Jump to content

Wikipedia:Reference desk/Archives/Science/2013 November 23

From Wikipedia, the free encyclopedia
Science desk
< November 22 << Oct | November | Dec >> November 24 >
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.


November 23

[edit]

Genetically engineering hair color and eye color

[edit]

Is it true that hair and [[eye color are coded for by a single gene, and therefore would be the easiest to genetically modify (compared to other traits)? ScienceApe (talk) 02:55, 23 November 2013 (UTC)[reply]

No. I have linkified the relevant articles if you want to read them. μηδείς (talk) 03:05, 23 November 2013 (UTC)[reply]

Redshift experiments

[edit]

The existence of dark energy AFAIK was deduced from differences between type-1 supernova derived distance and redshift-derived velocity of high-z galaxies. Surely all alternative hypotheses explaining these differences have been ruled out? So, for example, nonlinearity of redshift in the optical wave lengths at high-z velocities in general can be rejected how? Direct experiments? Which direct experiments measuring redshifts at the relevant velocities have been done? Any hint is welcome. --SCIdude (talk) 11:16, 23 November 2013 (UTC)[reply]

The origin of these hypotheses is Einstein's cosmological constant, which was explicitly introduced as a fudge factor rather than as part of a more "elegant" theoretical equation. The astronomical data you mention establish that the cosmological constant is, in fact, non-zero and has a definite measurable value. Dark energy is an hypothesis that attempts to explain why the cosmological constant is non-zero based on known existing physical phenomena - on metaphysical grounds, the answer "it just is, it's an inherent property of the vacuum that we can only measure rather than predict" isn't (apparently) acceptable. Tevildo (talk) 11:37, 23 November 2013 (UTC)[reply]
No, all alternate hypothesis cannot ever be ruled out. Your un-orthodox nonlinear redshift would require some huge revolution of physics though and is considered to have a negligible a priori probability. Dauto (talk) 14:12, 23 November 2013 (UTC)[reply]
So the Nobel prize is rather for the SN1A observation, like the one for CMB discovery by Penzias/Wilson. I misunderstood that, thanks. - SCIdude (talk) 18:31, 23 November 2013 (UTC)[reply]
To answer the other part of your question, about experiments that directly measure redshift at high velocities: Novotny et al. have done spectroscopy on ion beams at 0.33c, not too different from the recessional velocity of supernovae at z=1 or so [1]. This confirms the expected shifting of spectral lines. Moreover, synchrotron light sources and free electron lasers use ultrarelativistic electrons with Lorentz factor much higher than anything that comes into play for supernovae. If you try to somehow tweak the Lorentz transformations to give a different result for cosmological redshift, you will almost certainly break SSRL and FELs in a way that would be immediately noticeable. People do still perform experimental searches for Lorentz violation, but they're looking for tiny effects that would make no difference at all for the original data that provided evidence for accelerating expansion back in the 1990s. --Amble (talk) 22:35, 23 November 2013 (UTC)[reply]
Very good. Many thanks. Could you please augment the redshift article to include the relevant parts of your answer? - SCIdude (talk) 05:24, 24 November 2013 (UTC)[reply]
I'm afraid that might fall afoul of WP:SYNTH without some secondary sources to make the connection. --Amble (talk) 16:40, 24 November 2013 (UTC)[reply]

Structural elements

[edit]

Are there any good diagrams which show which part of a cantilever structure are trusses, supports and cantilever members? And which ones are beams?Clover345 (talk) 12:20, 23 November 2013 (UTC)[reply]

