Wikipedia:Reference desk/Archives/Science/2010 July 20

Science desk
< July 19	<< Jun | July | Aug >>	July 21 >

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.

July 20

Unidentified Insect

Hi. I was hoping for some help from anybody with knowledge of British entomology.

Earlier today, I had an insect fly into my house in Devon that I don’t believe I’ve ever seen before. It appeared bigger than a common wasp but rather smaller than a hornet. It had a round but, perhaps, slightly flat yellow body with two extremely well defined stripes circling the abdomen (this was the most immediately striking thing about it – the stripes were not blotchy and poorly defined as on most bees and wasps but perfectly defined, like you’d expect to see on a T-shirt and there were definitely only two stripes) and very big, bulbous grey eyes just above and either side of a bright yellow wasp-like mouthpart. Its thorax looked to be a dark brown colour (perhaps with a hint of red) and its legs (which seemed very long for a wasp) were most definitely reddish. It was so wasp-like that I assume it’s a member of hymenoptera but, given my lack of insect knowledge, I suppose it could conceivably be something else indulging in a bit of mimicry, although I doubt it. Any ideas?

Thanks Pantscat (talk) 00:54, 20 July 2010 (UTC)

Can you get a photograph of it?

See anthing similar in http://commons.wikimedia.org/wiki/Category:Wasps

Some mimic images here http://www.pbase.com/tmurray74/wasp_mimics&page=all

Hoverflys might have the flat abdomen - the eyes are distinctively 'flyish' though eg http://www.google.co.uk/images?q=hoverfly&um=1&ie=UTF-8&source=og&sa=N&hl=en&tab=wi&biw=1280&bih=937 178.78.65.223 (talk) 01:30, 20 July 2010 (UTC)

(edit conflict) I think that most "like a wasp but not a wasp" sightings turn out to be hoverflies, but there are so many different species of the critters that identifying the one that visited your house may be difficult. Do any of the ones pictured here look at all like your visitor? Deor (talk) 01:35, 20 July 2010 (UTC)

There's a list of common uk hoverflies here http://www.microscopy-uk.org.uk/mag/artmay07/cd-hoverflies.html I can't see your two stripe one, but your description of "big bulbous eyes" suggests a fly and not a wasp - which have big eyes - but I wouldn't call them bulbous.178.78.65.223 (talk) 01:40, 20 July 2010 (UTC)

I did get a photo of it but I'm not sure how good it was and am not at all sure how I would get it onto here even if it was a good one. I haven't yet found out what it was but I'm now reasonably confident that you guys are right and that it was a hoverfly (although I didn't realise they could be that size) as it looked similar (although not identical) to Epistrophe grossulariae. http://bugs.decemberized.com/images/Epistrophe_grossulariae_female.jpg

Must be a fairly close relative I would have thought. Pantscat (talk) 02:19, 20 July 2010 (UTC)

If you need more you can upload the image using the a button on the left hand margin - select "upload file" which should take you here http://en.wikipedia.org/wiki/Wikipedia:Upload - you'd need to give it a license since everything uploaded has to be shared. Uploading to commons is recommended http://commons.wikimedia.org/wiki/Commons:Upload , then just paste the link here...87.102.13.208 (talk) 02:41, 20 July 2010 (UTC)

How about this one - found nearish your region and has two yellow bands http://www.cornwalls.co.uk/photos/img2270.htm 87.102.13.208 (talk) 02:53, 20 July 2010 (UTC)

Yep, that's the fella. Looked just like that one and sounds about the right size. So I now have a name - Volucella zonaria or the Hornet Mimic Hoverfly.

Thanks, guys Pantscat (talk) 12:32, 20 July 2010 (UTC)

And we do have an article, with some other images. Deor (talk) 17:59, 20 July 2010 (UTC)

Elements in the human body

How many elements are necessary for human survival? So far I have listed:

• Hydrogen
• Oxygen
• Carbon
• Nitrogen
• Phosphorus
• Iron
• Calcium
• Sodium
• Magnesium
• Sulphur
• Potassium
• Iodine

Please list any more that I haven't named, and tell me if any of theses are not needed for humans. Thanks, --The High Fin Sperm Whale 01:38, 20 July 2010 (UTC)

Technically you need http://www.soils.wisc.edu/~barak/soilscience326/listofel.htm essential plant elements unless you want to starve. or Plant nutrition

See also Dietary mineral which I hope is a complete list.01:44, 20 July 2010 (UTC) —Preceding unsigned comment added by 178.78.65.223 (talk)

We need salt, and that has Chlorine in it. Maybe. I'm not a scientist, though. 142.104.215.130 (talk) 01:45, 20 July 2010 (UTC)

Both sodium and chlorine are in the list in dietary mineral.178.78.65.223 (talk) 01:47, 20 July 2010 (UTC)

You check out Composition of the human body yet? APL (talk) 01:50, 20 July 2010 (UTC)

Or better yet, Dietary mineral, which lists (in a convenient periodic table graphic) all of the elements typically in a human body (in reasonable amounts), and also what the different elements do. Buddy431 (talk) 03:22, 20 July 2010 (UTC)

That actually lists the minerals that people need to eat, which is an important distinction. Ariel. (talk) 13:54, 20 July 2010 (UTC)

I would be surprised if there were any reasonably common elements that we don't need for something. Either as a part of our bodies - or as a part of our food sources or some other subtle thing. Possibly the nobel gasses (helium, neon, etc) aren't necessary - but I wouldn't even want to bet on that. SteveBaker (talk) 04:26, 20 July 2010 (UTC)

Acording to the linked article: "An exception is aluminium, which is the third most common element in the Earth's crust (after oxygen and silicon), but seems to serve no function in living cells. Rather, it is harmful in large amounts." Rmhermen (talk) 04:46, 20 July 2010 (UTC)

Despite being linked to 3 (now 4) times the article Dietary mineral seems strangely inaccessible to some - to simplify I've copied the relevant data below:77.86.76.212 (talk) 20:35, 20 July 2010 (UTC)

H																		He
Li	Be												B	C	N	O	F	Ne
Na	Mg												Al	Si	P	S	Cl	Ar
K	Ca	Sc		Ti	V	Cr	Mn	Fe	Co	Ni	Cu	Zn	Ga	Ge	As	Se	Br	Kr
Rb	Sr	Y		Zr	Nb	Mo	Tc	Ru	Rh	Pd	Ag	Cd	In	Sn	Sb	Te	I	Xe
Cs	Ba	La	*	Hf	Ta	W	Re	Os	Ir	Pt	Au	Hg	Tl	Pb	Bi	Po	At	Rn
Fr	Ra	Ac	**	Rf	Db	Sg	Bh	Hs	Mt	Ds	Rg
			*	Ce	Pr	Nd	Pm	Sm	Eu	Gd	Tb	Dy	Ho	Er	Tm	Yb	Lu
			**	Th	Pa	U	Np	Pu	Am	Cm	Bk	Cf	Es	Fm	Md	No	Lr

The four organic basic elements

Quantity elements

Essential trace elements

Pervasive but no identified biological function in humans

Lemon Juice - alkaline or acidic?

