Wikipedia:Reference desk/Archives/Science/2016 December 28

Welcome to the Wikipedia Science 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.

December 28

Confusing pubmed article: "skin defect of hell"

Was not sure whether to ask this in reference desk language, but would anyone know what this surgery abstract in pubmed would refer to by the term "skin defect of hell" https://www.ncbi.nlm.nih.gov/pubmed/12078310 ? — Preceding unsigned comment added by 2604:2000:7011:5C00:91A0:39FC:5873:5B0E (talk) 12:08, 28 December 2016 (UTC)

Ehm, yeah. "Hell" should probably be "heel". I googled "skin defect of hell" -"Surgical intervention in the treatment of skin defect of hell". I found http://www.rrsurg.com/article/zgxfcjwkzz/2000/4/232 and then I put the Chinese text in Google translate and it said: "Perioperative Management of Primary Repair of Skin Defect in Heel". (((The Quixotic Potato))) (talk) 12:16, 28 December 2016 (UTC)

I do not understand Chinese but Google Translate tells me that 足跟部皮肤缺损 means Heel skin defect (足跟 means heel). I have used the contact form to make them aware of the situation, maybe they'll fix the typo at some point in the future. (((The Quixotic Potato))) (talk) 12:18, 28 December 2016 (UTC)

By way of additional confirmation that it should be "heel", note the list of "similar articles" as well as the mentions of leg anatomy. --69.159.60.210 (talk) 11:09, 29 December 2016 (UTC)

I received a response, but unfortunately it wasn't what I was hoping for:

The National Library of Medicine enters article titles exactly as they are published in the journal. NLM indexed the above issue from the print version. I have confirmed that the above article English title published in both the table of content and the article page as “[Surgical intervention in the treatment of skin defect of hell]” and is cited in PubMed exactly as it is published in the print journal . In order for us to make a change in PubMed, you will need to arrange to have the journal publish an erratum notice, on a citable (numbered) page of a future issue. Once that is published, we will change the citation and add a bibliographic link to the erratum notice to alert the scientific community.

Please see the URL on errata:

http://www.nlm.nih.gov/pubs/factsheets/errata.html

Thank you for your interest in the accuracy of NLM’s databases.

(((The Quixotic Potato))) (talk) 12:00, 31 December 2016 (UTC)

(short-term) Death Toll from catastrophic Dam failure?

Does anyone have (short-term) death toll estimates from the catastrophic failures of the Aswan Dam and the Three Gorges Dam? Note, I'm looking for short term, those directly caused by flooding and building collapse as opposed to those from disease such as cholera, typhoid, etc. or dehydration due to lack of clean water. If possible, I'd like to additionally get information for any Dam whose short-term expected death toll is greater than these two (I doubt there are, but I'd like to be sure) *and* the one with the highest death toll in the United States. (Most of the dams in the US Southwest on the Colorado don't have *large* cities downstream (Yuma, Arizona isn't large by these standards)). If this belongs in RD:Misc, let me know.Naraht (talk) 14:32, 28 December 2016 (UTC)

IDK about the Aswan Dam or the Three Gorges Dam, but for your last question, the deadliest dam failure in the USA was that of the South Fork Dam in 1889, which caused 2209 deaths (see also Johnstown Flood). FWIW 2601:646:8E01:7E0B:E5C0:72CB:AF6F:FD32 (talk) 03:41, 29 December 2016 (UTC)

Thank you. Studying actual deadly dam failures is a start. I would expect dam failures for the Aswan or Three Gorges to be at least two magnitudes higher.Naraht (talk) 15:29, 29 December 2016 (UTC)

The Banqiao Dam failure was the deadliest in history, causing an estimated 26,000 direct deaths and ~150,000 - 200,000 indirect deaths from famine and disease. In that case the immediate cause of the failure was a "once-in-2000-year flood" that significantly exceeded the dams design capacity, though deficient construction is also considered a possible cause. One of the interesting things to consider is that Banqiao involved cascading dam failures. The loss of the first dam sent so much water and debris down stream that subsequent dams also breached. Altogether 62 dams failed in a single flood. Dragons flight (talk) 12:30, 30 December 2016 (UTC)

Why isn't alcohol used to treat bacterial skin infections?

