Talk:Speed of light

See also: Talk:Speed of light/Definition of the metre

Template:WP1.0

Are we ready for FA?

When I returned to this article last week, after about a month I hadn't significantly edited it, I found it even better than I remembered. I am quite positive that it is ready for FA status, but can anyone find any issue with it before I nominate it again? ― A._di_M.^{3rd Dramaout} (formerly Army1987) 18:41, 10 July 2010 (UTC)

I have an issue with the 'Fundamental role in physics section. In this section there important facts about the subject of the article have been relegated to footnotes. Why are they not in the main text? Martin Hogbin (talk) 23:40, 10 July 2010 (UTC)

The Doppler effect and the Terrell rotation are quite irrelevant to the point being made, so I would keep them where they are. The Scharnhorst effect is so small that it might well be never observed during the lifetime of anyone around here, so it was agreed that it was undue weight to put it in the main text.

Anyway, I'm going to move the reference to the relativity of simultaneity and the tachyonic antitelephone back into the main text (being sent to a footnote only comprising one link distracts more than seeing it in the text, IMO). As for the note on one-way vs two-way speed, I think it should be in the main text too, but I seem to remember there once was an opposition to that and was moved to the footnote as a compromise. ― A._di_M.^{3rd Dramaout} (formerly Army1987) 13:56, 11 July 2010 (UTC)

I thought that the article was ready for FA last time, and it is somewhat better now than it was then. The FA review did not bring out significant criticism, and the points that were raised were dealt with (mostly by A. di M.). The review simply died with a whimper from lack of comment. Would it useful to put the article up for peer review before re-nominating it for FA?—Finell 10:24, 20 July 2010 (UTC)

The "Natural sciences and mathematics" section of WP:PR appears to be stagnating right now: I submitted the Physics article eleven days ago and so far only an editor saying " Doing..." popped up. A. di M. (formerly Army1987) (talk) 15:38, 20 July 2010 (UTC)

Does light always travel at c, or does it never travel at c?

These are not rhetorical questions. I'm curious, if c in is the speed of light in a vacuum, and pure vacuums do not exist in nature, why is it claimed that that physical light travels at c? Does light always travel at c as claimed in section 4, or does it never travel at c as claimed in section 5? Isn't it common knowledge that c is a concept distinct from "the speed of light", and that it's just called that for historical reasons? Is anyone willing to argue that c is in fact "the speed at which light travels"? The question is, does light travel at c (y/n)? NOrbeck (talk) 07:32, 19 July 2010 (UTC)

I think the article deals with the issues that you have raised. In current models of physics c is a fundamental constant representing, among other things, the speed of light in a vacuum. You might say, as you suggest, that there is no such thing as an absolute vacuum, indeed there can be no such thing except in a completely empty universe. However, most of physics is based on concepts that do not actually exist so the speed of light is no different in that respect.

Bear in mind also that the article is entitled 'The speed of light' and should therefore have some basis in the natural meaning of these words. The philosophy of this subject has been discussed at some length and we could continue to discuss it ad infinitum but I think the article in its current state reflects the current state of physics regarding the subject reasonably well. That is not to say it could not possibly be improved. Do you have any suggestions? Martin Hogbin (talk) 08:12, 19 July 2010 (UTC)

Re "The question is, does light travel at c (y/n)?" – y --Bob K31416 (talk) 08:24, 19 July 2010 (UTC)

Yes. We 'know' that light travels at c in a vacuum as well as we 'know' most other things in physics. Martin Hogbin (talk) 08:34, 19 July 2010 (UTC)

Hey this is a really good question Bob. The answer is, NO. The physicists are just hoping that their opinion in the matter is the right one. But, lets get real. Most of the time I calculate the speed of light I don't use the value as the physicists give it in their hopeful definitions. That is because I live in the real world and it isn't a vacuum. The statement that it is a constant of physics is simply wishful thinking. In my world I have to figure out what its velocity is really and account for that fact. —Preceding unsigned comment added by 72.64.52.188 (talk) 20:24, 19 July 2010 (UTC)

Re "In my world I have to figure out what its velocity is really and account for that fact." - Regarding having to "figure out" what the speed of light in vacuum "is really", could you give an example of what you mean? Thanks. --Bob K31416 (talk) 20:34, 19 July 2010 (UTC)

I'm unsure what the reader is after myself. On the one hand, it seems that the concern is that the speed of light is c only in free space, and free space is not real. I think the article deals with that issue in discussing the role of a medium upon the speed of light. On the other hand, the concern may be that c is a "defined" value, not a measured one. So saying c is 1 city blocks/ ns isn't too sharp when we don't know what a city block is. It doesn't help much to say a city block is c × 1 ns. That is what is going on with the definition: we don't know what the metre is. It's c × 1/299792458 of a second. Brews ohare (talk) 15:11, 20 July 2010 (UTC)

The way out of this dilemma is to specify circumstances under which the "standard speed of light" is realized. These circumstances are established by careful experiments. That amounts to specifying a "standard medium" and the appropriate correction to convert measurements to their equivalent in free space. Of course, the standard speed can be assigned any value, because all that does is establish the arbitrary unit for speeds. Brews ohare (talk) 15:39, 20 July 2010 (UTC)

The lead section needs expansion

Alright, so there are a few people getting ready to resubmit this to the featured article crowd. Cool, I hope it works out well. I suggest working on the lead section more before renomination, and doing so with a slightly different focus.

There are several knowledgeable editors who have reworked the text and debated all kinds of technical minutiae, and the result is a nice, reasonably comprehensive and stable presentation of the important aspects of the topic. However, the lead doesn't do its job of providing an adequate summary, and more importantly fails to reach out and grab the reader in an entertaining way.

At the moment, the lead has three paragraphs. The first has been through some bloody hand-to-hand combat, and looks pretty good. The second paragraph starts out okay, but quickly descends into equations and values for refractive indexes, which are unneeded in an intro and not particularly entertaining for a general reader. There are more interesting details from the 'propagation of light' section that should be included instead. The third, currently final paragraph of the lead, gives an overview of the history. It's not terrible, but is far too sketchy and misses out on several chances to grab the reader. This paragraph should be expanded. And, though admittedly a personal pet peeve which I tried to eliminate months ago, the first sentence of the paragraph is one of the cheesiest lead-ins to a history section possible — "For much of human history, it was debated..." — give me a break, that is unsourceable and virtually meaningless.

Besides revamping those last two paragraphs, there is additional material conspicuously missing from the lead. The article sections 'practical effects of finiteness' and 'measurement' are poorly represented, if at all, and those sections contain a lot of good material. The lead needs expanding with emphasis on a fun presentation, rather than a dry recitation of facts. Think of Feynman in his book QED, where he says something like "now you know that light not only does not move in straight lines, but doesn't go at the speed of light either". The basic facts are in the article, but it needs more evocative details in the lead, such as providing a mental image of a 17th century cannon firing in order to measure the speed of light. The ingredients are in the pot, now add the spice. Tim Shuba (talk) 09:46, 20 July 2010 (UTC)

In my opinion, the lead is in the best shape that it has been since I have been following this article. This is an encyclopedia article, not a paperback book. Expanding the lead would not be an improvement.—Finell 10:20, 20 July 2010 (UTC)

I agree with Tim that the lead is intended to be a summary of the article, that is one reason why I have previously suggested leaving it until the rest of the article is complete. Once the body of the article is complete, writing the lead should be relatively straightforward. One of the requirements for FA is 'brilliant prose'. I do not think the lead meets this standard yet. Ideally we need a copyedit (maybe rewrite) from a good writer of English under supervision from the techies here. Martin Hogbin (talk) 10:42, 20 July 2010 (UTC)

I disagree - the lead is a summary of the article, but very often articles, especially mathematical ones, cover a lot more ground in a lot more detail and depth than should be in the lead. The lead is meant to be an accessible introduction that encourages readers to read on. Trying to summarise all the content from a technical article will make it less accessible and readable and probably too long. Looking at it I can't see anything seriously wrong with it. --JohnBlackburne^words_deeds 11:00, 20 July 2010 (UTC)

The article (including the lead) might not be perfect, but IMO it is in very good shape, and it's not like featured articles are required to be perfect, nor like it is forbidden to improve an article once it is listed at FA. A. di M. (formerly Army1987) (talk) 15:42, 20 July 2010 (UTC)

BTW, Shuba, if you have ideas about how to improve the lead clear enough to write a draft in a sandbox, please do so... A. di M. (formerly Army1987) (talk) 15:55, 20 July 2010 (UTC)

Remove animated graphic?

Forgive me if this has been thoroughly hashed out before, but in my opinion the animated graphic doesn't really contribute to the article. Most people don't have any real sense of the distance between the Earth and Moon, so showing how long it takes light to travel from one to the other doesn't inform.—Finell 10:40, 20 July 2010 (UTC)

I agree. Martin Hogbin (talk) 10:47, 20 July 2010 (UTC)

I don't like most animations, but I like this one. It's hard to illustrate the speed of light, and while the general reader will have no idea how far the light is travelling in this animation, they know it's a long way. I think the animation provides a good human scale: the reader sees that light takes a bit over a second to travel to the Moon which they have some passing familiarity with. Johnuniq (talk) 11:16, 20 July 2010 (UTC)

I like it. It gives me a good feel of the speed of light for a distance that is just the right size to display the info. --Bob K31416 (talk) 13:50, 20 July 2010 (UTC)

But most people will have at least a rough sense of how big the Earth is, and the picture is to scale. A. di M. (formerly Army1987) (talk) 15:19, 20 July 2010 (UTC)

Agree with A. di M. Also, if people really think that more comparison is needed, why not add the fact that it took the Apollo spacecraft three days to cover the distance that light travels in 1½ seconds. Physchim62 (talk) 15:32, 20 July 2010 (UTC)

Physical constant

The lead says light is a physical constant. It is interesting that that article says:

“A physical constant is a physical quantity that is generally believed to be both universal in nature and constant in time. It can be contrasted with a mathematical constant, which is a fixed numerical value but does not directly involve any physical measurement.”