I've heard that although lemon juice in its raw form (squeezed from the lemon) is acidic, once it enters your stomach it somehow becomes alkaline and has an alkaline affect on the pH of your stomach - is that true? And if so, by what process is the citric acid converted into an alkaline substance? —Preceding unsigned comment added by 118.139.32.234 (talk) 01:59, 20 July 2010 (UTC)

Once cells metabolise the citric acid to carbon dioxide and water, it will leave behind the alkaline components in the juice such as potassium ions. The carbon dioxide will equilibriate to bicarbonate ion in the body, which is mildly alkaline. The biochemistry that handles this is the citric acid cycle. Graeme Bartlett (talk) 02:40, 20 July 2010 (UTC)

... and, of course, even the most acidic lemon juice has a pH of about 2 which is similar to the average pH in the stomach. The pH of Gastric acid at the point of secretion can be as strong as 0.8. Dbfirs 08:23, 20 July 2010 (UTC)

Don't forget that the carbon dioxide equilibriates to bicarbonate and protons. --Chemicalinterest (talk) 10:49, 20 July 2010 (UTC)

I'm pretty sure that CO2 dissolved in blood is acidic. John Riemann Soong (talk) 18:57, 20 July 2010 (UTC)

It is. Protons are much more acidic than bicarbonate is basic. --Chemicalinterest (talk) 19:36, 20 July 2010 (UTC)

But that's only once the acid has diffused out of the stomach, so it won't have any effect on the pH of the stomach. Even before that lemon juice isnt really going to change to the pH of your stomach - pH is measured on a logarithmic scale - this means that to change the pH of the stomach from 1 > 2 would require diluting the acid with 10 times as much water. So no, lemon juice can't change the pH of your stomach. Smartse (talk) 00:08, 21 July 2010 (UTC)

But most of the CO2 is breathed out losing that acid producing carbonic acid and leaving behind the potassium bicarbonate. But this is not in the stomach, it is in the body cells. That CO2 goes on to acidify the oceans instead. Graeme Bartlett (talk) 06:11, 21 July 2010 (UTC)

Does not... The plants take it and produce O₂, completing the cycle. If the CO₂ is breathed out, then the bicarbonate and carbonic acid are released because the equilibrium shifted. Why it is alkaline is because the K⁺ ion is basic in comparison to gastric acid with its H⁺ ion. --Chemicalinterest (talk) 14:44, 21 July 2010 (UTC)

Under normal conditions, the K+ ion is neither basic or acidic. It is ever so weakly Lewis acidic, even more than Na+. If you take the potassium salt of fluoroantimonic acid suddenly the normally stabilised K+ ion will become a strong Lewis acid indeed.

Also, breathing CO2 out does not leave behind bicarbonate. CO2 is breathed out by protonating bicarbonate. Blood near the lungs is more acidic than blood in the systemic capillaries. This does two things: diioxygen's affinity for haemoglobin increases, while bicarbonate is protonated and released from transporter complexes. All that is left behind is the water that the CO2 reacted with in the first place, at the actual site of metabolism. Bear in mind, blood pH is buffered by bicarbonate. John Riemann Soong (talk) 15:49, 21 July 2010 (UTC)

age

Can there is a medicine aur ane thing bye which the growth of child increse?? example like in movies now days actors age of about 15-17 look much younger than mormal .. —Preceding unsigned comment added by 115.248.45.13 (talk) 06:08, 20 July 2010 (UTC)

Nothing that would be remotely legal to administer to a child to alter their appearance for acting purposes. Film directors cast actors who look the right age for the part, and some people look older and some younger than the average, so an actor who looks young will be preferred (as they will have more acting experience than someone really that young), and such actors will specialise in such roles. Make-up also contributes, of course.

There are treatments to help the growth of people whose growth is less than it should be because of some medical problems, of which there are several. See, for example, Growth hormone. 87.81.230.195 (talk) 07:10, 20 July 2010 (UTC)

I'm confused why you want medicine to increase the growth of a child. Do you think actors age 15-17 need to be given it to look older? As 87 has said, this seems strange, actors who look young are generally preferred so there's no reason why anyone, either the director/s or the actor/s are going to want to take it. If you think actors age 15-17 have been given medicine to make them look young then logically it's medicine which reduces the growth of a child not increases. Nil Einne (talk) 20:48, 20 July 2010 (UTC)

Acting a part is more than just "being" the part. Early silent movie actress Lillian Gish said that director D.W. Griffith told his young adult actresses to hop and skip and jump around to look more like children. Today they just look like the are hyperactive or ate too much speed or sugar. Actors or comics today still use certain mannerisms when playing younger characters. Edison (talk) 19:58, 21 July 2010 (UTC)

Strange worms - Could anyone tell me what these are?

I've seen quite a few of these worms in covered areas (mostly on stored plush fabrics, boxes, bags, etc.) and have always wondered what they are. I most often find dead ones, but I've also noted a few live ones in some old carpets.

Most of the worms I've come across are about .3-.5cm in length, but, in the most extreme cases, I've seen them go up to 1cm.

Here is a picture of the one that I normally see: http://eepromeagle.dyndns.org/unknownbug/IMGP4319.JPG And a larger one: http://eepromeagle.dyndns.org/unknownbug/IMGP4322.JPG

EEPROM Eagle (talk) 06:35, 20 July 2010 (UTC)

It might help someone to make an identification if you tell us approximately where in the World you are, since insect species are very numerous and different continents and climates have markedly different populations. These look as if they may be the caterpillars of some kind of moth (so not a "worm" except in a very non-scientific sense), but beyond that we need geographical clues to narrow down the possibilities. 87.81.230.195 (talk) 06:57, 20 July 2010 (UTC)

Beetle larvae (not worms) - possibly Dermestidae Shyamal (talk) 07:01, 20 July 2010 (UTC)

Yes, it looks like the cast off shell of the larvae of the larder beetle. I get them in my flat, I wish I didn't. 213.122.7.114 (talk) 09:29, 20 July 2010 (UTC)

Heh, yes. I said worm just to say to what they kind of looked like to me. I was pretty sure they were larvae. To answer the above question, I'm in Massachusetts. Thanks for the information! EEPROM Eagle (talk) 22:01, 20 July 2010 (UTC)

would it be possible to duplicate, mechanically, the action of a modern CPU (obviously at up to minutes per clock cycle).