So alcohol kills 99.999% of bacterial within 15 seconds. Why isn't it used to treat bacterial skin infections instead of antibiotics? — Preceding unsigned comment added by 177.37.160.211 (talk) 16:40, 28 December 2016 (UTC)

Because skin infections are not usually on the surface of the skin - and if the patient either took the alcohol by mouth, or had it injected, it would be very quickly metabolised. Using alcohol swabs to clean the surface of a wound has a long tradition - but it is only effective on the surface. Wymspen (talk) 18:43, 28 December 2016 (UTC)

That figure is very optimistic, I think. But the main point there is that if the bacteria are not in you then it doesn't take much of a magic bullet to kill them. You could probably wash with soap or wave your hand over a Bunsen burner and kill/remove a lot of bacteria also. But in between live cells, your body can't have 70% ethanol, no more than it could put up with Bunsen burner temperatures or dish soap. It would be like trying to fight terrorist infiltrators with nuclear weapons! Wnt (talk) 18:51, 28 December 2016 (UTC)

If it's not causing disease, a bacterium on your skin is a colonization, not an infection. You don't want to kill them all; you only want to kill the ones likely to cause infection. The other ones - by being present - can help prevent the potentially pathogenic ones from predominating and causing other problems. In other words, the skin has a "normal flora" - bacteria that are supposed to be there - the absence of which can cause problems. - Nunh-huh 18:57, 28 December 2016 (UTC)

All true, but the OP's premise was the case of an infection. Although many antibiotics have some specificity, for example targeting Gram-positive cell walls, there are not many things, apart from a few drugs against weird bacteria like isoniazid, or phage therapy, that really go after the bad actors and leave the bacterial bystanders alone. Wnt (talk) 21:25, 28 December 2016 (UTC)

Neosporin ointment is routinely used in health care providers colonized with MRSA in order to kill the MRSA and permit recolonization with methicillin sensitive Staph aureus, which is less likely to harm those they care for. - Nunh-huh 20:40, 29 December 2016 (UTC)

One might potentially debride a recalcitrant sore with drunken maggots. μηδείς (talk) 04:19, 29 December 2016 (UTC)

Insects are amazingly tough, but I think 70% ethanol would also be too much for them. Wnt (talk) 12:41, 29 December 2016 (UTC)

I simply said drunken, not pickled in 140 proof vodka. μηδείς (talk) 20:01, 29 December 2016 (UTC)

My anecdotal experience is that 70% ethanol will kill ants and roaches just fine, but not mites. Someguy1221 (talk) 08:52, 31 December 2016 (UTC)

I wish I were toasting at your house tonight, so long as you don't also have bedbugs. μηδείς (talk) 01:05, 1 January 2017 (UTC)

Artificial intelligence

Many sci-fi films depict scenarios in which humans are outsmarted by artificial intelligence but is this actually a possibility in reality? 2A02:C7D:B937:6300:B46E:5B0E:8B82:3412 (talk) 21:13, 28 December 2016 (UTC)

It is not possible at the moment. It might become possible in the future. Stephen Hawking is certainly concerned about that possibility - http://www.bbc.co.uk/news/technology-30290540 Wymspen (talk) 21:24, 28 December 2016 (UTC)

How about a nice depressing match of Computer chess? Really, we'll set it low, you might have a chance. Wnt (talk) 21:26, 28 December 2016 (UTC)

This has some merit. My garden-variety PC outsmarts me with alarming frequency, so AI can be said to be able to outsmart a human in tasks for which some higher level of cognition isn't required, but just an extreme ability to assess possibilities (i.e. a chess match in which a computer can just map out possible near-term positions based on the board's setup). Tyrol5 [Talk] 00:34, 29 December 2016 (UTC)

History of artificial intelligence and Philosophy of artificial intelligence are two nice long articles with good references. Then there's Artificial_intelligence. See also Weak AI, Hard AI, and AI-complete. As Wnt alludes, computers are already better than us at chess. And now even playing go. And all kinds of other things; machine learning is now entering a phase where any kid with some coding skills can jump in and make computers do crazy "smart" things [1]. But these are generally seen as "soft" or "narrow" forms of AI, and we don't have anything yet that we think can truly think. It gets hard to talk about. Intelligence itself is hard to define, and can mean many rather different things. Hard problem of consciousness outlines some of the thoughts we've had about thoughts and consciousness. SemanticMantis (talk) 21:55, 28 December 2016 (UTC)