That contrast can be compared with the last statement in the article:

“As a result, in the SI system of units the speed of light is now a defined constant.[7] Improved experimental techniques do not affect the value of the speed of light in SI units, but do result in a more precise realization of the SI metre.”

An even clearer statement comes from one of sources cited:

“One fallout of this new definition was that the speed of light was no longer a measured quantity; it became a defined quantity. The reason is that, by definition, a meter is the distance light travels in a designated length of time, so however we label that distance - one meter, five meters, whatever - the speed of light is automatically determined. And measuring length in terms of time is a prime example of how defining one unit in terms of another removes a constant of nature by turning c into a conversion factor whose value is fixed and arbitrary.”

— James Jespersen, Jane Fitz-Randolph, From sundials to atomic clocks: understanding time and frequency, p. 280

(my underscore). There appears to be a lack of clarity here. Brews ohare (talk) 11:52, 20 July 2010 (UTC)

Not at all, you are being deliberately obtuse again Brews.

Let's take another example. We have strong reason to believe that all atoms of carbon-12 have exactly the same mass. We can say that the mass is exactly 12 atomic mass units, or we can say that the mass is about 11.996186 "old atomic mass units": the mass of carbon-12 atoms didn't change in 1960 when carbon-12 became the basis of the atomic mass scale in place of oxygen-16! Our units of measurement are mere human constructs, we can choose whatever suits us: the speed of light is the same if you express it in metres per second or in furlongs per fortnight. Physchim62 (talk) 12:08, 20 July 2010 (UTC)

I read that quote when looking at your new article, and I'm not sure what Mr Jespersen is getting at. Nothing has been removed, or changed. The older less accurate ways of doing things are still there if people want to use them. All that's happened is we've through better understanding unified length and time, replacing two arbitrary scales (the second and metre) with one. But he's written a popular science book not a text book or research paper, so hardly a reliable academic source, and I don't think it's of use as a reference here.--JohnBlackburne^words_deeds 12:32, 20 July 2010 (UTC)

John, this source is a NIST publication, and so I'd hesitate to discount it as "a popular science book". It is that, but it is written by an authority on the subject. You are right that the old methods can be resorted to, and maybe they will one day if some revolution in technique occurs to make wavelenght a more accurate measurement than frequency, though it's hard to imagine at the moment. However, the logical basis of the units has to be made clear, and there is a shift in this regard. Brews ohare (talk) 14:21, 20 July 2010 (UTC)

A somewhat more careful wording is this, from Hall and Ye:

redefinition of the International Meter in 1983 in terms of an adopted value for the speed of light, thus reducing by one the number of independent, base physical standards.

— John L. Hall and Jun Ye (2003), Optical Frequency Standards and Measurement

Brews ohare (talk) 13:48, 20 July 2010 (UTC)

Yes, that's better, and agrees with what's in the article. Again you can think of instead as reducing the number units, i.e. scales, which we have to define to do physics, i.e. one less arbitrary thing. That was my objection to the first quote: the only thing removed is one independent and so arbitrary measure. Now there's only one, the second, with the metre defined in terms of it. That definition also includes c and so fixes it in terms of our units. We could if we wanted to do away with the metre altogether and measure everything in seconds, but the metre is far more convenient and familiar so we keep it. But c is also physical constant - it's a property of our universe, not a purely mathematical constant like π or e. We can hypothesise about a universe with a different c, or that c might change over time, but we can't do so with π. That I think is the difference.--JohnBlackburne^words_deeds 14:55, 20 July 2010 (UTC)

In response to Physchim62: The issue here is not what scientists are free to do about standards. The issue here is the choice of wording in a WP article. The term "physical constant" is irretrievably related to measurement in the linked article physical constant. Jespersen says the redefinition has "removed a physical constant". Hall and Ye say it "reduced by one the number of base physical standards". Personally, I like Hall & Ye's terminology better. But whatever the approach chosen, we need a terminology that is clear and consistent, and that does not seem to be the case here. Brews ohare (talk) 14:40, 20 July 2010 (UTC)

The dimensionless number 299,792,458 "does not directly involve any physical measurement", but the dimensionful quantity 299,792,458 m/s does involve some measurement (for example, how long a second lasts). Can you see the difference? A. di M. (formerly Army1987) (talk) 15:22, 20 July 2010 (UTC)

A physical constant has an existence independent of any attempt to measure it. The speed of light is still the same as it was in Rømer's time: or are you denying that, Brews? Physchim62 (talk) 15:29, 20 July 2010 (UTC)

Response to Physchim62. No, not at all. I would hope you would understand this matter yourself, particularly in view of your very apt summary of the situation when you said “Basically, you exchange a measurement uncertainty in the value of a physical constant for a realization uncertainty in the corresponding unit.” To me, this statement of yours very neatly sums the situation where one defines c as an arbitrary value, setting the unit for speeds, and then specifies the circumstances under which the unit will be realized. Brews ohare (talk) 15:51, 20 July 2010 (UTC)

Response to A di M: I do not believe you present the situation usefully by making a distinction between the number 299,792,458 and the definition of c = 299,792,458 m/s. There is no measurement involved in setting c = 299,792,458 m/s, and, for example, Jespersen and also Sydenham say this definition is independent of measurement and arbitrary. I believe Physchim62 has explained the matter very well: one sets up the circumstances under which c is realized; under those circumstances the unit is produced, and what value is assigned to the unit is a matter for the CGPM. Brews ohare (talk) 15:58, 20 July 2010 (UTC)

I believe that a simple formulation like Physchim62's should appear in the lead and that would go a long way to clarify what is meant by an "exact" value for c and the removal of c from the list of "independent, base physical standards". Brews ohare (talk) 16:09, 20 July 2010 (UTC)

I can't see what you mean, i.e. what formulation, and what precisely "like". Don't expect us to work it out for ourselves, post what you think should be changed or inserted if you want other editors' opinions on it.--JohnBlackburne^words_deeds 16:20, 20 July 2010 (UTC)

John: Physchim62's formulation is spelled out in the previous remark. Here it is again: “One sets up the circumstances under which the standard speed of light is realized; under these circumstances the standard speed of light is produced, and what value is assigned to this unit is a matter decided by the CGPM. This procedure exchanges a measurement uncertainty in the value of a physical constant for a realization uncertainty in the corresponding unit. Specifically, c = 299,792,458 m/s exactly, and this standard unit of speed is realized in free space, which as a practical matter, can be realized only approximately.” Brews ohare (talk) 16:43, 20 July 2010 (UTC)

The definition says "vacuum", not "free space". Physchim62 (talk) 17:06, 20 July 2010 (UTC)

I'm not sure how that would fit in - it more an answer to a question, and does not add anything as that information's already in the article, i.e .the article already covers this very well.--JohnBlackburne^words_deeds 16:52, 20 July 2010 (UTC)

You are quite right Brews, the metre standard can only ever be realised approximately, for all the usual 'measurement' error reasons. This point is probably more appropriate for the metre article. Martin Hogbin (talk) 16:59, 20 July 2010 (UTC)

The problem remains that the use of the term "physical constant" contradicts the usage in the article physical constant and the sources I have quoted. Brews ohare (talk) 15:50, 21 July 2010 (UTC)

I agree here, this can be clarified in the section about moden physics. My opinion is that in that section it should be made clear that from the point of view of relativity, c is a conversion factor, see e.g. here

The possible time variation of dimensionless fundamental constants of nature, such as the fine-structure constant $\alpha$, is a legitimate subject of physical enquiry. By contrast, the time variation of dimensional constants, such as $\hbar$, $c$, $G$, $e$, $k$..., which are merely human constructs whose number and values differ from one choice of units to the next, has no operational meaning.

Count Iblis (talk) 18:06, 21 July 2010 (UTC)

And, yet again, we get the Duff paper coming out, dispite the fact that it fails WP:RS (it was rejected from a peer-reviewed journal) and without any mention of the refutations that the author honourably includes as appendices to the preprint. We could simply look at standard usage of the term "physical constant" as in, for example, the list of values regularly published by CODATA (the 2006 list is available at ‹The template CODATA2006 is being considered for deletion.› Mohr, Peter J.; Taylor, Barry N.; Newell, David B. (2008). "CODATA Recommended Values of the Fundamental Physical Constants: 2006" (PDF). Reviews of Modern Physics. 80 (2): 633–730. arXiv:0801.0028. Bibcode:2008RvMP...80..633M. doi:10.1103/RevModPhys.80.633. Archived from the original (PDF) on 2017-10-01.

and here): the speed of light is not only included in the list, but it is even the first on the list in most printed summaries! To pertend that the speed of light is not a physical constant would be to give undue weight to a minority point of view, regardless of the intellectual merits of that point of view. Physchim62 (talk) 19:07, 21 July 2010 (UTC)

I agree. Martin Hogbin (talk) 19:12, 21 July 2010 (UTC)

I don't agree. It is a reliable source despite being rejected from Nature. To see this, you have to look at the citations of the paper, how it is cited, by whom it is cited and how much those citing papers themselves are cited. What Duff was saying in his paper is now something most people who look at time variation of constants take very serious; they take great care to make sure the quantity they are sudying is dimensionless. In contrast, if you look at the paper by Paul Davies, and evaluate that in the same way, you see that his argument which also the Referees made who rejected Duff's paper, have little support in the physics community. People are not looking at whether c or e is changing (if alpha is changing).