A modern latest CPU like the Intel i7 can have 8 megabytes of cache, which is 67 million bits (or a square array of 8185 by 8185 bit-representing elements). Then it has 731M transistors. That's a square array of 27038 x 27038 elements. However, they are constrained to work on a very, very small scale, as they must operate at close to or somewhat over 3 GHz. Now, in 3 GHz, light only travels 10 centimeters, so obviously the overall size can't be huge.

Our mechanical design has no such constraints: it's fine to be many feet by many feet (by many feet, since you can use vertical space more easily when assembling wooden parts by hand).

If "speed is no object", and it is fine to take as long as you want to go through one clock cycle, is it feasible to duplicate the action of this CPU mechanically, such that if you mechanically set the input of the "wooden processor-equivalent" it will churn and eventually give the same output as the actual Core i7?

Secondly, if you imagine this mechanical processor, what do you think the fastest spinning gear is, and how fast does it have to spin to complete one "clock cycle" in wood in, say, 30 seconds? Then, if we take that RPM and do simple multiplication, what RPM is the Intel i7 at 3 GHz equivalent to in wood and metal rpms? Say, 18 trillion RPM? Thanks. 84.153.247.76 (talk) 11:00, 20 July 2010 (UTC)

Charles Babbage nearly did something similar with his Difference engine and Analytical engine in Victorian times. 92.15.7.17 (talk) 11:18, 20 July 2010 (UTC)

(Well, he did make working 'designs' - but none of his designs were actually completed until the British Science Museum recreated one of them a couple of years ago). SteveBaker (talk) 11:54, 20 July 2010 (UTC)

Please read the third word in the sentence above. Incomplete parts of the machines were made under his supervision. 92.15.7.17 (talk) 12:49, 20 July 2010 (UTC)

I did read what you wrote - but "nearly" is a pretty weak explanation and I think it needs amplification. What Babbage achieved is widely misunderstood. He produced a complete design for the difference engine (which is NOT a computer - but more like a fancy printing calculator) and he made many partial designs for the analytical engine (which would have been, recognizably a "computer" - but for narrow technical reasons is not properly "Turing complete"). During his lifetime, only about 10% of the difference engine design was ever actually constructed - but it was fully completed recently by the science museum in London - and with a simple correction to the original drawings (where the orientation of some part was reversed) it was made fully functional. But the design for the analytical engine - which would have been "the world's first computer" is not complete enough to be constructed. What I think is most impressive is that the first ever computer program was written for the analytical engine (although there is some controversy about whether Babbage or Ada Lovelace actually wrote it). SteveBaker (talk) 00:18, 21 July 2010 (UTC)

You certainly wouldn't "duplicate" the way a modern CPU actually works. But it's certainly possible to build mechanical computers - and many people have done that. (I actually own one - you can see a photo I took of it in our Digi-Comp_I article). There is a mathematical law called the Church-Turing thesis that says that any computer that passes a certain minimal level of capability (a "Turing-complete" computer) is - in principle - capable of the same things as any other Turing-complete computer...providing you have enough time and memory to complete the calculation. So we could build a relatively small, fairly simple computer (like this one - which is made from Lego) and it could simulate an Intel i7 - eventually - and given enough memory...which is just a matter of being patient enough, and owning enough Lego! SteveBaker (talk) 11:54, 20 July 2010 (UTC)

you say "you certainly wouldn't "duplicate" the way a modern CPU actually works" (my emphasis). Why not? You say that it's not possible if the input, off the bus, is fed via braille, to get the same output as the i7, having taken the same intermediae processing steps??? Well why not... aren't all the little transistors basically just doing boolean logic? THere's nothing "magic" or analogue about an i7 that you couldn't do with another, logically equivalent, design: after all, before they ever built one, Intel surely modeled the i7 in software... 84.153.247.76 (talk) 12:30, 20 July 2010 (UTC)

I Imagine Steve's point is that although you could (in theory) duplicate the 731 million transistors of the i7 in a mechanical computer, that wouldn't be a rational way to design a mechanical computer. With a mechanical computer, costs of fabricating and maintaining individual components become significant, and factors such as friction increase with the number of components, so it would be better to go for a design that minimised the number of components, even though it did less in parallel and took more clock cycles to process each instruction than a mechanical i7 clone would. Gandalf61 (talk) 13:08, 20 July 2010 (UTC)

Well, I don't know what you guys think I might want with a mechanical computer, but it's nothing at all. I have absolutely no need for a mechanical computer, I am only interested in the Intel i7, and the only reason for asking about a mechanical computer is whether one could, theoretically, at great cost and impracticality, be created that was logically equivalent to it. If so, how long could we estimate it would take to "render" (or "virtualize" or in some way be equivalent to) one clock cycle on the i7? Going from that, how many RPM would it go at to, say do an i7 clock cycle in 30 seconds? Thanks. 84.153.247.76 (talk) 13:39, 20 July 2010 (UTC)

A mechanism that was "logically equivalent" to the i7 (i.e. gives equivalent outputs when presented with equivalent inputs) would not need to replicate the i7's internal design. So - do you really mean "logically equivalent", or do you mean a mechanical copy of the i7, where you could identify each i7 transistor with a specific mechanical widget ? Gandalf61 (talk) 14:10, 20 July 2010 (UTC)

There's no easy way to know this without actually spending the rest of our lives designing the machine. The mechanisms of an electronic chip (Transistors, etc.) do not exactly translate into mechanical constructions. Babbage's proposed computers used gears, but he was designing a new system from the ground up, taking full advantage of the parts he had available. It's not at all clear to me that an exact duplicate of the i7 could be built that way. In fact, pneumatic systems would be a closer analog to electronic circuits than gears would.

This is an important point because how that mechanism is crafted will be a large portion of what determines the speed that your machine runs at.

