Wikipedia:Reference desk/Archives/Science/2010 December 7

= December 7 =

How to identify something I had photographed in the sky this morning?
Dear RefDeskers, this morning, using my 40D with a 24-135 mm lens, I took a shot at a bright object in the morning sky. Turns out it's probably natural in origin, not artificial, because it came out vaguely round. How to identify what it is? Will anyone be able to tell me if I tell you my geographical location and the direction I was looking in? It's Gdańsk, Poland, object almost straight due East, twentyish degrees above the plane of the horizon. Curiosity, yes, nothing more! Cheers and thanks, Ouro (blah blah) 05:01, 7 December 2010 (UTC)


 * Yourhorizon at Fourmilab shows Venus low on your SE horizon about now, that's probably it. Jupiter is just above it.   Wiki Dao  &#9775;  (talk)  05:29, 7 December 2010 (UTC)
 * Wow, useful tool! Would be Venus then, though it's much more E than SE over here. Brighter than stars in the sky. Thanks, WikiDao! --Ouro (blah blah) 06:04, 7 December 2010 (UTC)
 * Earth is about as close to Venus as it ever gets: 0.435 AU at 2010-12-07 6:35:47 UTC. This is almost the closest pass during the entire year.  KStars said Venus is around a Magnitude -4.3; Sky and Telescope says -4.8.  (Very bright).   Nimur (talk) 06:38, 7 December 2010 (UTC)
 * I'm gonna go ahead and mark this as resolved. Thanks loads, guys! Was worth it getting up early today :) Ouro (blah blah) 06:57, 7 December 2010 (UTC)
 * There exists lots of software that can show what the sky looked like at a particular time. I like Stellarium (computer program), which is quite user friendly in allowing you to choose a location and a time to see what's in the sky. Buddy431 (talk) 15:50, 7 December 2010 (UTC)

LED Street Lights
My city (Seattle) recently installed new "energy-effecient" LED street lights on my block. They seem awfully bright and I was wondering if they'll dim over time as they get nearer their end-life or if the spectrum would likely be at all adjustable, maybe some way to reduce the "spillover" into bedroom windows and such? Definitely have their upsides it sounds like, but I do find myself suddenly nostalgic at seeing the contrasting dull yellow glow of the old sodium lights. Ah, well. —Preceding unsigned comment added by 67.160.87.224 (talk) 05:09, 7 December 2010 (UTC)


 * Yes, the primary failure mode of LEDs is to gradually get dimmer. However, they have a very long lifespan.  See LED and List of LED failure modes.  Red Act (talk) 05:33, 7 December 2010 (UTC)


 * New LED streetlights may well be part of a street light control system.  If you apply to your relevant local authority they may well be able to dim your nearest lights for the hours that you find most annoying. There also seems to be a feedback page at seattle.gov that you could use to register your opinion.  Blakk   and ekka 15:14, 7 December 2010 (UTC)


 * This brings up an interesting Q: Could they compensate for dimming of LEDs over time by putting in more than are needed, initially, but leave some off (or at lower light levels) at present, to keep from blinding people ... then gradually turn more on as they dim, so that the final brightness is constant, until they are scheduled to be replaced ? StuRat (talk) 04:51, 8 December 2010 (UTC)

Testing for tobacco use
Settle an argument; how long after tobacco use would evidence still be present in a blood sample? --John (talk) 06:46, 7 December 2010 (UTC)
 * The in vivo half-life of nicotine is only about 2 hours. However, blood tests for tobacco usage also test for cotinine, the primary metabolite of nicotine, which has an in vivo half-life of about 20 hours, and is detectable in the blood for days (up to about a week) after the use of tobacco.  See Nicotine and Cotinine.  Red Act (talk) 07:37, 7 December 2010 (UTC)
 * Thank you! --John (talk) 14:55, 7 December 2010 (UTC)

michael douglas
will michael douglas die —Preceding unsigned comment added by 66.66.92.152 (talk) 10:54, 7 December 2010 (UTC)


 * Yes. Dismas |(talk) 11:03, 7 December 2010 (UTC)
 * Eventually. TomorrowTime (talk) 11:06, 7 December 2010 (UTC)
 * We're all dying, in some sense of the word. Shadowjams (talk) 11:08, 7 December 2010 (UTC)
 * For reference, this is probably in regards to Michael Douglas, who has said that he has Stage IV Throat Cancer. That is, the cancer has metastasized, or spread through the body.  If I understand correctly, cancers that have metastasized are rarely curable, but, depending on the circumstances, may be treated with a varying amount of success.  I don't know enough about Mr. Douglas' case to say how it's going to end, but he'll almost certainly die with the cancer, and probably because of it.  How long this will take is dependent on a variety of factors. Buddy431 (talk) 15:46, 7 December 2010 (UTC)

Climate change
From a synoposis of BBC documentary series "How the Earth Made Us": "Having documented the planetary forces that have shaped the development of civilisations, Iain Stewart discovers how this relationship has been reciprocated. He examines how the advent of farming prevented the world from slipping into a new ice age, and demonstrates why humanity unwittingly caused a catastrophic mudslide in Indonesia." I'm interested in knowing the scientific standing of the idea that the introduction of farming had an effect on global climate. It must be based on at least one scientific paper, but how recent, how established, is this idea? Itsmejudith (talk) 12:51, 7 December 2010 (UTC)


 * There is this paper:how did humans first alter global climate?. Also we have an article on Climate change and agriculture.--Aspro (talk) 15:15, 7 December 2010 (UTC)
 * Thanks. The Ruddiman paper is what I was looking for. I should have remembered we would have an article but it doesn't go into the Ruddiman hypothesis. Itsmejudith (talk) 15:40, 7 December 2010 (UTC)