So, I would say that to not mention that c is regarded to be a rather irrelevant conversion factor, leaves out the very important POV of many theoretical physicists. But I won't make a big issue out of this, for a FA review this is not a stumbling block. Count Iblis (talk) 21:24, 21 July 2010 (UTC)

B.t.w., I think there is another paper by someone else that makes a simlar argument as Duff did that did get published in Nature... Count Iblis (talk) 21:24, 21 July 2010 (UTC)

I agree with Physchim62 and Hogbin on this point.—Finell 23:01, 21 July 2010 (UTC)

Then you are wrong as well. I'm not saying that the article should not mention that c is a physical constant (as that is what many sources say), but simply that it is a wideheld opinion in the theoretical physics community that it isn't, i.e. that it is nothing more than an irrelevant re-scaling constant that has no operational meaning whatsoever. There are tons of articles that say this in so many words. My opinion is that the section on modern physics is misleading because it doesn't mention this. It gives the (sort of) high school POV on units that actually causes students at university to unecessarily struggle a bit when they have to learn Natural Units. Sometimes a mental block prevents them from seeing that c = 1 is not "wrong", "inconsistent", "paradoxical", "just a trick as you ultimately have to put back c to its correct value", or whatever.

I'm not going to make a big issue out of this; I won't change the article on this point, but the editors here have to consider that the irrelevance of c as a genuine physical constant is a wideheld opinion, whether they agree with it or not, and that the article does not discuss this properly. Count Iblis (talk) 14:42, 22 July 2010 (UTC)

It's a respectable opinion that held by a significant number of physicists, but it's not the only opinion and so we shouldn't present it as if it is. This is an article that made it past the journal referees and that expresses both Duff's opinion on the matter and two other contrasting points of view. Physchim62 (talk) 23:25, 22 July 2010 (UTC)

The problem remains that the use of the term "physical constant" contradicts the usage in the article physical constant, is not used consistently throughout Speed of light, and contradicts the sources I have quoted, for example, Jespersen. Brews ohare (talk) 12:08, 22 July 2010 (UTC)

Maybe it's the physical constant article that needs tweaking on this point (and others) ;) Physchim62 (talk) 23:25, 22 July 2010 (UTC)

Whatever our physical constant article claims, the standard usage of the phrase doesn't exclude values which are exact in a particular system of units, or conversion factors. See [1] for example. A. di M. (formerly Army1987) (talk) 16:01, 23 July 2010 (UTC)

Count Iblis has accurately summarized the situation. Perhaps the underlying problem here is that (IMO) there is in reality a phenomenon called ‘light’ and it travels at fixed speed (according to relativity, anyway). That phenomena underlies the set-up of the SI units. However, it is easy to confuse the usage of c in the SI units, which is simply a conversion factor for that system of units, with the speed of travel of the fundamental phenomenon ‘light’. The ability of the mind to cope with the same name for different things is very limited. Brews ohare (talk) 15:01, 22 July 2010 (UTC)

Unfortunately, this article has to cope with the fact that the speed of light is much more than, well, the speed of light! It's a fair point that the role of the speed of light in defining the metre is physically insignificant, as the metre is simply a human construct that we can define as we wish. On the other hand, we do have to give the numerical value for c, as that's what many readers will have come looking for. And once we give the numerical value we have to explain that it is fixed by definition and so has no measurement uncertainty in SI units. And we have to say a lot of other things about the speed of light, so we cannot give undue weight to its role in the SI.

But it remains that there is this phenomenon called 'light', and that it travels at a fixed speed (as far as we can tell). In that sense, the speed of light is a "physical constant". The speed doesn't depend on what day of the week it is, nor on what unit system we choose to express it in. On the other hand, we have to express it in some unit system to be able to compare it with anything else, and SI units are the conventional choice. Physchim62 (talk) 23:25, 22 July 2010 (UTC)

In a ‘time-of-flight’ system of units based upon a standard speed, there is no requirement that one know what that speed is, or whether it is fast or slow. The only requirement is that one be able to identify situations where that standard speed is actually realized. The choice of units decides the numerical value assigned to the speed in that system, and that value is “fixed and arbitrary”, to quote Jespersen. Brews ohare (talk) 15:21, 22 July 2010 (UTC)

Brews'edits

Brews, since 12:00 UTC you have made 9 edits to this article. I have not looked at them all but they seem to be a mixture of uncontentious edits and ones making more of a point. This was the kind of thing that got you into trouble before. In the light of the fact that a lot of work has gone into this article to try to get it ready for FAR, can I suggest that you slow down a bit to let others review your edits. Idealy, I would suggest that, at this stage, all edits are discussed first. Martin Hogbin (talk) 13:35, 20 July 2010 (UTC)

Hi Martin: I think the edits are all uncontentious, involving a bit of added detail from the sources already cited. I don't think any of them are "making a point".

I haven't made any edits on one point that might be considered. There is some lack of clarity about just how an interferometer can be used to "measure a wavelength". Basically, the interferometer can compare path lengths. Thus, to measure a wavelength, a known length has to be compared against.

A tricky issue is this: let's say you start with a metre bar as the standard, and you want to establish that the metre is 1 650 763.73 wavelengths in vacuum of the krypton-86 atomic transition. The standard metre is not known to the kind of accuracy suggested by this number. Thus, the standard metre is being replaced by a more precise standard. Evidently, the new standard can be selected at any value within the precision with which the old standard is known. I don't know the actual error bars on this one, but suppose the old standard is between 1 650 762 wavelengths and 1 650 765 wavelengths, a wavelength being about 605 nm. Then I'd say the committee is free to pick any value in this range. That automatically decides the speed of light too, given λ = cf.

The article hasn't mentioned these issues of upgrading a standard. Brews ohare (talk) 14:03, 20 July 2010 (UTC)

The point here is that you can't measure c to greater precision than the standard allows, and a more precise value for c implies a more precise standard of length. Of course, this all refers to the "old" methodology, as is discussed in the section on interferometry. Brews ohare (talk) 14:29, 20 July 2010 (UTC)

What do you mean by 'the precision with which the old standard is known'? At the time, it was the standard. Martin Hogbin (talk) 14:45, 20 July 2010 (UTC)

HI Martin: The idea is this: the standard meter bar is the distance between a couple of scratches on a bar kept under "standard conditions". That definition has an inherent vagueness related to the role of the "standard conditions" on the separation of the marks (are the marks reproducibly at the same separation at all times) and also the marks are physical scratches, let's say, and so just where they are located is a ± issue too. Of course the krypton source has its problems that way too, but they are on a finer scale. The modern laser has its problems too, but they are on an even finer scale. Each time the standard is refined, we have a more precise version of the metre. Consequently , we also have a more precise value for c when c is thought of as a distance/time or as λf. Is this responsive to your query? Brews ohare (talk) 15:00, 20 July 2010 (UTC)

OK, you mean the precision with which the old standard can be transferred to the new system. (I am not sure that I have used the right technical terms there.) I guess this is always the case, when the equipment is upgraded there is some judgment to be used when setting the new standard, right up to and including the SoL standard that we now have. I am not sure that this is all relevant to a SoL article. Martin Hogbin (talk) 16:52, 20 July 2010 (UTC)

The relevance is apparent in discussion of the evolution of technique: one cannot obtain a more precise value for the speed of light without introducing a more precise standard for the metre. Of course, the modern approach is to define λ =c/f so the metre automatically is refined as the frequency measurement is improved. That was not so previously. Brews ohare (talk) 17:57, 20 July 2010 (UTC)

You mean apart from the sentence "Improved experimental techniques do not affect the value of the speed of light in SI units, but do result in a more precise realisation of the SI metre."? Never mind, if you're happy all the better! Physchim62 (talk) 08:28, 22 July 2010 (UTC)

Too much personal criticism

There's too much personal criticism on this article's talk page IMO. It can be disruptive. That's why there is the policy WP:NPA which says,

"Do not make personal attacks anywhere in Wikipedia. Comment on content, not on the contributor."