Not only that, mechanical linkages have problems that electronic ones don't. (Biggest issue : friction adding up over all those gears.) So something will have to change regardless for that to be fixed. APL (talk) 14:54, 20 July 2010 (UTC)

To be clear, there is no theoretical reason you couldn't build a mechanical copy of an i7. In practice, mechanical computers tend to work differently though because binary logic is not easily modeled with gears. (Think: What is the mechanical analog of a transistor, or an AND gate, etc.? Such things could be created, but the individual elements would tend to be more complex than is practical.) You'd also have to design so that the total friction in the system was manageable and the amount of force between individual gears was never so great that the gears would break. I would speculate that if you tried to follow i7 logic directly using wooden gears (as you suggest), then you'd probably fail both of these constraints. In practice, it would be much more reasonable to think about designing a system that takes i7 inputs and gives back i7 outputs, but was internally redesigned to operate mechanically. As Steve mentions, there are a wide variety of ways that could be accomplished. However, questions like "how fast is the fastest gear" and "what RPM" are largely pointless without considering a specific design and such values could be changed by orders of magnitude based on engineering choices that have no impact on overall performance. So I don't think there is any useful answer to those questions without considering a specific model. You could look at some of the historical mechanical calculators and try to figure out how much longer they take to operate than a modern CPU given the same operation, and that is perhaps the most useful comparison I could suggest. Dragons flight (talk) 13:48, 20 July 2010 (UTC)

It might be theoretically possible, but it would be an engineering wonder, and easily the most complicated machine ever constructed. Googlemeister (talk) 13:52, 20 July 2010 (UTC)

If the chip has 700 million transistors, and each one requires a few moving parts, you're talking billions of moving parts (compare to the Space Shuttle's 2-3 million mostly non-moving parts). I think it would be a statistical impossibility that they would all be functioning at the same time, even for a moment. --Sean 15:09, 20 July 2010 (UTC)

Let's think about this the way a computer-designer actually thinks about this. Every few years, a new VLSI process is introduced. Sometimes, they just make this year's process smaller; but sometimes they fundamentally change something - like the order that individual chemical layers get deposited. When those changes happen, the chip designers don't sit around thinking "darn, now we have to translate the i7 core, device by device, 731 million times." This is critical - a modern processor has 731 million pieces (or whatever) - so to build it, you must have another machine to macro-ize construction operations 731 million times. Instead, they just "recompile" their design for a new process, and then spend some effort "tuning" for performance. Here's how it works. First, you ask yourself how to build one transistor in the new process. And you come up with a transistor macro; and then you consider how you need to interconnect two transistors; and then after some time, you end up with a set of macros to represent a lookup table, a bit of a register, a multi-bit vector, and so on. Now, the i7 core is described somewhere deep in the vaults of Intel's research and development lab; and it is described in a language like verilog or VHDL (or more likely, some bizarre combination of both that gets woven together by incomprehensible perl scripts and passed down by an elite cabal of engineers and computer architects). Then it gets run through a different machine that maps each high-level interconnection description into a series of device-layer connections - called place and route or VLSI synthesis. So - let's get back to the horrifyingly huge wooden contraption. How do you build one transistor in mahogany and teak? Well, let's think about what a transistor actually is - and remember, we have the luxury that we really only care about digital logic, so we really just want a high-gain inverting transistor. This is a simple device that takes one input and produces an output at the opposite value of the input. So, let's think of your wooden device as some kind of lever or piston or something; it has to take a high- or low- input (maybe it gets poked by the position of another lever or piston, and then pokes the next lever with the opposite polarity). Of course friction has been brought up many times; transistors deal with this too, it is called the "static current draw" or the quiescent current. Even the best amplifiers waste some energy. If you look at a circuit schematic for a transistor, you will always see both a signal input and a power input. Thermodynamics requires that power be consumed. The difference is that one transistor is ridiculously close to the theoretically most-efficient system; whereas your wood/pneumatic piston is not. But oh well; nobody really worries about energy problems in modern computers anyway. Let's say you just attach an air-compressor to every piston, and power all the pneumatics using fossil fuels, (keeping the analogy to modern electronic computers). So we now have one wooden transistor, and we need to interconnect it to others to form basic building-block devices, like truth tables, clocks, registers, and arithmetic units. Then, we design the macros for each of these, and provide the list of available macros (and their description in terms of boolean logic) to the VHDL synthesizer software. That will crunch out our design for about 24 hours, and finally spit out a list of all 731 million wooden/pneumatic devices - including where to place them, and what to connect them. Now, you could require those placements to be exactly in the same location and connection pattern as an i7 core - but that would be stupidly inefficient. The i7 core has been optimized for the speed of propagation of 32-nm silicon CMOS. The number of cache-lines, and the number of bits per cache-line, for example, are directly determined based on real latency to main memory in the particular technology it was designed for. Much of the complexity comes about from these specialized hardware and process tuning. But, for the sake of exploration, let's say that your wooden memory timings are exactly scalable to those of a modern DRAM. Now you just need a way to fabricate 731 million pneumatic elements. In CMOS, this is easy - we call it photolithography - but you can probably find some way to easily obtain all the parts, lay them out, and operate on them in bulk. Don't forget to build redundancy into your design - because in 731 million wooden pistons, termites will bore out the innards in 1 in a million wooden pistons. So, at the wooden CPU factory, you need to be able to build everything - test everything - and then automatically reroute to a redundant piston any time a sub-unit doesn't work. Alternately, you can spend your whole life trying to de-termite your fabrication-process to obtain a 100% yield; but that is less efficient. So, in summary, yes it could be done. If the explanation above has served any purpose, it is to convince you why microfabrication and silicon are the only realistic ways to do processor design using reasonable quantities of time, resources, and engineering effort. VLSI is amazing stuff - it's got room for innovation - but it really is centuries ahead of any other mechanical manufacturing process. Nimur (talk) 16:02, 20 July 2010 (UTC)

Wait, I notice a couple of times the question-asker uses the phrase "logically equivilant". If all you want to do is mimick the inputs/outputs of the i7 then Steve's method above would work just fine. Any Turing compatible CPU can be made to mimic the inputs/outputs of any other Turing compatible CPU.

All you'd need is an additional mechanism to map certain parts of the mechanical Turing machine's 'memory' to the input/output pins, and run the 'memory' through that device at the end of every simulated cycle. APL (talk) 20:47, 20 July 2010 (UTC)

My first answer is still the right one. Let's refer to the actual wording of the question: "would it be possible to duplicate, mechanically, the action of a modern CPU". The "action" of a modern CPU can been seen at various levels. At the level of software - the Church Turing thesis says that, yes, any Turing machine could be programmed to accept a computer program for an Intel chip and produce the same numerical answers. At the level of voltages on pins of the ship, obviously a purely mechanical computer fails because it can't accept electrical signals without some kind of tiny low current/voltage motor.

