "Anaerobic and near freezing temps don't preserve steel" my point was that 50 years ago they didn't understand this. I don't get why you're assuming it will leak out slowly though. I imagine once a canister is sufficiently weakened the huge pressure differential will basically rupture the vessel while spilling most of the contents out. Anonymous.translator (talk) 23:00, 21 March 2012 (UTC)


 * Once it develops a pinhole leak, seawater would quickly pour in and equalize pressures. At this point, the contents would only slowly leak out. StuRat (talk) 23:18, 21 March 2012 (UTC)


 * I have a hard time picturing anybody thought the canisters would last forever - it's not like the Navy lacks experience with corrosion. This sounds like "light fuse and get away" implemented on a career timescale. Wnt (talk) 02:37, 22 March 2012 (UTC)


 * And to think these are the responsible adults whom we entrust our nuclear weapons to...Anonymous.translator (talk) 11:53, 22 March 2012 (UTC)


 * The dumping was done on the assumption of eventual harmless 'dilution' of the chemical warheads. The historic legacy of this disposal is still under current revue. Example:  www.fas.org/sgp/crs/natsec/RL33432.pdf  And  www.unidir.ch/pdf/articles/pdf-art2963.pdf  This reminds me, I've got a beyond consume date jars  of really, really  hot of chillies, so rich in Capsaicin that they put most cans of Mace to shame and thus  that probably need some ecological safe form of disposal. Maybe I should sell them on Ebay. --Aspro (talk) 19:31, 22 March 2012 (UTC)
 * Well, if dilution was regarded as safe and effective, why drop them in canisters? Why not leave a hose trailing behind a ship and pour the stuff down the hose at a slow, steady, controllable rate? Wnt (talk) 20:12, 24 March 2012 (UTC)


 * Why? Cost, labour, effort and danger of empting the munitions only to dump the contents into the sea. Who would have to pay for this ?– the tax payer naturally. What extra benefit would this proposed and expensive form of   'hose' of dispersal realize?   I don't know – you proposed it, so you tell me. --Aspro (talk) 22:11, 25 March 2012 (UTC)

Convention when drawing molecular ions.
When drawing the structure for a molecular ion such as $$N(CH_3)_4^+$$, does convention dictate that you include the charge of the ion or not? Widener (talk) 23:30, 21 March 2012 (UTC)


 * It does dictates that you do. Plasmic Physics (talk) 00:17, 22 March 2012 (UTC)




 * Typically the charge is drawn localized on one atom (the nitrogen in this case) as a formal charge, though sometimes it is applied to the whole species, using brackets. Buddy431 (talk) 03:34, 22 March 2012 (UTC)


 * Depending on the context, you may want to include unpair electrons in the same style as the charge. Plasmic Physics (talk) 06:44, 22 March 2012 (UTC)

Is it possible for a person to function as a human ZIP/RAR data compressor?
Is it possible for a person to function as a human ZIP/RAR data compressor? Memorize a small amount of data and then use an algorithm to extract it?

Ok so I had a friend who stole the answer-key to a huge multiple-choice final exam, which took 3 hours to complete, and must have had hundreds of questions. So I said memorizing that many letters in a row, must be harder than the test itself.

But he had a system. Where he could memorize a relatively small amount of data. Just a paragraph. And then he could plug that paragraph into an algorithimic formula, which would extract the full pattern of answers. In effect he was functioning as a human RAR. He was memorizing 40kbs of data but compressing it into 2kbs of data, and then using a formula to extract it. He was going through the same step by step process that WINRAR goes through when it takes the compressed data of an RAR file and then unpacks it.