I feel that an article talk page should be used for discussing the article, rather than discussing editors personally. If an editor has some useful criticism to make of another editor, perhaps it is best to make it on the subject editor's talk page. Also, if an editor doesn't like another editor's change of the article, simply be bold and revert it, rather than criticize an editor personally on this talk page. Regards, --Bob K31416 (talk) 14:13, 20 July 2010 (UTC)

Hi BobK: A most temperate observation, and in keeping with the objectives of the Talk page to be a forum to improve the article, not its contributors. Thank you. Brews ohare (talk) 14:18, 20 July 2010 (UTC)

Brews is known for his disruptive editing on this and related subjects, including engaging in tendentious debates and soapboxing: so sayeth ArbCom. He has only recently come back from a topic ban covering all physics-related topics. Any sign of a resumption of his disruptive behaviour on the very page that was the nucleus of his previous activities is of great concern. It is up to Brews to demonstrate that he can edit constructively on the subject, not the other way round. Physchim62 (talk) 15:20, 20 July 2010 (UTC)

Hey, some tolerance is called for here. This particular article is a tough one, and involves a lot of nice points and subtle distinctions. It is natural for editors to become a bit frustrated over attempts to express things right, because it isn't simple, and iteration is required. However, I hope that suggestions for clarifications can be taken as exactly that. Brews ohare (talk) 15:46, 20 July 2010 (UTC)

It would help if you were to discuss things first though rather than make a whole bunch of edits in one hit. Martin Hogbin (talk) 16:54, 20 July 2010 (UTC)

Bold, revert, discuss

As this article is hoping to be ready for FAR soon, can I suggest that everyone here adopts the edit, revert, discuss rule. That is to say make your change, but not too many at once, if your change is reverted, discuss it before putting it back. Martin Hogbin (talk) 17:18, 20 July 2010 (UTC)

And can we try not to discuss too many points at once, especially if they all come from the same editor? All editors have the right to comment on the article, but there are also procedures availble for ensuring that single individuals do not crowd out others. Physchim62 (talk) 00:30, 22 July 2010 (UTC)

Lead sentence

Present lead:

The speed of light, usually denoted by c, is a physical constant representing the speed at which light and all other electromagnetic radiation travels in vacuum. Its value is exactly 299,792,458 metres per second (approximately 186,282 miles per second).

Proposed replacement:

Light and all other electromagnetic radiation travels in an idealized vacuum at the speed of light, usually denoted by c. In the ideal vacuum, the speed of light is defined to have the value c = 299,792,458 m/s exactly (approximately 186,282 miles per second), setting the unit of speed against which all other speeds are compared. In practice, the ideal vacuum is only approximated by real-life media, and measured speeds are corrected to refer to the ideal vacuum. Thus, the speed of light observed in a real-life medium never is known exactly.

Purpose: The speed of light having a defined value is a bit hard to swallow. The notion that it can have an exact value in an ideal medium is easier to swallow, and in real life measurement is always involved. Brews ohare (talk) 17:24, 20 July 2010 (UTC)

No, the current version is much better: mostly as it's grammatical and clear, unlike your suggested replacement. And please read WP:LEAD#First sentence for why it should start "The speed of light is...".--JohnBlackburne^words_deeds 17:35, 20 July 2010 (UTC)

Proposed replacement:

The speed of light, usually denoted by c, is the speed of travel of light and all other electromagnetic radiation in an idealized vacuum. In the ideal vacuum, the speed of light is defined to have exactly the value c = 299,792,458 m/s (approximately 186,282 miles per second), setting the unit of speed against which all other speeds are compared. In practice, the ideal vacuum is only approximated by real-life media, and measured speeds are corrected to compare with the ideal vacuum. Thus, the speed of light observed in a real-life medium never is known exactly.

Brews ohare (talk) 17:51, 20 July 2010 (UTC)

You seem to be the only person who finds the speed of light having a defined value is a bit hard to swallow. It is not at all clear what, 'measured speeds are corrected to compare with the ideal vacuum' means. This is not an article about vacua. Nothing that is not a defined quantity is ever known exactly. Martin Hogbin (talk) 18:07, 20 July 2010 (UTC)

I see we have reached the point of diminishing returns. The problem remains that the use of the term "physical constant" contradicts the usage in the article physical constant and the sources I have quoted. Brews ohare (talk) 19:01, 20 July 2010 (UTC)

I see no contradiction. Martin Hogbin (talk) 21:38, 20 July 2010 (UTC)

For clarity, would changing the existing text to read "Its value is defined as exactly..." to emphasize that this is a defined, exact quantity be helpful? I am inclined to argree with Martin and John's concerns that the proposed changes offered are confusing and unnecessarily wordy. The lede is not a good place to attempt to introduce subtle points about definition versus measurement. TenOfAllTrades(talk) 19:18, 20 July 2010 (UTC)

I don't think so: I think the fact it's defined rather than measured is too subtle a point for the first paragraph. After all it's defined to the exact value to nine significant figures that was found by measuring it, so that's also the measured value. And it's as much a definition of the metre as of the speed of light. So I would leave it as it is. Very smart readers might pick up on the "is exactly" as unusual, and then read on (to the third paragraph initially) to find out why, but most will probably ignore the "exactly" and pick up the first two things everyone should know: it's a constant and it's a very big number.--JohnBlackburne^words_deeds 19:48, 20 July 2010 (UTC)

Also it is not, strictly speaking, defined, it is fixed by definition, of the metre. Martin Hogbin (talk) 21:38, 20 July 2010 (UTC)

The notion that one should let slide a point like this because only "smart readers" will notice and everyone else will be stunned by the fact it is a "very big number" is patronizing and also a poor basis for content. Brews ohare (talk) 20:57, 20 July 2010 (UTC)

The detail is given on the body of the article, as always. Martin Hogbin (talk) 21:38, 20 July 2010 (UTC)

Everyone else can read on: the details are there, starting in the third paragraph. But the lead is meant to be accessible, and it certainly won't be if we try and load too much into the first sentence. The relationship between the definitions of light speed and the metre is of only historic interest to most readers and so much less important, and is anyway well covered already in the lead.--JohnBlackburne^words_deeds 22:08, 20 July 2010 (UTC)

Brews's version would suggest that the ideal vacuum is defined in terms of the speed of light (as David Tombe claimed), rather than the other way round. A. di M. (formerly Army1987) (talk) 13:37, 21 July 2010 (UTC)

Brews is still flogging his dead donkey of a hypothetical standard state – note how he uses "ideal" or "idealized" four times in a single paragraph! The definition of the metre simply says "in vacuum", and so should we. The important point is that it is not the speed of light in air. It is the speed of light in outer space, as near as can be measured, given that outer space is non-dispersive. It is also the physical constant relevant to the interaction between a proton and an electron in a hydrogen atom. Or maybe Brews thinks that a few pesky air molecules might get between the electron and the proton... Physchim62 (talk) 14:30, 21 July 2010 (UTC)

Regarding the part of the proposed change, "In practice, the ideal vacuum is only approximated by real-life media, and measured speeds are corrected to refer to the ideal vacuum." - In the most precise measurement of the speed of light, could anyone give the excerpt from the corresponding journal article that discusses that work's correction for not having a perfect vacuum? Thanks. --Bob K31416 (talk) 15:09, 21 July 2010 (UTC)

Physchim62 & A di M: Gentlemen: My wording is less simple and straightforward than it might be.

Physchim62, I thought we were on the same page and I actually quoted you word for word. You pointed out that the exact definition of c and its removal from the sphere of measurement came at the expense of specifying the circumstances where this standard defined speed could be realized, and the measurement error of the older method was replaced by an uncertainty in how accurately these ideal, (excuse the word), ideal circumstances were realized in practice.

Now how about forgetting about nitpicking, and finding some way to get this point across in a simple fashion in the introduction? Eh? Brews ohare (talk) 15:52, 21 July 2010 (UTC)

BobK: It would indeed be helpful to find some such statement. If one looks at the NIST literature you can find discussions of corrections made for measurements in air. They take the form of determining the refractive index of the air. As for measurements made in partial vacuum, it appears that the errors introduced by fringe counting are dominant, and they are discussed in the literature on interferometers. These errors due to imperfections in the optical paths, imbalance in the paths, mirror problems, and other errors due to the failure to obtain a plane wave, failure to achieve a monochromatic source. These issues are a bigger concern than the inadequacy of the vacuum. Brews ohare (talk) 15:26, 21 July 2010 (UTC)

Does that mean that you aren't aware of any discussion of a correction for not having a perfect vacuum, in the journal article on the most precise measurement of the speed of light? --Bob K31416 (talk) 15:52, 21 July 2010 (UTC)

Hi BobK: That is the case. I haven't done an exhaustive search. The paper by Evenson et al. from 1972 is the only one I've looked at closely. It doesn't mention the error due to the non-ideality of the vacuum. What it does mention in calculation of the errors is: (i) the asymmetry of the Kr line: modeling of this asymmetry is needed to interpret the fringe pattern (ii) the incoherence of the Kr source. They refer to a paper detailing the error analysis to be published by two of the authors (Barger & Hall), but so far as I can determine this error analysis paper never was published. Brews ohare (talk) 17:12, 21 July 2010 (UTC)

Follow-up: The only explicit reference I've found to the quality of the vacuum is in spectroscopy, where a line is observed as the vacuum is pumped down, and the vacuum is considered adequate when the line properties cease to depend on the partial pressure. The inferred speed of light is then compared to the standard as a reassurance that the vacuum is good, but no effort is expended to account for the discrepancy: all that is wanted is that it is ‘small’. Brews ohare (talk) 19:06, 21 July 2010 (UTC)

I'm getting the impression that there is negligible effect when the vacuum is "adequate" for the measurement of the speed of light with laser beams. Perhaps the concentration of photons in a laser beam is so much greater than the concentration of residual atoms in the space occupied by the beam, that only a negligible percentage of photons in the laser beam interact with the residual atoms, and to the precision of the measurement has no effect on the result for the measured speed of light. --Bob K31416 (talk) 20:26, 21 July 2010 (UTC)