The fuzzy answer of whether you could simulate an Intel chip by somehow simulating each of it's internal logic gates out of mechanical systems is tough...I don't think that's relevant to the original question though - you don't have to do a 1:1 simulation to fulfil the requirements to simulate that chip perfectly. For example - when the engineers at Intel are designing a new CPU, they don't draw out this vast circuit - then manufacture one - then see if it works. Turning a paper design into a chip costs millions of dollars - you don't do that until you have a known-working design. So they start out by simulating the new design - first at the level of registers and such - then at the level of gates, then at the level of transistors. Those simulations are 1:1 identical in function to that of the real chip - so that when they manufacture the chip, it should work the first time. Since that exact simulation of the chip is a computer program - the Church-Turing thesis says that we can precisely simulate the chip (right down to measuring the voltages at the inputs and outputs of transistors) on any sufficiently large Turing machine. We could even have our mechanical Turing machine have a mechanical display of some sort (imagine lots of multicolored cubes that rotate on rods to show different colored faces or something) - and display the interior logic of the Intel chip. So yeah - we can do a perfect simulation on any Turing machine.

Another branch of this discussion has been talking about the size and speed of a mechanical computer. But let's think out of the box here. A mechanical computer could be vastly faster than the fastest hardware we have today. Calculations on the possible speeds of a nanotechnological computer in which the mechanical parts were (let's say) 100 atom carbon chain push-rods, levers and cams suggest that it would be vastly faster than the fastest machine we could build out of electronics. The fastest transistors we have can switch at perhaps 100GHz - but moving 10 atoms a distance of just a few times the size of an atom can happen WAY faster than that. There is much debate about whether such machines are possible - but if they are, they will almost certainly be teeny-tiny mechanical turing machines.

SteveBaker (talk) 23:59, 20 July 2010 (UTC)

Did you mean to type "simulating each of ITS internal logic gates" ? Cuddlyable3 (talk) 15:44, 21 July 2010 (UTC)

It's an interesting idea and I certainly hope someone looks into it. But I'm not going to invest any money I can't afford to lose. Intuitively it seems to me the big problem is going to be reliability. How many of your enormous number of rods and cams and levers are going to get stretched, or squashed, or broken, in the course of a computation?

By the way I also think that's why your big version made out of wood probably won't work. Babbage never tried to represent a billion bits. At some point, changing one bit will require changing so many others, that the force you have to exert at the input level will simply break something along the way. But as I say, that's intuition; if you have a design that gets around the problem it would be interesting to see it. --Trovatore (talk) 00:32, 21 July 2010 (UTC)

That's why you need a wooden buffer amplifier - you must expend power in order to isolate each stage from the others. But this means that any individual wooden lever does not have to push a billion others - it just has to drive its own output. The friction of later stages is totally isolated. Each stage consumes power from an external power-supply, so that the signal input does not have to provide power input as well. Nimur (talk) 00:58, 21 July 2010 (UTC)

Hmm, might work. I'm still skeptical that it can be done reliably enough to simulate a modern computer. --Trovatore (talk) 01:15, 21 July 2010 (UTC)

Babbage had worked out some clever tricks to avoid the problem of changing one digit and thereby causing a whole bunch of gears to turn at once. The obvious worst case would be when if you added 0000000001 to 8999999999 - resulting in 9000000000 and the force of one tiny gear wheel changing would have to turn (in that case) 10 other wheels. Babbages difference engine used something like 30 decimal digits of precision (VASTLY more precise than a modern computer in full double-precision mode!) - so in his case, turning one wheel might cause 30 others to have to turn - which would have caused serious mechanical issues. Those kinds of problems had beset previous efforts at building mechanical calculators - resulting in them being restricted to low precision calculations and making them highly temperamental. He came up with a rather clever way to fix that problem so that the amount of force required to do a worst-case 'carry' was no different than for a simple non-carry addition. However, the way he did it in the Difference Engine was rather slow - basically, the turning of a result wheel from 9 to 0 tripped a little "carry" lever and didn't propagate the carry on to the next digit - then this little spiral gizmo went down the column of levers one at a time rotating the wheel by one unit whenever the previous wheel's carry lever was sticking out - resetting the carry lever and possibly tripping the next carry lever down the stack. If you watch the machine in the London Science Museum doing this while it's adding random numbers together, it's kinda hypnotic! The machine performs about one addition per second - so these things happen rather quickly! For his later "Analytical engine", he came up with an even more devious mechanism that avoided the mechanical force issues and didn't consume significant extra time by doing a kind of 'carry prediction' trick. The interesting thing about that was that it made the adding mechanism so freaking complicated that he gave up the idea of every stack of gears that represent a number having the "ADD" capability. Instead it would copy the numbers from a "memory register" into the "mill" where the addition would take place and the result copied back into the register. This is eerily similar to a modern CPU where numbers are stored in dumb memory and all of the math happens in one central unit rather than having full arithmetic capability at every single byte of RAM! SteveBaker (talk) 21:36, 21 July 2010 (UTC)

You could have small wooden balls to simulate electrons. It would then be the size of a skyscraper. 92.28.255.176 (talk) 08:59, 21 July 2010 (UTC)

For that you must have very small balls. Making DRAMs would be difficult. Cuddlyable3 (talk) 15:48, 21 July 2010 (UTC)

Skyscraper? No, it would be a few miles on a side at least. The i7 has a minimum feature size of 32nm. If you make the reasonable assumption that your wooden device will have a minimum feature size of half an inch, you wind up with a device roughly 400,000 times the original. APL (talk) 15:56, 21 July 2010 (UTC)

On early vacuum tube computers, the mean time between failure of the few thousand tubes (valves) was a practical limit on complexity (as was power consumption). What is the MTBF for a high speed wooden gear and pinion, or lever action? How many times can it repetitively click and clack at high speed before it wears or warps? There may be a practical limit on what fraction of a modern computer chip could be physically modelled in wood, and how many clock cycles it could complete before repairs/replacements/adjustments were necessary. As complicated wooden mechanism, various early automata, the Jacquard loom, the [Tracker action| tracker organ] and other musical instruments, and the wooden clock come to mind, so somewhat complex wooden gadgets can function thousands of times, or tens of thousands of times, with levers moving levers or gears turning pinions. A chip goes at many millions of operations per second and many seconds are required for some calculations or programs. It would be instructive to make a mechanical analog of an amplifier, an invertor, basic logic gates, or a flipflop. But I would not expect to see a wooden computer which would click and clack and invert a large matrix, say. Edison (talk) 19:40, 21 July 2010 (UTC)

A Pneumatic Transistor

Here's how to make a pneumatic (P) transistor (reverse the spring and gate for the N equivalent). You'll also need pneumatic "wires" (copper pipe will do the trick and give the whole thing an attractive old-world charm) and a few other simple device equivalents such as resistors (constricted pipes) etc. Once you've got the handful of fundamental devices, you can build and model a standard cell library, and enter it into your synthesis libraries. Run through the standard EDA flow as described above to get a fully mechanical pneumatic CPU.