Two very thoughtful web sites on the question of AIs taking over:

• https://what-if.xkcd.com/5/
• http://idlewords.com/talks/superintelligence.htm

I highly recommend both. --Guy Macon (talk) 23:57, 28 December 2016 (UTC)

• Not quite as good as the above two, but still quite interesting: https://www.explainxkcd.com/wiki/index.php/1002:_Game_AIs --Guy Macon (talk) 00:13, 29 December 2016 (UTC)

• Intelligence is what a baby has; it learns about its environment spontaneously, beginning with perception and recognition, then moving to concept formation and induction, often without formal training, and certainly without digital programming. When you are beaten at chess by a computer executing an algorithm, you are being beaten by the programmer, just at a distance, using the programmer's algorithm executed at a higher speed by an electronic device that would never have developed from a single fertilized transistor to a motherboard running software with mechanical devices capable of discreet inputs and outputs. The Coming Grey Goo! 04:15, 29 December 2016 (UTC)

Another very relevant article is Turing test Wymspen (talk) 11:02, 29 December 2016 (UTC)

To some extent. But the restrictions put on that test compromise the idea behind it. ←Baseball Bugs ^{What's up, Doc?} carrots→ 13:14, 29 December 2016 (UTC)

The Turing test is the same nonsense. It proves nothing about the "intelligence" of the machine. It shows whether the person conversing with the computer is stupider than the programmers. See the Chinese room. μηδείς (talk) 20:00, 29 December 2016 (UTC)

If you do, you will find that "The overwhelming majority, still think that the Chinese Room Argument is dead wrong". The Turing test is basically a simple instance of the Duck test. You are appealing to some non-observable special property - as in "a well-educated <ethnic group of choice> can behave just like a human, but of course they have no soul". That is both unscientific (because you require some hidden, unobservable property) and violates Occam's Razor (because doing that multiplies entity without good reason). --Stephan Schulz (talk) 02:24, 30 December 2016 (UTC)

Seriously? Your answer, Stephan, is to call me a racist? I'll take that ad hominem as an admission of defeat on your part, and ask you to please play with your straw men in private. I have defined my idea of intelligence in my comment of 04:15, 29 December 2016 (UTC), and it is quite observable.

I am also quite aware that there are many people who are taken in by the Turing Test, just as there have been behaviourists and Chomskyans. But they are either confused or equivocating. I am quite sure that some artificial construct that possesses intelligence might some day be possible. But the Turing Test tells us absolutely nothing about the consciousness or intelligence of a machine running a program simulating intelligence. It only tells us how clever the programmer is, and how gullible the test-administrator is. μηδείς (talk) 19:39, 31 December 2016 (UTC)

It might be "unscientific" in the sense of not having a scientific explanation, but it is eminently scientific in terms of bowing to observation. There is, simply put, no law of physics that says that under condition X, when a photon strikes an electron, the electron actually feels it. We experience the phenomenon of qualia and we don't have a scientific explanation, and so way have no scientific way of evaluating whether it is duplicated. But although the property is inexplicable, it is not unobservable; it is indeed everything we have observed, and accounts for basics like why we're observing an atypical planet with life rather than being rock trying to comprehend a typical uninhabited world based on how the vibrations in the crust affect ourself. Wnt (talk) 15:09, 30 December 2016 (UTC)

Playing a chess game on a gross-master's level by a computer has nothing to do with artificial intelligence. Discovering a wave equation or Special Relativity are. Machines will never do anything like this. --AboutFace 22 (talk) 16:42, 29 December 2016 (UTC)

Nor can most humans. {The poster formerly known as 87.81.230.195} 2.122.62.241 (talk) 20:26, 30 December 2016 (UTC)

Also see Automated Mathematician for a program that creatively (?) discovered new mathematical theorems. Playing chess well was generally seen as one of the hallmarks of high intelligence - until machines managed to do it with a combination of brute force and heuristics. But if we declare anything we do understand to do as "not real intelligence", then we are in the God of the gaps situation, where "real Intelligence" is shrinking every time AI makes some progress. See no true Scotsman. --Stephan Schulz (talk) 00:06, 31 December 2016 (UTC)