In effect, what Brews is saying is that there is no evidence that the speed of light as measured in a laboratory vacuum is different from the limiting value. That is that the limiting value will be within the measurement uncertainty of the determinations. And so there is no justification for his tirades about ideal vacua and the such like. Brews should stop his soapboxing and tendentious editing, on this aspect and others. Physchim62 (talk) 20:43, 21 July 2010 (UTC)

Please see Too much personal criticism. Thanks. --Bob K31416 (talk) 21:02, 21 July 2010 (UTC)

And please see my reply to your previous comments there. There are a remarkable number of editors replying to Brews' even more remarkable number of "queries". Physchim62 (talk) 21:24, 21 July 2010 (UTC)

It is not the concentration of photons relative to the density of atoms that is important. If you have a certain density of atoms, then you'll have a certain index of refraction that determines the speed of light. I think that in ultrahigh vacuum you'll typically have a residual plasma and then you can treat the situation as presented here. Count Iblis (talk) 21:03, 21 July 2010 (UTC)

Well Count Iblis, I suppose I could elaborate on my point and further discuss your point, but for the purpose of developing this article we need to use reliable sources. Do you know of any reliable source that says that the most precise measurement of the speed of light made a correction for not having a perfect vacuum? Regards, --Bob K31416 (talk) 21:16, 21 July 2010 (UTC)

If you have a certain number density of atoms you will, in theory, have a refractive index which is not equal to one. On the other hand, below a certain number density of atoms, you will have a refractive index that is experimentally indistinguishable from one. Brews' reading of interferometer literature suggests that this latter number density is attained in practical laboratory vacua, not to mention in outer space. I hardly see how the ionosphere is a good model for a vacuum in terms of its interaction with electromagnetic radiation (as the Count seems to suggest) – quite the opposite, in fact! Physchim62 (talk) 21:24, 21 July 2010 (UTC)

Physchim62 is sort of correct. The point is that the errors due to various problems with interferometry make the error in measuring wavelength so large that the added error due to imperfect vacuum is secondary. However, that is not to say that because of practical difficulties one should simply discard the point that one wants to have a perfect vacuum in principle. The logical basis for the whole matter is based upon Physchim62's earlier point that the speed of light is defined to appear in vacuum, and so one should try to obtain a vacuum, or make theoretical corrections for the medium actually used so as to compare the results with the standard speed. I'm left with the notion that Physchim62 wants to back away from this view he expressed earlier.

In any event, if one wishes to understand the statement of an exact speed of light, it is inescapable that one has to accept that it occurs under specific ideal circumstances that can be ensured only imperfectly. The fact of imperfection has nothing to do with its being small: the logic of the situation is separable from the practice. Just like geometry: Geometry is the science of correct reasoning on incorrect figures. Brews ohare (talk) 22:06, 21 July 2010 (UTC)

I'm not "backing away" from my previous position at all. Brews has admitted that, in a decent laboratory vacuum, the corrections due to the presence of residual gas are insignificant compared to the other measurement uncertainties. Note that this can be tested: you can flush the apparatus with helium or with sulfur hexafluoride prior to the experiments and then take it down to the same pressure as you did from air to see if you get the same results. In whatever case the experimenter has two options:

1. if the effect cannot be completely removed, to get as many data points as possible and extrapolate to zero interference; or
2. if, after extrapolation as above, it is possible to work in a régime of negligeable interference, to work in said régime.

Brews' own comments admit that working in a decent laboratory vacuum is working in a régime of negligeable interference for length measurements, and hence for SoL measurements. Again, this is simple to test: you measure the wavelengths of various sources (against an internal standard) and see if the different ratios of wavelengths differ from the ratios of their frequencies.

Brews seems to wish that "vacuum" in the definition of the metre means something different from "limit of zero pressure": it seems obvious to me that it doesn't. Physchim62 (talk) 00:21, 22 July 2010 (UTC)

I have been quite clear. Physchim62 is addressing nothing of what I have said. Brews ohare (talk) 05:44, 22 July 2010 (UTC)

Physchim62 has addressed everything that you have said that is relevant to the lead sentence. Martin Hogbin (talk) 08:21, 22 July 2010 (UTC)

Brews, Drop it. There's a clear consensus against your changes, you're never going to change that, your continual efforts simply waste everyone's time, including your own.--JohnBlackburne^words_deeds 08:43, 22 July 2010 (UTC)

Martin & John: The unaddressed issue is: If one wishes to understand the statement of an exact speed of light, it is inescapable that one has to accept that it occurs under specified ideal circumstances (measurement in vacuum) that must be approximated. The approximation is still an approximation, no matter how accurate it may be: the logic of the situation is what underlies the use of an exact definition. I will insist no further, but I am dismayed at your unwillingness to come to grips with this simple point (despite Physchim62's clear enunciation of it), and instead to digress upon irrelevant details of the implementation of the practical approximation to vacuum. Brews ohare (talk) 14:38, 22 July 2010 (UTC)

Paragraph in interferometry

Present paragraph:

The main difficulty in measuring c through interferometry is to measure the frequency of light in or near the optical region; such frequencies are too high to be measured with conventional methods. This was first overcome by a group at the US National Institute of Standards and Technology (NIST) laboratories in Boulder, Colorado, in 1972.

Suggested replacement:

A major difficulty in measuring c through interferometry is to measure the wavelength. The accuracy of this measurement improves as the wavelength is shortened, but shortening the wavelengths places higher demands upon the frequency determination. As the the frequency of light is increased from the microwave region (Froome used 72 GHz) into or near the optical region, such frequencies are too high to be measured with conventional methods. This was first overcome by a group at the US National Institute of Standards and Technology (NIST) laboratories in Boulder, Colorado, in 1972.

Reference

K D Froome (1958). "A new determination of the free-space velocity of electromagnetic waves". Proc R Soc London Ser A. 247: 109–122.

Purpose:

The point of this text is that accuracy is improved by going to shorter wavelengths because the interferometer fringes are clearer at short wavelengths, allowing greater precision. The earlier measurements were made at microwave frequencies, and it is pointed out here that the Boulder group was able to extend frequency determinations to the optical region, allowing the use of short wavelengths. IMO the present paragraph inadequately conveys the main point about short wavelengths, and jumps to the frequency issue without the necessary preamble, thereby missing the entire point of the discussion. Brews ohare (talk) 17:44, 20 July 2010 (UTC)

Issues concerning the accuracy of optical interferometry, which was used for accurate length measurements, were around for years before the Boulder measurements These were all well understood and dealt with in the best way known. The real breakthrough referred to in this paragraph was the ability to measure the frequency of optical lasers and thus relate optical frequency interferometric measurements directly to the speed of light. Martin Hogbin (talk) 17:55, 20 July 2010 (UTC)

Martin: Here again we have differences that are largely semantic. The source “Speed of Light from Direct Frequency and Wavelength Measurements of the Methane-Stabilized Laser” (Evenson, Wells, Petersen, Danielson, Day, Barger and Hall) reads as follows:

At shorter optical wavelengths the accuracy of the wavelength determination increases. ... The 100-fold improvement in the presently reported measurement comes mainly from the increased accuracy possible in the measurement of the shorter wavelength.

As you know the speed of light is c=λf. The error is thus the sum of the errors in λ and in f. That error is dominated by the error in λ at longer wavelengths, and by the error in f (prior to this work) at shorter wavelengths. So the authors prefer to compare the two cases in terms of the errors in each factor, and see that the error in the product is now dominated by the error in λ, just as it was before, but of course this error is less at shorter λ.

The authors' view, and the one I express in the above text replacement, is that the breakthrough in frequency measurement would be useless if it were not that the error in λ is much less. This reason for the success now is stated, which is omitted entirely in the present paragraph, and the nature of the breakthrough is described just as it is in the present paragraph. Brews ohare (talk) 18:23, 20 July 2010 (UTC)

I cannot read that paper from the link. What year was it and what wavelength was the laser. Your quoted sentence is ambiguous. Martin Hogbin (talk) 21:33, 20 July 2010 (UTC)

Phys Rev Lett vol 29 No 19 pages 1346ff Nov 1972 methane stabilized laser at 3.39um. It is the paper cited in footnote 93 of the article. Brews ohare (talk) 00:34, 21 July 2010 (UTC)

Thanks, footnote 94 in the article gives the whole story. It seem to be as I thought. At the time of the Boulder measurements, which were a few years before the change in definition of the metre, the situation was that the speed of light had been most accurately measured by using microwaves of known frequency and measuring their wavelength. The main accuracy limitation was the measurement of the 72 GHz microwave wavelength. Calibration of distance would have been against a physical standard, using (visible) optical interferometry before 1960 and by interferometry using the specified krypton line after that time.