The simulation part of the EDA flow will give you an estimate of the highest reliable clocking frequency as well as static and dynamic current (pressure) consumption. The MTBF of a pneumatic transistor is likely much lower than for a semiconductor one which will shorten your CPU's lifetime accordingly (perhaps to less than the time required for a single operation). David Carron (talk) 14:39, 22 July 2010 (UTC)

The first computers built by Konrad Zuse were mechanical. They were rather specialized as they worked in floating point but the design could be changed pretty easily to do what modern computers do except you'd need a lot more memory and they'd be very slow. The only electrical bits were the motor and tape reader. The mechanical memory was more compact than anything else for another twenty years. Dmcq (talk) 14:35, 24 July 2010 (UTC)

how did the Greeks discover the earth was round?

How did the Greeks discover the Earth was round? 84.153.247.76 (talk) 13:36, 20 July 2010 (UTC)

Well, the first guy who estimated the diameter was named Eratosthenes who did this around 250BC or so, but I think they knew it was round before that. Googlemeister (talk) 13:47, 20 July 2010 (UTC)

Because it looks round. My history of science tutor believes it was pretty obvious to any seafaring nation that the horizon was curved, and it was not a 'discovery' at all. The idea that it was flat wasn't one ever held by the Greeks, or by western civilisation at all, but found only in civilisations constrained to land (egypt, babylon, the western populace prior to civilization per se), where a smooth horizon is hard to come by, and so because the surface of a sphere locally approximates to euclidean, the ground appears flat. —Preceding unsigned comment added by 129.67.116.122 (talk) 13:58, 20 July 2010 (UTC)

I doubt the Greeks had Wikipedia back then, but we do now, and it has a page specifically about the issue: spherical earth. DMacks (talk) 14:15, 20 July 2010 (UTC)

Perhaps by considering the fact that the shadow at noon is not vertical in places to the north and south. But another explaination for that could be a flat earth and a nearby sun. 92.29.123.248 (talk) 13:00, 21 July 2010 (UTC)

Late response, but an early proof was the curved shadow on a lunar eclipse. Awickert (talk) 02:53, 24 July 2010 (UTC)

Website which shows raincloud movement for the next few hours (in UK)?

In the Netherlands, there is a well known website that lets you see where rainclouds will be moving over the next few hours. It's pretty accurate, 'cause it's only predicting the next few hours based on radar data of cloud movements. Do we have such a website for the UK? ----Seans Potato Business 13:40, 20 July 2010 (UTC)

Have you tried the Met Office web site ? This app shows cloud cover forecasts in map form, and this one shows rainfall forecasts. Both show snapshots every 3 hours for next 24 hours, then two snapshots per day for following 3 days. Gandalf61 (talk) 13:50, 20 July 2010 (UTC)

You might also try http://news.bbc.co.uk/weather/ - click on the more uk weather link to access various interactive time-lapse displays of recent actual and future forecast rain etc, though unfortunately the scope and options of these seem to have been restricted quite recently. 87.81.230.195 (talk) 18:24, 20 July 2010 (UTC)

induction

would it be possible to induce a polar or overall charge on an incoming bullet usign static electricity or someother means and then using a high power magnetic field to deflect it away?--91.103.185.230 (talk) 14:01, 20 July 2010 (UTC)

In theory, yes. In practice, no. — Lomn 15:07, 20 July 2010 (UTC)

I agree. To elaborate, start by reading muzzle energy. A modern rifle round carries thousands of joules of kinetic energy. To "stop" this bullet, you need to dissipate that energy; to deflect it, you need to impart enough momentum to meaningfully change its trajectory. If you consider the energy involved, we have no machinery that can practically or safely deliver that amount of energy in a focused way to induce current or magnetic force on a bullet. An easier way to deflect such a bullet would be to detonate a large explosive in its trajectory. Alternately, you can dissipate bullet energy effectively by hiding behind a sandbag barrier. Because the bullet impacts inelastically with sand, sandbags are a better defense against rifle and machine-gun rounds than steel-plate armor. When you consider alternatives, you have to keep thinking about how much energy you need to dissipate, and/or how much momentum you need to null out or deflect. Nimur (talk) 16:26, 20 July 2010 (UTC)

It depends on what is meant by 'deflect'. I can conceive of a laboratory-demo-type experiment in which a bullet is fired, is given a static electric charge, and then passes through an intense magnetic or electric field and suffers a small deflection. In that situation, a shot aimed at a relatively distant target might be deflected by a few inches — enough to miss the person at whom it was aimed. Mind you, I can't think of a practical use for this sort of scheme; a small-angle deflection is only useful if it can be applied far, far from the target. TenOfAllTrades(talk) 18:26, 20 July 2010 (UTC)

(ec) If the bullet is linear dielectric it would be attracted to a point charge with a force that scales r^-4. This would make it practically hard. —Preceding unsigned comment added by 92.8.29.89 (talk) 18:31, 20 July 2010 (UTC)

Suoermagnetman recommends using magnets to break the motion of metallic bullets (and so do I) ( http://www.youtube.com/watch?v=iABmUEH5s0k&NR=1&feature=fvwp )(see also Eddy current brake). The motion is close to critical damping, see Damping. 77.86.76.212 (talk) 19:33, 20 July 2010 (UTC)

You could do a calculation to determine how much momentum you need to impart to yield a 12-inch deflection on a round that is traveling at 900 m/s. If you begin applying an acceleration orthogonal to the round as it exist the muzzle, you have less than 1/3 of a second before it reaches its target (at maximum range). How much force is required to accelerate a 5 gram object so that it gets a 0.25 meter displacement in 0.3 seconds? This is elementary physics: I'm getting a force appoximately equal to 1 newton. Does anyone know a way to focus an electromagnetic or electrostatic field at a distance of 300 meters and impart a 1 newton force on a lead/copper bullet? As far as I know, there is no such technology capable of that. Take a look at our article on magnetic force. Electrostatic acceleration is just implausible; magnetic induction is equally unlikely (you must first produce eddy currents, and then subject the magnetized object to an additional external field to apply a force to it). The physics just is not on your side with this one. Even if you fired the round into a specially-constructed laboratory setup, some kind of tube or tunnel surrounded by magnetic coils... that's still a lot of force to produce strictly by induced magnetic interaction. The idea is fun to toy with in theory, but when you start looking quantitatively at feasibility, it is no surprise that today's best defense against bullets is still low-tech sandbags. The laws that govern energy and momentum are unchanged by our ability to harness electromagnetism. Nimur (talk) 20:13, 20 July 2010 (UTC)