This page goes through the terminology at a very basic level. Tevildo (talk) 12:53, 23 November 2013 (UTC)[reply]
I've looked for such sources a few times myself. Questions like what's the difference between beam and girder (http://www.eng-tips.com/viewthread.cfm?qid=164292).
Cantilever seems to be the easiest to identify: it's an overhanging part of a construction. A beam, a truss or any other construction can be supported at both ends; but you can also support it for example at the left end and somewhere in the middle, in which case the part of the construction to the right of that point is considered a cantilever.
A truss is a structural "unit" made up of straight elements connected as triangles. List of truss types should give you a good idea. You could consider it a metal plate where much of it is cut out to save weight but still keep the stiffness. The term is mostly used for a two dimensional structure, not sure if for example an antenna tower is considered a truss or a construction made of several trusses. The Allen truss bridge shown here has three trusses (or six, when you cross a bridge you have a truss to your left and a truss to you right) not one, because the parts are connected at only one point, so the bridge as a whole isn't stiff.
A support can mean different things. For example Truss#King_post_truss calls the top chords of the truss "supports". Cantilever uses the word for the points where the structure is "supported", the "external" connections you could say, which is the most common meaning I think. It can also mean the support column a truss rests on.
A beam is usually an element that has to bear a bending load: for example, a bridge can consist of two trusses, in between them is the deck that carries the traffic, this deck rests on top of beams that are connected to the two trusses. While the elements that make up the trusses only have to withstand either tensile or compressive forces (and a minor bending force due to their own weight), the beams are only or mainly subjected to bending forces.
Vocabulary of trusses explains some of the terminology and has diagrams. Keep in mind that the terms aren't always used consistently. Ssscienccce (talk) 17:28, 23 November 2013 (UTC)[reply]

How (non-rad) hard would it be to turn every cellphone into a radiation detector?

[edit]

https://medium.com/war-is-boring/21dc0b023f1d Homeland Security Agency Wants to Turn People Into Nuclear Tripwires

Why not just mandate that this be used:

http://www.researchgate.net/publication/253023219_Detection_mechanisms_employing_single_event_upsets_in_dynamic_random_access_memories_used_as_radiation_sensors

To have every cellphone track the Single event upsets in its own chips and report to the cellphone network provider (and hence of course the NSA), whenever the rad counts hop over the background level? Hcobb (talk) 20:08, 23 November 2013 (UTC)[reply]