The point is that, long before the change to the SoL standard for the meter, distance was routinely being accurately measured by optical interferometry. The big step forward was the ability to measure the frequency of the light used in interferometry against the caesium frequency and time standard. The fact that the wavelength of light could be determined much more accurately than that of microwaves was well-known long before the Boulder experiments. Martin Hogbin (talk) 11:21, 21 July 2010 (UTC)

To put it another way, if you're working with microwaves, the problem is to measure the wavelength because measuring the frequency is easy; if you're working with visible (or near-visible) light, the problem is to measure the frequency because measuring the wavelength is easy. I think we should really change the title of the section to something like "Heterodyne frequency measurements" to make that even clearer. Physchim62 (talk) 14:39, 21 July 2010 (UTC)

Gentlemen: Everybody is on the same page that the measurement of frequency was a roadblock that was lifted. What you also both know is that fringes are clearer at short wavelengths. Martin says that was well known long before this paper, and of course that is why the push was on to improve frequency measurements. Everybody agrees about that too. Without the improvement in frequency measurement a better value for the speed of light would not result. The error in c reduced because it became possible to take advantage of the lower error in fringe counting and that advantage was opened up because of the advance in frequency measurement.

Now, I think we all know these two facts. What I think is a problem is that the WP paragraph assumes the reader knows how these two things work together, and doesn't say all this. The article stresses that frequency measurement improved, but doesn't bother to say that the significance of this improvement was that it opened the door to the better regime for wavelength measurement. The quote I provided from the paper is the authors' expression of this fact, and it is missing from the WP article. Brews ohare (talk) 15:03, 21 July 2010 (UTC)

I reworked it a bit. It was not clear why the paragraph was speaking about microwave wavelength measurement being problematic, since there was no statement that microwave interferometry had been attempted to measure c. I tried to get the whole idea into the text, and moved the quotation to the footnote as background. I also corrected some transcription errors in the quotation and added the last sentence of the quoted paragraph, which puts the whole idea together. If someone wants to add more history, more could be said in the text about the 1958 experiment that Brews's source discusses; this addition is not necessary, in my opinion, but I wouldn't oppose a succinct sentence or three about it. I also added the page number, which the MOS requires (when there is one) for all direct quotations.—Finell 02:58, 22 July 2010 (UTC)

Finell: Looks good to me; thanks. Brews ohare (talk) 05:47, 22 July 2010 (UTC)

Reasons for the CGPM decision

The section Speed of light#Increased accuracy and redefinition of the metre says the following:

“This made its wavelength, and hence the length of the metre, uncertain, because the definition did not specify what point on the line profile (e.g., its maximum-intensity point or its centre of gravity) it referred to.[Note 8] To get around this problem, in 1975, the 15th Conférence Générale des Poids et Mesures (CGPM) recommended using 299,792,458 metres per second for "the speed of propagation of electromagnetic waves in vacuum".”

This paragraph suggests that the main problem addressed was the lineshape of the krypton source. Although this was certainly a problem, the CGPM resolution (which should be linked, as it is the full statement of the decision) itemizes a number of other considerations. In particular, the improvement in stability and coherence of laser sources made the krypton discharge lamp obsolete.

Supposedly, the CGPM could therefore simply have replaced the krypton source with a more coherent and monochromatic laser source, thereby avoiding the problems with the krypton lamp. Instead of simply paralleling the old definition of the metre and substituting a better source, they avoided specifying any particular source or transition. Instead in their Mise en pratique (which also should be linked) they specified a “CIPM list of approved radiations for the practical realization of the metre, 1997: frequencies and vacuum wavelengths”. These various sources were related to one another “exactly by the relation λf = c₀, with c₀ = 299 792 458 m/s, but the values of λ are rounded.”

The better accuracy of frequency measurements compared to wavelength measurements meant that relating the sources through their frequencies was much more accurate than an attempt to relate their wavelengths. This simple and accurate method for interrelating sources was the main driver behind the switch to a new definition. It was recognized that the most accurate source would change over time, and that different applications would be better served by different sources. This accurate comparison coupled with its flexibility in selecting a source is the underlying reason for the change to a defined value for c.

Of course, I expect a lot of debate over this matter. However, we should understand the issues involved. Brews ohare (talk) 13:47, 22 July 2010 (UTC)

Perhaps the link to the mise en pratique should be to the more recent version: CI-2002. It contains this language indicative of the wide range of considerations behind the standard:

considering

• that science and technology continue to demand improved accuracy in the realization of the metre;
• that since 1997 work in national laboratories, in the BIPM and elsewhere has identified new radiations and methods for their realization which lead to lower uncertainties;
• that there is an increasing move towards optical frequencies for time-related activities, and that there continues to be a general widening of the scope of application of the recommended radiations of the mise en pratique to cover not only dimensional metrology and the realization of the metre, but also high-resolution spectroscopy, atomic and molecular physics, fundamental constants and telecommunication;
• that a number of new frequency values with reduced uncertainties for radiations of high-stability cold atom and ion standards already listed in the recommended radiations list are now available, that the frequencies of radiations of several new cold atom and ion species have also recently been measured, and that new improved values with substantially reduced uncertainties for a number of optical frequency standards based on gas cells have been determined, including the wavelength region of interest to optical telecommunications;
• that new femtosecond comb techniques have clear significance for relating the frequency of high-stability optical frequency standards to that of the frequency standard realizing the SI second, that these techniques represent a convenient measurement technique for providing traceability to the International System of Units (SI) and that comb technology also can provide frequency sources as well as a measurement technique;

Brews ohare (talk) 16:30, 22 July 2010 (UTC)

At a minimum, this section should be modified to include links to the decision and the mise en pratique, and the discussion of the krypton source modified to avoid the impression that its lineshape was the sole factor behind the switch in CGPM's approach to the speed of light. Brews ohare (talk) 14:50, 22 July 2010 (UTC)

No, I think those are fair points, and your summary of the reasons behind the change is pretty much the same as my understanding of the matter. We should certainly cite the two CGPM resolutions (1975 and 1983), and I'll do that in a moment if nobody beats me to it ;) I'm a bit more wary of citing the mise en pratique, because it changes regularly and because it is the sort of technical detail that would be better in the metre article.

As for other changes, I'll see what others have to say. I think the article already covers the most important point, which is to explain why the krypton source was considered no longer satisfactory. The stability issue is actually quite secondary – there are other ways of getting round a definition which is no longer a good practical method for measuring the quantity concerned, such as the International Temperature Scale or conventional electrical units. My first choice for something to add (if we're going to add anything, we also have to consider the weight of the various topics in the article) would be the comment in the 1983 Resolution "these various forms, making reference either to the path travelled by light in a specified time interval or to the wavelength of a radiation of measured or specified frequency [...], have been recognized as being equivalent". Physchim62 (talk) 17:00, 22 July 2010 (UTC)

The mise en pratique of [2002 recapitulates the earlier resolutions and standards, and adds to them. Therefore, it is a better source than the more recent versions that are simply lists of additions to the standard radiations or revisions of accuracies. I believe a link to the 2002 version of the mise en pratique is suitable because (i) it provides details of what the standards are, (ii) provides a very detailed rationale, (iii) alerts the reader to where such info can be located, and (iv) provides a more up-to-date perspective than the 1983 resolutions. It is therefore a useful link and adds to the value of the WP article. Brews ohare (talk) 17:31, 22 July 2010 (UTC)

Table heading

I just reverted the change to the table heading as it is the history of measurements: from the first to the decision to fix it to match the then current value: that put an end to needing to measure it, as the text makes clear, so ends the sequence of measurements, while a table of values should include other earlier estimates. Also the change wasn't grammatical and for no obvious reason broke a perfectly good link.--JohnBlackburne^words_deeds 18:50, 22 July 2010 (UTC)

John Blackburne has made this reversion of a change in title of this table. His reasons are: “it is the history of measurements, not values. Rv ungrammatical change and bad edit)”

Unfortunately, not all the items in this table are measurements. The last entry is a definition. A more correct title is therefore: History of values of c. I don't see much quarrel about this. If it an ungrammatical title, at least it is accurate. Referring to a ‘definition’ as a ‘measurement’ is not accurate. Brews ohare (talk) 18:48, 22 July 2010 (UTC)

Measurement section intro

The section by this name states:

“The techniques described here are therefore no longer needed to measure it, though they can be used to confirm its value.”

The difficulty with this statement is that it suggests the defined value of c=299 792 458 m/s can be confirmed by measurement. That is an incorrect statement, as a defined value can be changed only by changing the definition, which is the prerogative of the CGPM, not of measurement.

Perhaps what could be said instead is:

“The techniques described here are therefore no longer needed to measure it; it may be noted, however, that they produce values the same as the defined value to within the present-day accuracy of such measurements.”

Brews ohare (talk) 19:59, 22 July 2010 (UTC)

Wouldn't it though be confirming it's value, measuring it though you know the answer? The same way I can confirm the value of π by comparing the diameter of a wheel to the distance travelled as it rotates once, or confirm the date of Christmas this year by looking on a calendar, but I cannot change either π or the holiday. I did when writing it think it simplifies it somewhat. You might rather be measuring the speed of light to gauge the errors in your time and distance measurements, or to test your methods or equipment. The point is there's no need to measure the speed of light today, as at all you can do is get the existing value and confirm it - or get a different value which indicates a problem in your methods.