So is what my friend did actually possible? Is it really possible to be a human ZIP file? --Gary123 (talk) 23:46, 21 March 2012 (UTC)


 * Yes, most mnemonic techniques are analogous to data compression in symbolic systems. 70.59.28.93 (talk) 00:47, 22 March 2012 (UTC)
 * What you are asking is can a human perform mental Lossless data compression in a test environment. I think the answer would be no. The algorythms used for compression are not trivial. The main thing is, i'm pretty sure for any algorythm I can find that you can't, for example, JUST extract the answer to the 1st question, then extract the 2nd anwser, then the 3rd, you have to decompress the entire block.. I expect someone MIGHT be able to decompress some data if they could work with pen and paper and had a long time and a lot of space, but to decompress 2kbs of data would require a LOT of effort and I can't imagine you could do it in your head, unless you are alread a savant and then remembering 40k of data might not be such a big deal. There might be more simple, less efficient algorythms that I'm not taking into account, but i find it hard to believe a simple algorythm could be used to compress data 20 fold unless there were some really obvious "patterns" in the data, which seems like something that an answer key would want avoid. Vespine (talk) 00:57, 22 March 2012 (UTC)
 * Hmm I suppose in a limited sense, mnemonics could be used to compress data, but I doubt it would acheive 95% compression, unless you're substititing whole words for letters, like BODMAS, but that would certainly NOT work with something like an answer key. IF the answer key was made of purely binary questions (yes or no, true or false), you could break it up into maybe 16 "half byte" chunks, like yyyy, yyyn, yyny, yynn, and assigne each a letter and remember the string of letters, that would give you a quarter compression. If the answer key is multiple choice of 3 or 4 answers, or more, and if it is NOT all uniform (some have 3 answers some 4) I think that would introduce a level of complexity which would make that method unusable. Vespine (talk) 01:16, 22 March 2012 (UTC)


 * You can encode longer strings of letters using a systematic patern and use a single letter or word to represent a string of 4 or 5 letters. For example there are 24 combinations of strings of four (ABCD, ABDC, etc). So to remember the answers to 120 questions you would only need to remember a sequence of 30 words of code. Sky Machine   ( ++ ) 01:17, 22 March 2012 (UTC)
 * Are you neglecting that 4 answers need not be unique and could be AAAA,AAAB,AAAC,AAAD... That would make far more then 24 combinations of strings of 4. Vespine (talk) 01:24, 22 March 2012 (UTC)
 * Thinking about it just a little more, yes or no questions are binary, a ABC or D question is Quaternary, the limit of a 4 digit quartenary number is 255. Vespine (talk) 01:28, 22 March 2012 (UTC)
 * Yeah my bad. A further encoding/decoding step is probably necessary then. Sky Machine   ( ++ ) 01:44, 22 March 2012 (UTC)
 * Maybe using a shorter string would be better, there are 64 combinations of three letters (AAA, AAB, AAC, AAD, etc) encode each with a letterpair (1=AB, 2=AC, 3=AD, 4=AE, 5=AF, ... 10=BA, .. 64=GE) use a wordpair associated with each letter pair AF=airforce etc. Then you just need to remember a 60 wordpair (or syllable pair)sequence (of 64ish possible wordpairs) for a 180 question multiple choice test. Although depending on which is easier you could break the string of 60 up into shorter mnemonics say 6 x 10 word sentances.  Sky Machine   ( ++ ) 03:35, 22 March 2012 (UTC)


 * Contestants of memory championships don't compress information, they do the opposite. To quote from a Slate article: When Cooke sees a three of clubs, a nine of hearts, and a nine of spades, he immediately conjures up an image of Brazilian lingerie model Adriana Lima in a Biggles biplane shooting at his old public-school headmaster in a suit of armor. Btw, final exam with 20 000 questions? Wow...84.197.178.75 (talk) 01:36, 22 March 2012 (UTC)


 * Assuming a 4 choice multiple-choice test, 20kb of data would equal 10000 questions. Given 3 hours, a (presumably human) test taker would have roughly one second per question. Most people can't even fill in their Scantron sheet that fast. I suspect your friend greatly embellished his story. Anonymous.translator (talk) 02:13, 22 March 2012 (UTC)


 * From the story above, I'm thinking the friend had free access to a computer with the answer key on it and could freely do a compression on it, but had to face some kind of search to get it out. But does anybody really do that?  I have a hard time picturing ETS or whoever getting out the rubber glove and looking for hidden notes... Wnt (talk) 02:34, 22 March 2012 (UTC