That would lead to too many false positives (from radon exposure, cosmic rays, even thunderstorms). Cell phones CANNOT reliably detect radiation that way -- this is a job for a dedicated instrument (such as that which DHS wants to use). 24.23.196.85 (talk) 21:06, 23 November 2013 (UTC)[reply]
I have to agree with 24.23.196.85 in part on this. At the current sophistication of cell phones they may be able to detect ionizing radiation but to be useful the Signal-to-noise ratio would need to be clear. However, having said that, if a thousand cell phones all reported an indication of a higher level of ionizing radiation above background, then yes, they could be used as trip wires.--Aspro (talk) 21:46, 23 November 2013 (UTC)[reply]
The memory chip area available for use as a detector in a cell phone would, for a start, be far too small, meaning no detection events for months at least, even with significant increases in radiation levels. Because the chip is enclosed in the cellphone case and hidden by internal structures, it would essentially only detect cosmic rays - not what Homeland Security would be most interested in. The electronics industry / electronics hobbyist magazine Elektor carried a series of articles last year or ealy this year on making a home made radiation detector using a standard commercially available semiconductor device as a detector, instead of a gieger tube. Read that article, and you'll see that a detector chip has to have sufficient area (of the order 1 x 1 mm - vastly bigger than even a large number of DRAM memory cells), and be enclosed in special pakaging that blocks light completely while letting low energy particles through. 60.228.240.47 (talk) 00:12, 24 November 2013 (UTC)[reply]
Memory uses error correction and detection. It used to be parity detection, but they grew the word line and sophistication so a 256 bit line can add 10-12 bits to correct either single bit or multiple bits. Usually this goes unseen to the user as the 270 or so bits still return a 256 bit corrected line. The ECC is flagged on access, not when the event happened. It cannot identify the cause of failure. An unreliable cell or noise could cause the error. As long as the device can correct it, it continues. Multi-bit failures that exceed the ECC capacity generate the parity type of halt/BSOD hardware error. The next question is how separate is neutron radiation from alpha particle radiation? Alpha particles dominate SEU from ionizing radiation in semiconductors not the least of which is the use of lead in packaging (although being phased out). I do not know if neutron radiation will cause enough ionization failures over background alpha radiation to be reliably detected. Feature size also makes cumulative damage difficult to measure in a single device. The mission for memory is to correct, not accumulate errors. To be honest, I think the best starting points are more like the CMOS digital camera arrays with methods to filter and isolate the radiation being sought. --DHeyward (talk) 04:44, 24 November 2013 (UTC)[reply]
Note that most phones do not have ECC. I'm not aware if there is even a LPDDR3 (or any LPDDR) standard for ECC, our article Mobile DDR doesn't mention one nor do I find anything from a quick search.
And it's perhaps worth remembering that most desktops and laptops don't have ECC memory either. (Remember you will generally need a Xeon on the Intel side. Although on the AMD side there is limited support with some/many?/most?/all? of the AM3+ CPUs but many of the motherboard manufacturers still don't bother to official support it.)
The CPU cache may or may not have ECC I'm not totally sure.
Either way, most desktops and laptops, let alone mobile devices don't even have the capability to clearly detect many errors. This isn't surprising, as many above have indicated, in normal circumsances these errors are so rare, they just aren't worth worrying about for non mission critical systems. (Some people even use servers an workstations without ECC.)
Sure you can do some sort of test with a known result like people do for stress testing to detect errors, but that won't do wonders for your battery life.
(Just to emphasise the earlier point, if you have done stress testing, you would know whether one of the memtest variants or whatever, with a stable system with nonECC RAM you can run for over a week or more with no errors on any test. Incidentally as far as I know from such tests, you can generally obtain reports of ECC failures from the CPU if you want to.)
Now I guess you could mandate ECC, some may even be happy for it for the obvious other benefits. But to force a feature which is not going to work very well for what you're want to achieve, when you could at least force something which is designed for the purpose, just seems dumb. Although I share concerns with our resident cranky Perth engineer who likes to pretend to be multiple people, that you still won't be able to achieve much given the size and design limitations and perhaps cost ones as well (edit: for clarification I mean even if you require chip purposely designed to monitor radiation levels).
Nil Einne (talk) 13:15, 24 November 2013 (UTC)[reply]
ECC is not just DRAM. SRAM on the microprocessor incorporates ECC. It's inherent and invisible to the system or user. ARM and Intel both have ECC on L2 Caches. --DHeyward (talk) 05:10, 25 November 2013 (UTC)[reply]
Not to mention that cellphone manufacturers really want to MINIMIZE the number of RAM errors - not MAXIMIZE them! What would be needed would be some kind of additional detector chip - and the cost of that makes this proposal impractical unless somehow a law were passed to require it - which seems overwhelmingly unlikely. SteveBaker (talk) 16:48, 24 November 2013 (UTC)[reply]

Oil for plastics

[edit]

I think petroleum products are used in the manufacture of certain polymers. When viable oil runs out, will we see certain types of plastic that we still have now disappear? Are there already alternatives in place? — Preceding unsigned comment added by 78.148.107.181 (talk) 23:04, 23 November 2013 (UTC)[reply]