Your life does get easier if you assume a muzzle velocity of 300 m/s (typical of a handgun) rather than 900 m/s (a supersonic rifle round). If we assume a 300-meter range, then the constant force over that distance to achieve a 0.25 meter deflection of a 5-gram object is just 0.0025 newton. If we apply a constant one newton force, then we only need to apply it over the first 0.5 meters (roughly) to generate a 0.25-meter displacement 300 meters downrange. TenOfAllTrades(talk) 21:00, 20 July 2010 (UTC)

TOAT, handguns at 300 yards? You must have watched different Westerns than I did! Physchim62 (talk) 23:03, 20 July 2010 (UTC)

I didn't say it was a good idea, just that a 'best-case' scenario for some sort of deflection device. I'll admit that expecting a handgun to be accurage within a foot at that distance is pretty unlikely. TenOfAllTrades(talk) 01:47, 21 July 2010 (UTC)

Unlikely? Try miraculous. Googlemeister (talk) 14:17, 21 July 2010 (UTC)

Some electromagnetic and electrostatic effects on projectiles could in principle make the difference between hitting or missing a target, but might require expensive equipment. A deflection applied near the muzzle, or over a large portion of the path, will have a larger deflection than one applied only near the target. How hard does the wind push on a bullet? Marksmen spend enough time adjusting for windage, which would otherwise deflect the bullet at distances rifles commonly shoot. A rubber comb, with a small static charge can deflect a falling stream of uncharged water dramatically. Static electricity should be able to deflect a bullet some if the bullet were charged. Electric guns firing ferrous projectiles illustrate how powerfully induction can accelerate a projectile longitudinally. A copper disc dropped between the poles of a strong magnet seems to stop in its descent before descending slowly due to induced currents. A charged round projectile like a BB would retain charge better than a bullet shaped one, and would follow a curving trajectory when it passed through a strong transverse magnetic field. As a thought experiment, if a BB gun were charged from a Van de Graaf to 100kv, then fired at a target, the aim could be made such that the drop due to gravity placed it in the bulls eye at some distance, like 10 meters. Now consider if the first meter it travels between vertical plates either side of the path, charged to plus and minus varying voltages, spaced just far enough apart they did not arc. How many cm of deflection could be achieved? It is like a mechanical analog of an oscilloscope. An uncharged steel bb could be similarly deflected by passing through a magnetic field. Edison (talk) 19:24, 21 July 2010 (UTC)

how do you tell if a plant cell has apoptosed?

I can see stuff through the cell wall (DIC microscopy) and it looks like a mess. However, it's not quite like the arrangement when animal cells die (they give off apoptosis "bubbles"), so I'm not sure. They are also onion cells, so I'm not sure if those are simply lots of storage vesicles clumped together. John Riemann Soong (talk) 16:31, 20 July 2010 (UTC)

Ok I used a 700 nm filter (for gold nanoparticles lol) and I can see distinct intact nuclei...wow I didn't think that reducing the light levels by 20 times would help me see inside the cell. Because I don't fully understand DIC -- I want to ask, is it because cellulose doesn't diffract 700 nm light very well? (I get lost when it comes to diffraction and polarisation and not simply absorption/emission). John Riemann Soong (talk) 16:45, 20 July 2010 (UTC)

odorless mineral spirits

How strong (harsh, corrosive) is odorless mineral spirits compared with 70% rubbing alcohol? I would like to use the spirits to clean off kitchen countertops, and formica kitchen table. —Preceding unsigned comment added by Tomjohnson357 (talk • contribs) 17:41, 20 July 2010 (UTC)

The article White spirit has some information about toxicity. It's not a good idea to wipe down counter-tops with white-spirits unless you are going to thoroughly wash them again straight afterwards (presumably you're asking because Formica doesn't work well with household Bleach]?). ny156uk (talk) 18:15, 20 July 2010 (UTC)

Is there something particularly horrible you're trying to clean off? White vinegar does a decent job of disinfecting surfaces and is obviously okay for food handling areas. I'd be leery of using anything as toxic as paint thinner on something I was going to leave food on. Matt Deres (talk) 18:53, 20 July 2010 (UTC)

Do any of these solvents contain OH groups? John Riemann Soong (talk) 18:54, 20 July 2010 (UTC)

will odorless mineral spirits damage the surfaces how strong is it —Preceding unsigned comment added by Tomjohnson357 (talk • contribs) 19:52, 20 July 2010 (UTC)

The strength of household white vinegar will vary from brand to brand, but you're typically going to find stuff that is around 5% to 8% acetic acid. As with any other acid, I would hesitate to use it on marble (which is essentially calcium carbonate), as it would eventually etch the surface. I use vinegar as my standard cleanser on glass, porcelain, chrome fixtures, and my countertops. I also use it to disinfect my plastic and wooden cutting boards (see here). I know a lot of people prefer diluted bleach for disinfecting surfaces, but I prefer the smell of vinegar and I like the convenience of not having to worry about getting the dilution correct - you just use it straight up. After applying vinegar, you need to wipe it off afterwards as it can leave a residue if you let it puddle. Matt Deres (talk) 23:46, 20 July 2010 (UTC)

Note: This follow-up question read "will it damage ..:" when Matt Deres responded. It was later changed to "will odorless mineral spirits damage ..."

Few countertops (worksurfaces in the UK) are made of marble in the strict sense, because it's not good to have a polished surface that you can easily etch with spilt lemon juice, vinegar etc. Most of them are some form of igneous rock (although very little of what's sold as granite is genuinely granite), which will probably survive most of the chemicals mentioned. Mikenorton (talk) 12:21, 21 July 2010 (UTC)

note the OP asked about a 'formica' worksurface. . OP - This link http://www.naturalhandyman.com/iip/infcountertop/infcleaningformica.html may be of use. ny156uk (talk) 17:41, 21 July 2010 (UTC)

Anything science still can't do?

Is there anything, after all these years, that science still can't do?