 * His story specifically mentioned performing the "decompression" in a test environment. Anonymous.translator (talk) 11:51, 22 March 2012 (UTC)


 * As others said, mnemonic devices like memorizing letters which imply words is probably the closest thing to this that could be done by an average human. I've actually done this myself a few times for various reasons when it's necessary to memorize a handful of words or a phrase. In theory, someone with a pretty good memory could take this a step further by memorizing several acronyms, say CARS, ATLAS, TREES, and STICK, which represent words meant to be committed to memory, and then further "compressing" the information by arranging those acronyms into the level-two acronym CATS. Obviously, this is an overly simplistic representation of how it could be done, but maybe that gives you an idea. I suppose that, for someone with a truly eidetic memory, this isn't much of a problem. Evanh2008 (talk) (contribs) 03:13, 22 March 2012 (UTC)
 * Actually our article on eidetic memory presents a very skeptical view of it. It would certainly provide an uncommonly straightforward resolution to a great mystery... Wnt (talk) 03:43, 22 March 2012 (UTC)


 * Is your friend an idiot savant? Clarityfiend (talk) 05:22, 22 March 2012 (UTC)


 * I don't see why you'd need 20kb of data for a multiple choice test.
 * I you take a more reasonable example of 100 questions, you could probably craft a paragraph that encodes that data with no problem. Personally I'd go with Length_Of_Word%5. So that "I" "SAFETY" are both A, "IS" and "CENTURY" are both B, and so on.  Then you'd only have a 100 word paragraph to learn. Tough, but doable. Especially if the paragraph is funny or rhyming. APL (talk) 08:46, 22 March 2012 (UTC)
 * You're forgetting the fact that a human brain is not a computer! The objective in data compression is to reduce the amount of data.  APL's method is a good way to increase the amount of data, but to place it in a mnemonic format that is specially crafted to take advantage of the unique noise model of human cognition.  Unlike most machine storage system, the bit error rate of human memory is not fixed; it is directly correlated to the type of information being stored.  This means that an increased length of data may actually be easier to remember if it is about something that you can process cognitively, as opposed to "20 kilobits of semirandom alphanumerics."  As a good example, many people memorize religious texts verbatim, albeit with years of practice and repetition.  I strongly suspect that these religious participants could not memorize a equal-length random sequence of words and numbers, no matter how long they studied.
 * A critical requirement of lossless data compression is that it must not lose data on any data set (even if the compression ratio is greater than one). Most commonly-used lossless-data-compression algorithms are tuned to compress low-entropy human-like information, and actually perform quite poorly on statistically-random streams.  But, they guarantee correctness, and this is predicated on the perfect storage of the compressed form.  On the other hand, mnemonics "may" fail - partially.  So, a crucial element of a well-designed "cheat-on-the-test" mnemonic would be robust stream recovery.  Corruption in an early portion of the data stream should be isolated, so that it does not forward-propagate.  Designers of satellite television and audio CD data streams know how to do this!  First, data interleaving, such as the Reed-Solomon method, allows recovery from short "bursts" of poor memory ("I forgot that entire sentence of my mnemonic... but I remember the next phrase!")  Techniques to recover the lost data exist.  Robust, sophisticated, multi-bit checksumming allows verification that the decompressed data is not in error, with a high degree of certainty.  These techniques are what makes it possible to receive television video data from satellites in space, even though the signal must travel thousands of miles through clouds, rain, Earth's magnetosphere, and pigeons.
 * Who would have thought that so much computer-science, rocket-science, and electronics engineering would be inspired from "trying to cheat on a test"! In fact, the same motivations about saving time and space, while guaranteeing correctness, drive test-cheaters and rocket-scientists.  It so happens that rocket-scientists tend to perform better on tests; perhaps they understand the techniques better than the average cheater.  Nimur (talk) 17:47, 22 March 2012 (UTC)

I don't think you can assume that the goal is strictly lossless compression and storage. If 90% of the data came through OK that'd still be an A. APL (talk) 20:20, 22 March 2012 (UTC)
 * Similar discussion Suraj  T  08:41, 24 March 2012 (UTC)