Bioplastic. -- Finlay McWalterTalk 23:44, 23 November 2013 (UTC)[reply]
Hey, thanks. I was aware that there were already polymers that could be manufactured using non-crude-oil-derived plastics but I was wondering whether there were certain applications not yet covered. — Preceding unsigned comment added by 78.148.107.181 (talk) 23:58, 23 November 2013 (UTC)[reply]
The earliest plastics were made from cellulose; see: Parkesine. -- Also: History of Plastic ...And from milk (casein); see: Casein / The Plastics Historical Society. ~E:71.20.250.51 (talk) 05:57, 24 November 2013 (UTC)[reply]
See also Fischer–Tropsch process. In addition to carbon monoxide, it is also possible to turn methane or carbon dioxide into synthetic hydrocarbon compounds, which could then be used to make traditional polymers. So running out of fossil fuels will not make traditional polymers disappear, but it would make them more expensive (digging something out of the ground is often far cheaper than making it from scratch). Due to this increased expense, you may still wind up seeing bioplastics instead. Someguy1221 (talk) 06:13, 24 November 2013 (UTC)[reply]
Indeed, we can make them, doesn't mean we will. Hard to find any Bakelite today. Unless you go to a chinese webshop. "Replacement Bakelite Saxophone Mouthpiece", two belgian inventions for the price of one! Ssscienccce (talk) 09:57, 24 November 2013 (UTC)[reply]
Firstly, I don't think we'll ever run out of oil. If we keep burning it at the rate we are, we'll cook the planet long before we run out of the stuff. However, if we somehow did, we can make perfectly good oil-substitutes from plants, bacterial mats, algea and so forth. The problem is that it takes a lot of energy to do that...often more than you get back from burning the resulting oil. However, for the purpose of making plastics, the energy consumption isn't likely to be the primary issue.
What might be a problem is that as we burn through the easy-to-dig-up oil we'll have to start being more agressive about mining oil shale and oil sands - which will certainly push up the cost of oil - possibly to the point where we don't want to use it for making plastics anymore.
Fortunately, there is no shortage of alternatives. We already make plastics like PLA (Polylactic acid) from corn starch or sugar cane...and PLA is a perfectly useful plastic for many applications. You can make sheets of the stuff, extrude it in a 3D printer, injection mold it and so forth. It's also biodegradable and recyclable - which is useful - and it's easy to adjust vrious chemical properties to vary the time before it degrades, the melting point and so forth.
There are many other useful plastics that can be made from plant material without going through the intermediate step of converting it to oil. Cellophane is made from wood pulp, cotton or hemp. Plastarch is like PLA but better for high temperature applications. Then there is PHB (Polyhydroxybutyrate) - which is made using micro-organisms. Then there are whole families of exotic materials like amorphous metal glasses which can be stronger than titanium and yet heat-formed just like a plastic.
Hence, I don't think we have anything to be concerned about here.
SteveBaker (talk) 16:43, 24 November 2013 (UTC)[reply]
I don't think he's concerned about running out of oil, he's asking if there are certain kinds of plastics that will disappear once the oil used to make them is no longer available. As you note, it's much more likely that we'll render oil too expensive to extract before we literally run out of it, but that doesn't alter the premise of the question. For example, polyvinyl chloride is the third most widely used plastic (according to our article) and it's made from the vinyl chloride monomer which is derived from petrochemicals. Once oil becomes too pricey, will we no longer have any PVC? I don't know, but that would seem to be the implication. Polyethylene is the most widely used plastic and our article notes that it's available in bioplastic form, so presumably, it's not one of the forms that would disappear (not that plastic really disappears, but you get the drift). Our article on polypropylene (the second most widely used plastic) is so full of jargon, I have no idea how it's actually made in real-world terms and so can't even register a guess. Matt Deres (talk) 18:35, 24 November 2013 (UTC)[reply]
PP, PE and PVC can also be made from natural gas or from coal -- so they won't "disappear" even if we run out of oil. 24.23.196.85 (talk) 03:09, 25 November 2013 (UTC)[reply]
Vinyl chloride is made from ethylene, which can be produced by steam cracking of fossil fuels, but the same process can also be applied to bio-ethanol. Basically, thermal cracking, steam cracking, hydrocracking apply high pressure and temperature to create a mixture of lots of chemicals, catalysts are added to increase the yield of useful products, the resulting mix is separated by distillation, unwanted chemicals are put through another cracking process. After all, heat and pressure is what produced the oil in the first place. It may be more costly when starting from biofuel, but that's no reason why it wouldn't be used. Look at gasoline production: light crude oil contains a lot of gasoline and diesel that can be extracted by simple fractional distillation, but we're also mining tar sands and using energy intensive processes to produce them. Ssscienccce (talk) 19:03, 25 November 2013 (UTC)[reply]