And this is the second section you've started titled "Measurement" on this page, so I've appended "again" to make it appear with a different name in edit summaries: otherwise it will be difficult to find on a very long talk page.--JohnBlackburne^words_deeds 20:22, 22 July 2010 (UTC)

--JohnBlackburne^words_deeds 20:22, 22 July 2010 (UTC)

Well, John, I have a lot of trouble understanding the need for convoluted meaning in what is a simple matter. The CGPM chose to define the SI c as a value near the value used at the time so as not to cause too much dislocation. But that wasn't a logical requirement, it was just expediency. And if it turns out one day that the last 3 decimal places in c determine whether the universe obeys quantum gravity or not, the defined value of c won't change. It just won't be confused with the "speed of light" any more, it'll be obvious to all that it is just a conversion factor. So I totally object to "confirming" a definition by means of measurements. It is a Humpty Dumpty usage, and turns the meaning of words upside down. Brews ohare (talk) 23:37, 22 July 2010 (UTC)

The wording at present is awkward in the very least. What are we trying to say here? IMO, we want to say something like "we don't actually measure c any more, but here are some of the methods that have been used in the past." Physchim62 (talk) 00:07, 23 July 2010 (UTC)

Just checking something regarding the views here. Is it only the physical constant c that is fixed and not the result of a measurement by someone in some lab for the speed at which light travels in a vacuum? Thanks. --Bob K31416 (talk) 00:50, 23 July 2010 (UTC)

Well, to measure the speed at which light travels in a vacuum, you need a length standard. If your working in SI units, the speed of light is your length standard, so the measurement becomes rather pointless. You're free to choose another length standard, of course, but it wouldn't be normal to express your results as a "speed of light" in that case: you might be measuring the length of your standard in metres, or measuring the ratio of the speeds of light at two different frequencies, for example. Physchim62 (talk) 01:13, 23 July 2010 (UTC)

Not, I hope, to disagree with Physchim62 but to say things in my own words: To answer your question directly: the "physical constant" is not fixed. (It is to get mired down to explore what that term actually means.) Rather, the standard unit of of speed is fixed. The standard unit is not established by measurement but by fiat by the CGPM. The standard unit of speed can be realized approximately in a lab by creating an approximate vacuum (a "real" vacuum is beyond realization, but the CGPM will help you set up a good approximation). However, the purpose in doing so is not to "measure" the standard; it is to get a measuring stick with which to compare other unknown speeds to the standard. Brews ohare (talk) 02:10, 23 July 2010 (UTC)

Once you have the standard speed set up in your lab, you also can compare two lengths by simply comparing the two times-of-flight of your approximate standard speed along the lengths. Brews ohare (talk) 02:17, 23 July 2010 (UTC)

Another point to remember is that when we say that c is "fixed" by the definition of the metre, we're usuing a sort of shorthand: c is "fixed" as long as relativity, quantum electrodynamics and a whole range of other physical theories are correct. If any of those theories are not correct, then our equations won't add up, although it may take a long time before our measurement techniques are good enough to realize that the equations don't add up. If the equations don't add up, we will have to either introduce a correcting factor and then figure out what that correcting factor actually means physically, or we will have to come up with a whole new theory: such is the stuff of scientific revolutions, as Thomas Kuhn pointed out. But, at the moment, the equations add up to within experimental uncertainty, at least most of the time! Physchim62 (talk) 03:13, 23 July 2010 (UTC)

I think the latest fix is lame. All the techniques still work; they just measure the meter instead of the speed of light. If you start with a meter bar, the measurement in the end gives you a correction to it. The only thing that has changed is what you take as fixed in the algebra. Dicklyon (talk) 03:24, 23 July 2010 (UTC)

Last two sentences of section Increased accuracy and redefinition of the metre

In the section Increased accuracy and redefinition of the metre, the last paragraph is,

The effect of this definition gives the speed of light the exact value 299792458 m/s, which is nearly the same as the value 299792456.2±1.1 m/s obtained in the 1972 experiment. The CGPM chose this value to minimise any change in the length of the metre. As a result, in the SI system of units the speed of light is now a defined constant. Improved experimental techniques do not affect the value of the speed of light in SI units, but do result in a more precise realisation of the SI metre.

Could the last two sentences be trimmed in the following way?

The effect of this definition gives the speed of light the exact value 299792458 m/s, which is nearly the same as the value 299792456.2±1.1 m/s obtained in the 1972 experiment. The CGPM chose this value to minimise any change in the length of the metre. As a result, in the SI system of units the speed of light is now a defined constant. Improved experimental techniques do not affect the value of the speed of light in SI units, but do result in a more precise realisation of the SI metre.

--Bob K31416 (talk) 12:51, 23 July 2010 (UTC)

I'd say that is sticking your neck out. The statement with SI units in it is correct. The statement without it is very broad, and appears to encompass all systems of units. It may be that some of these are not based upon a standard speed, but instead continue to use a standard length. In any event, one can imagine such systems. Brews ohare (talk) 13:56, 23 July 2010 (UTC)

Re "one can imagine such systems" - Could you give an example of an existing system where the proposed change would be incorrect ? Thanks. --Bob K31416 (talk) 15:31, 23 July 2010 (UTC)

I assume by "existing system" you mean one that is used today. In that case, the astronomical system of units, as Brews mentions in the next section. If we're allowing any conceivable system, then it would be any system of units (including SI) from before 1983. Physchim62 (talk) 15:43, 23 July 2010 (UTC)

I don't see the point, but I won't pursue this question because I don't think I could make any progress. Thanks anyhow. --Bob K31416 (talk) 16:57, 23 July 2010 (UTC)

It's metrology-wonkishness, I agree, but there are plenty of subject-wonks at WP:FAC as well ;) Physchim62 (talk) 17:11, 23 July 2010 (UTC)

I don't think it's a matter of wonkishness. It doesn't appear that these points are made in reliable sources. If they were, there wouldn't be any difficulty. I could just read the reliable source. Cheers, --Bob K31416 (talk) 17:29, 23 July 2010 (UTC)

Light doesn't give a damn about what metre means, so whatever the speed of light in itself was in 1982 still is now. It was the value of the speed of light in metres per second which changed status. A. di M. (formerly Army1987) (talk) 18:07, 23 July 2010 (UTC)

And similarly the value of the speed of light in any units that metres per second can be converted to, which changed status? --Bob K31416 (talk) 18:40, 23 July 2010 (UTC)

Only in units that metres per second can be exactly converted to. This includes miles per hours (for example), but excludes astronomical units per day. The value of c in AU/day had to be measured before 1983 and has to be measured now. A. di M. (formerly Army1987) (talk) 20:14, 23 July 2010 (UTC)

The value of c in AU/d wouldn't be exact if the conversion between m and AU isn't exact. Nevertheless it seems that, along with the status change of the value of c in m/s, the status changed for the value of c in AU/d that is obtained when m/s is converted to AU/d (using for example, 1 AU = 149,597,870,691 meters), because the value of c in AU/d would also not be changed by any future measurement of c. Regards, --Bob K31416 (talk) 22:58, 23 July 2010 (UTC)

I don't follow you. The value of c in AU/d is not exact and continues to be measured, and it is 173.144632674(3) where the digit in parentheses is the uncertainty; the length of 1 AU in metres can then be computed (neglecting nitpicking about the metre being supposed to be a unit of proper length) as 299792458×86400/173.144632674(3), and the result is 149597870700(3) m. What the metre has to do with measurements of c in AU/d I can't understand. A. di M. (formerly Army1987) (talk) 01:20, 24 July 2010 (UTC)

I think that you are saying that the speed of light c in units of AU/day is not derived from c=299,792,458 m/s. Is that what you're saying? --Bob K31416 (talk) 10:08, 24 July 2010 (UTC)

What do you mean by "derived from"? Certainly it can be measured without reference to the metre... A. di M. (formerly Army1987) (talk) 10:31, 24 July 2010 (UTC)

By "derived from" I meant c = 299,792,458 x 86,400 / 149,597,870,700 AU/d , using the inexact value for AU in metres that you gave and is given here. I may have to rethink this. What I just wrote, considering your comments, looks kind of circular. I'm open to any helpful comments from you and any reliable sources that you may know of that discusses this. --Bob K31416 (talk) 12:17, 24 July 2010 (UTC)

I don't think there's any fundamental difference between computing the value in AU/day and then converting the result into AU/metre or the other way round; our Astronomical unit article suggests it's the former, but ultimately (AFAICT) is just a matter of conventions. A. di M. (formerly Army1987) (talk) 12:54, 24 July 2010 (UTC)

A di M & BobK: Isn't the underlying issue here whether a change in the phenomenon of light propagation (due some mysterious cause) would change the speed of light in AU/day, but will not change the speed of light in SI units? And for the same reasons, the AU conversion to metres would change, because the metre changed length. Brews ohare (talk) 14:54, 28 July 2010 (UTC)

It depends on which mysterious cause does that. In any event, that's wild speculation. (Also, for the length of the AU to change it would suffice for the Sun to change mass.) A. di M. (formerly Army1987) (talk) 15:47, 28 July 2010 (UTC)

A di M: You have introduced a red herring in mentioning the change in AU due to other reasons. Of course, it is known that the AU does increase with time, and it is speculated that is due to the changing mass of the Sun. However, the subject here is not an error analysis to determine whether an observed effect is due to a change in the travel of light or due to other causes, nor is it a discussion of how large a change in travel of light would have to be to be clearly identified. Nor is it about what coauses the change in the way light travels. Rather, a question of logical issues in principle is at hand here.