Railgun vs. Coilgun efficiency
Which is more efficient at turning electric energy into kinetic energy? ScienceApe (talk) 14:40, 7 December 2010 (UTC)
 * My guess is coilgun, but I think these devices are too individual for any overall answer. Especially the coilgun that recycles the energy from each cycle (see the article) would be pretty efficient. Railguns waste a ton of energy as heat - but on the other hand can handle way more power. Ariel. (talk) 12:14, 8 December 2010 (UTC)

Standard practice rounding significant figures
I read in Schaum's Outline of Beginning Chemistry the following: If the first digit to be dropped is a 5, and there are no digits or only zeros after the five, we change the last digit remaining to the nearer even digit. The following numbers are rounded to one decimal place: 5.550 -> 5.6, 5.55 -> 5.6, 5.450 -> 5.4, 5.45 -> 5.4

Is this standard practice? I never heard of this before, and would have thought to make 5.450 and 5.45 change to 5.5 20.137.18.50 (talk) 16:16, 7 December 2010 (UTC)
 * That's the "Round half to even" method discussed in Rounding. See the associated discussion there of the other ways to do it, the pros and cons of them, etc. DMacks (talk) 16:31, 7 December 2010 (UTC)


 * (ec)Always rounding the mid-point ("xxx.5") up leads to a systematic error - e.g., the "apparent" average value of a collection of numbers is now slightly larger than the true average value if had been computed without rounding. This is problematic because it affects the statistical analysis of huge sets of data.  It also affects numerical or computer algorithms because it introduces an error that "grows".  This can turn a problem that is analytically stable into one that is computational unstable.  We have a complete section at rounding to integer discussing several algorithms, such as "always round up"; "always round toward zero" (... think about what "rounding up" means for negative numbers); "always round toward the nearest even-number"; "pick a random direction to round"; and so forth.  In IEEE-754 (the standard for computers to represent fractional numbers), the programmer can specify a rounding algorithm to guarantee a specific behavior.  Nimur (talk) 16:32, 7 December 2010 (UTC)


 * That being said, this is an issue that has pretty much lost relevance in this age of ubiquitous computation -- no sensible programmer would use the rounded data for any kind of important analysis, you'd always use the actual numbers stored to the maximum precision of your database, and if rounding matters at that level, you ought not to be doing that analysis in the first place. Looie496 (talk) 17:05, 7 December 2010 (UTC)
 * Numerical analysis is the science off determining how to compute accurately - which means correctly accounting for the errors you describe - deciding what precision needs to be stored. In a simple case, like financial mathematics, you are right - a programmer can simply use integer-math where necessary.  But in other cases, like scientific computing or controls, this issue becomes much more complicated: suppose you're building an autopilot; sampling flight-data at 10 MHz at 12 bits.  If every 10th sample has a 1-bit rounding error, and you're integrating the result for a control algorithm, you accumulate an error of 1 million samples every second.  If you didn't realize this, you might start pitching the plane up.  Entire textbooks; entire careers, even, are made on the subject of correct numerical handling to make sure you are not running up accumulated error.  Rounding Errors in Algebraic Processes, Error, Accuracy, Stability] (from the infamous Numerical Recipes).  In this era of ubiquitous computing, no sensible programmer should assume that "the computer will just do math correctly" - for mission-critical systems, programmers need to actually understand the way calculations are performed and how to manage them.  Nimur (talk) 18:09, 7 December 2010 (UTC)
 * In this day and age, I would think any analysis with a large enough amount of data for rounding rules to matter would be done by a computer. Most likely the numbers will be in binary form when the rounding is performed. So when analyzing how different rounding rules might affect a result, the radix of the number at the time of rounding must be considered. Jc3s5h (talk) 17:16, 7 December 2010 (UTC)


 * This has the interesting(?) side effect that, in Windows, UNIX and OpenVMS at least, if you add 252 (4,503,599,627,370,496) to a number then subtract it again, the result is the number rounded to the nearest whole integer. Gzuckier (talk) 13:52, 8 December 2010 (UTC)

Transduction, transformation and transfection
Do these terms actually apply only to particular cell types (e.g. pro- vs eukaryotic)? If so, is there a pragmatic reason? --129.215.47.59 (talk) 21:44, 7 December 2010 (UTC)
 * Our transfection article explains these distinctions right at the top. Looie496 (talk) 22:30, 7 December 2010 (UTC)

Sun-Jupiter Barycenter
How far above the surface of the sun is the Sun-Jupiter barycenter? ✍ (talk) 21:58, 7 December 2010 (UTC)
 * About fifty thousand kilometres, as I just calculated but is anyway stated at barycenter. Algebraist 22:04, 7 December 2010 (UTC)
 * I don't see it. ✍ (talk) 14:35, 9 December 2010 (UTC)
 * The examples table at barycenter has a Sun-Jupiter row where r1 - R1 = 742,000 km - 696,000 km = 46,000 km. By the way, it seems the center of mass of the whole Solar system moves between inside and outside the Sun (where is it now?). According to my quick calculations it will very rarely be on the opposite side of Jupiter inside the Sun. PrimeHunter (talk) 19:20, 9 December 2010 (UTC)

Bermuda rigs
How good are Bermuda rigs when heading directly downwind? Are lateen sails or gaff rigs any better for this? --  T H F S W  (T · C · E) 23:32, 7 December 2010 (UTC)
 * Ask a commercial concern such as http://www.riggingonly.com Ginger Conspiracy (talk) 18:56, 8 December 2010 (UTC)