Any number of things; check out science fiction. — Lomn 21:02, 20 July 2010 (UTC)

It apparently can't make everyone know that there is no "end of science". See our Science article. Comet Tuttle (talk) 21:03, 20 July 2010 (UTC)

Well it depends what you mean 'science still can't do' since since isn't something that does stuff. But anyway at a guess at what you mean the answer is a arguably an unlimited number of things. We still haven't mastered nuclear fusion for electricity generation, we still haven't made spaceships capable of 0.1C... You could include things believed impossible like faster then light travel. Nil Einne (talk) 21:05, 20 July 2010 (UTC)

There are certainly things scientists can't do. APL (talk) 21:08, 20 July 2010 (UTC)

Science seems to find it difficult to explain why so many supposedly well educated people, totally willing to use the fruits of its discoveries, reject it publicly in favour of mystical explanations for many perfectly natural phenomena. HiLo48 (talk) 21:12, 20 July 2010 (UTC)

Science find it difficult to explain why so many supposedly well educated people, totally willing to use the fruits of its discoveries, reject it publicly in favor of unscientific random changes for many perfectly natural phenomena. --Chemicalinterest (talk) 21:48, 20 July 2010 (UTC)

Actually, human susceptibility to superstition seems to be reasonably well understood. APL (talk) 22:29, 20 July 2010 (UTC)

Try List of unsolved problems for starters.

Ben (talk) 21:15, 20 July 2010 (UTC)

BTW, thank you for that article / category; it is great. Comet Tuttle (talk) 23:48, 20 July 2010 (UTC)

I was expecting the link to economics' unsolved problems to just be a redirect to economics. :) Matt Deres (talk) 00:00, 21 July 2010 (UTC)

Go back in time. Otherwise, science could do whatever science CAN do at any time. Googlemeister (talk) 19:53, 21 July 2010 (UTC)

It's common to confuse science and technology. Although they're usually pretty closely linked, they're not the same thing. Dear OP, if you were you asking whether there's anything that science still doesn't understand, the answer is yes, and I sincerely hope that it always will be. Science doesn't ever end because most — if not all — any scientific advances let us ask more questions that we wouldn't have known to ask previously. For example, we couldn't begin to address the cure for cancer without first developing cell theory. – ClockworkSoul 22:00, 20 July 2010 (UTC)

Well, if you want a specific example, science can not explain why animals sleep. There is discussion on memory and stuff, but even flies sleep. Ariel. (talk) 22:36, 20 July 2010 (UTC)

Science will never explain why I love my partner, and I prefer it that way ;) everything thing in its right place! Physchim62 (talk) 22:59, 20 July 2010 (UTC)

It actually can probably give a variety of good accounts for that. --Mr.98 (talk) 23:54, 20 July 2010 (UTC)

I think that science actually can explain why animals sleep. To my mind the "synaptic homeostasis" theory of Giulio Tononi is very strong, and I have the impression that it's gradually gaining favor in the community. The basic idea behind the theory is that learning and memory produces increases in neural connection strength that destabilize brains, and it's necessary to have an off-line period in which memory is switched off so that the problem can be fixed. Flies need to sleep because flies have memory -- using mechanisms surprisingly similar to the ones in humans. Looie496 (talk) 04:13, 21 July 2010 (UTC)

The consequence of not sleeping, Fatal familial insomnia, can at least allow science to infer some things, e.g. that the brain has a critical need to rest periodically (as does the heart). ←Baseball Bugs ^{What's up, Doc?} carrots→ 04:29, 21 July 2010 (UTC)

"synaptic homeostasis" doesn't really answer the question though. You also need to explain why it's not possible to build a brain without this need. In an animal that must not sleep (dolphin), instead of changing the brain (as you might expect), the brain sleeps one half at a time. Ariel. (talk) 12:14, 21 July 2010 (UTC)

There are plenty of problems that are considered outside the purview of scientific practice. Of course, it has been a long struggle over the centuries as various problems considered "outside" of science have found their way into it. --Mr.98 (talk) 23:54, 20 July 2010 (UTC)

Ah, but were they considered outside science by scientists or by other people that wanted to keep control over them? --Tango (talk) 23:58, 20 July 2010 (UTC)

Both. But of course the lines are blurry a lot of the time. Nobody wants to control like the status quo. --Mr.98 (talk) 13:14, 21 July 2010 (UTC)

Mankind can put a man on the moon but forget about plugging a hole in the ground. Also explain the interactions of certain subatomic particles. Also brains in general... ZigSaw 13:11, 21 July 2010 (UTC)

For the record Ziggy, plugging an oil well a mile underwater is much more an engineering problem then a science problem (and last I heard it WAS plugged). Googlemeister (talk) 14:15, 21 July 2010 (UTC)

See Opinion: The Limitations of Science - TIME (Friday, May. 07, 1965).—Wavelength (talk) 01:44, 22 July 2010 (UTC)

Measuring temperatures in Ireland in the late 17th century

When I was a boy, I spent a summer in Northern Ireland; that year, 1995, had what we heard was an all-time high temperature for Belfast, at 86ºF. From what we were told, they'd kept records since the Glorious Revolution. Since the Fahrenheit scale was only proposed in 1724, how was temperature measured in Ireland before that time? Nyttend (talk) 21:58, 20 July 2010 (UTC)

Thermometer#Early_history "However, each inventor and each thermometer was unique—there was no standard scale. In 1665 Christiaan Huygens suggested using the melting and boiling points of water as standards, and in 1694 Carlo Renaldini proposed using them as fixed points on a universal scale. In 1701 Isaac Newton proposed a scale of 12 degrees between the melting point of ice and body temperature" and Category:Units of temperature (earliest about 1701)

I would guess that they had records of comparative but non-standardised temperatures. Galileo is reported to have had a type of thermometere [1] (not to be confused with the Galileo thermometer) and predates the glorious revolution - so maybe they used something like that.77.86.76.212 (talk) 22:54, 20 July 2010 (UTC)

This site is quite interesting http://www.mastco.net/home16.html in that it mentions alcohol in glass thermometers as early as 1641, and refers to one used by Robert Hooke in 1644. It's decribed in his micrographia [2] or [3] or [4] (different editions) It seems likely that they may have used Hooke's scale if they were associated in some way with the Royal Society .77.86.76.212 (talk) 23:09, 20 July 2010 (UTC)

By the way do you have any clues as to who had the records - eg Queen's University Belfast opened 1845, but Ireland was then politically one so maybe Trinity College, Dublin which dates back to 1592. Or perhaps it was some seafaring organisation or similar in Ireland that kept the records? or an amateur clergyman or gentry?? any clues? That might help find the answer. 77.86.76.212 (talk) 23:19, 20 July 2010 (UTC)

Sorry, no clue. As I said, I was just a boy; my parents told me that the all-time-record thing was from the Telegraph. Nyttend (talk) 17:00, 21 July 2010 (UTC)