I would assume that you would agree that, for example, should it happen that the travel of light doubled in speed, the metre would half, the AU wouldn't change, and the speed of light in AU/day would double, while c in SI units would stay exactly the same. Brews ohare (talk) 16:00, 28 July 2010 (UTC)

Astronomical units

The AU apparently is a measured quantity given by:

$A^{3}={\frac {G\ M_{S}}{k^{2}}}\ ,$

where A = AU, k=0.01720209895 AU^3/2 d^-1 exactly, and d is a day of 86400 s, and M_S is the solar mass, G the gravitational constant. G and M_S are measured quantities, introducing uncertainty in the AU. Its value in metres at the present time is 149597870700 ± 3 m (same source). It increases at a rate of (15 ± 4) cm/yr. The origin of the increase is unclear, but attention is given to trying to determine what is happening to the mass of the Sun.

For the purpose of this article, there are two issues. One is that a system of units based upon a length is not a system based upon a standard unit of speed.

The second issue is whether one can attach any significance to the speed of light as measured in this system of units based upon the uncertainty in the AU. In other words, is the error bar here too big to make such a determination interesting?

What should the article say about these matters? Brews ohare (talk) 14:49, 23 July 2010 (UTC)

I'd say the article hits the main points at the moment, but fails to clearly contrast this approach to the SI units approach as logically separate methods. Brews ohare (talk) 14:54, 23 July 2010 (UTC)

I'm glad you think it hits on the main points, because I think I wrote that paragraph! I'll see if I can add a footnote to cover the points you raise, but I'm not sure it merits more than that. As for the accuracy of the measurement, I think it's impressive (possibly even a little optimistic, but probably order-of-magnitude correct): we couldn't do much better on Earth for the moment, even if we defined a length scale based on the best practical wavelength for laser interferometry. Physchim62 (talk) 16:49, 23 July 2010 (UTC)

I added the Nieto source, and a few words of emphasis. Brews ohare (talk) 14:34, 28 July 2010 (UTC)

Folk relativity withdraw syndrome. Struggling only makes it worse.

Relativity: "light is always propagated in empty space with a definite velocity c" -- Einstein
Wikipedia: "light is always propagated with a definite velocity c".
Does anyone (else) object to eliminating this discrepancy? Before you answer, recall your fifth year of college when you were informed that condensed matter physics is not the domain of special relativity.

Also, this statement needs to be justified with references: "The physical photon is the bare photon plus all sorts of (formally infinite) corrections and this physical photon travels at c".
To the contrary:
A wave packet containing a single dressed photon travels at the group velocity through the medium "In the spirit of Einstein’s original model of light quanta in the vacuum, each dressed photon carries energy ℏω (k) according to (19). Furthermore, one can show that the appearance of the group velocity in the normalization factor in (20) guarantees that a single-photon wave packet, propagating at the group velocity, carries the energy ℏω (k) associated with the carrier wave.", "the dressed photon might be regarded as a relativistic particle with velocity vgr(k)"

"In general terms, a photon always propagates in a medium at less than the speed of light. In the language of modern quantum field theory it can be considered a 'dressed' photon

"The photon in plasma (dressed photon) has rest mass = wp/c.

Photons in matter (Wikipedia) "light that travels through transparent matter does so at a lower speed than c, the speed of light in a vacuum. X-rays, on the other hand, usually have a phase velocity above c ... The factor by which the speed of light is decreased in a material is called the refractive index of the material. In a classical wave picture, the slowing can be explained by the light inducing electric polarization in the matter, the polarized matter radiating new light, and the new light interfering with the original light wave to form a delayed wave. In a particle picture, the slowing can instead be described as a blending of the photon with quantum excitations of the matter (quasi-particles such as phonons and excitons) to form a polariton; this polariton has a nonzero effective mass, which means that it cannot travel at c. Alternatively, photons may be viewed as always traveling at c, even in matter ... a light wave made up of these photons does travel slower than the speed of light. In this view the photons are "bare", and are scattered and phase shifted, while in the view of the preceding paragraph the photons are "dressed" by their interaction with matter, and move without scattering or phase shifting, but at a lower speed."

A bare photon plus the corrections is called a dressed photon, dressed photons do not travel at c.

Next on the agenda, more heresy! The speed of light on Earth (as compared to light in interstellar spare) is slower in January than in July. Physical light always travels slower than c, etc, etc. Unless of course, Einstein was wrong when he said all of these things. —Preceding unsigned comment added by NOrbeck (talk • contribs) 08:05, 24 July 2010 (UTC)

The resolution of this problem is simple. Light as a classical wave travels at c in (the idealised medium of) free space and more slowly in real media. Photons are a quantum entities and it is not valid to apply classical concepts of velocity, trajectory, or position to them, as this leads to paradoxes, problems, and arguments such as we have here.

The answer is never to talk of the speed or path of a photon or photons. Whatever the original intention may be, someone will always imaging the photon as a little ball travelling from A to B down a fixed trajectory in a precise time. Martin Hogbin (talk) 09:13, 24 July 2010 (UTC)

I agree. The article at 100 KB is already long enough, there's no point in adding tons of esoteric discussions which would be better suited for (e.g.) Metre, Interferometry, or Quantum electrodynamics which would only interest (and be understood) by about 1% of the readers of this article. :-) A. di M. (formerly Army1987) (talk) 10:38, 24 July 2010 (UTC)

It is not correct to say that dressed photons do not travel at c. This is only the case when you define dressed photons by taking into account interactions with a medium (which is not Lorentz invariant and thus leads to the dressed photon travelling at a speed different than c). But the phrase "dressed photon" is also used to take into account interactions the electromagnetic field has with the fields of other paticles in vacuum. So, the dressed phton is then simply the ordinary physical photon in vacuum which travels at c. Note that in the Scharnhorst effect the physical photon travels faster than c, which happens because the Casimir vacuum between the two plates is not Lorentz invariant. Count Iblis (talk) 15:36, 24 July 2010 (UTC)

Great, I wholeheartedly agree. In the change log I wrote: "dressed particles do not travel at c, except in a vacuum.". Above I meant to say "dressed photons do not necessarily travel at c". I also agree that we should avoid mentioning photons whenever possible.

This statement from the article is is egregious: "c is the speed at which all massless particles and waves, including light, must travel."

The tone is classic pseudo-relativity, no theory dictates how reality "must" behave. Of all the relativity myths that cause confusion and frustration, this is one of the worse.

I would have been satisfied to append "in a vacuum" to the end of the sentence, but that still would have been a lie. The truth is real physical light never travels at c. I don't think this fact is controversial or contestable. Nevertheless, despite two attempts, the claim is still there.

The article often seems to be intentionally misleading the reader. Is this lying to children? I wonder, what do people really believe?

Prove your faith. The questions are:
Does real physical light travel travel at c? (No)
Is c is the limiting speed approached, but never reached, in partial vacuums. (Yes)
Is the local instantaneous speed isotropic and invariant in inertial frames? (Yes!)
Does real physical light actually move from point a to point b at a constant speed? (No)
Is the real physical (coordinate) speed slowed by the everpresence of matter; so that even in a vacuum, the speed of light is still not c. (Yes)
Is the speed of light slowed by proximity to other light? (Yes)
When viewed from outside, does the speed of light in a room speed up when the lights are dimmed? (Yes)
Is the speed of light equal to c in a universe with no matter? (No)
Is the speed of light equal to c in a universe with no matter and no light? (Yes!)

Please vote, but let's not waste effort debating this. I'm just trying to see to what extent my expertise will be useful here. NOrbeck (talk) 03:37, 25 July 2010 (UTC)

What is real physical light?

What is the 'instantaneous' speed?

What is the 'real physical' speed?

How do you know what the speed is in a universe with no matter?

Is this philosophical speculation appropriate for this article or even this talk page? Martin Hogbin (talk) 09:08, 25 July 2010 (UTC)

1. Special relativity dictates that all real massless particles and waves have light-like momentum, which is an obvious consequence of the definition of massless; thus, their velocity (if defined as c²p/m) does equal c in magnitude. (It doesn't dictate that light must be massless, though.)
2. It is true that "travel at" isn't supposed to mean "have a well-defined position at every single instant", but such nuances are better suited to such articles as photon, quantum electrodynamics, etc. than here: c is the fundamental conversion factor of special relativity, and special relativity doesn't depend on quantum mechanics. (Matter of fact, it was formulated before quantum mechanics).
3. The answer to the last of Hogbin's questions above is "No, it isn't." A. di M. (formerly Army1987) (talk) 10:50, 25 July 2010 (UTC)

Re Martin Hogbin's remark "Is this philosophical speculation appropriate for this article or even this talk page?" - I would add to this remark that Wikipedia is an encyclopedia which summarizes existing published information, and is not meant to be a source of information that has not already been published. I think that the article should only include material that appears in reliable sources and this should be demonstrated by citing them. If there is a question as to whether a cited reliable source contains the material, then the relevant excerpt from the reliable source should be given here on this talk page where editors can decide whether the reliable source supports the material.

So with this in mind, what reliable source or sources are we trying to summarize here? On the other hand, is there a reliable source for the material in the article that is being questioned? If not, perhaps at least a {{Citation needed}} template should be placed. I'll leave that for the other editors here to decide. --Bob K31416 (talk) 12:10, 25 July 2010 (UTC)

My point was not so much, 'Are there any sources which support these views?', but 'Do the questions asked have any meaning?. Martin Hogbin (talk) 13:37, 25 July 2010 (UTC)

BobK has the right approach to this matter, especially for this article that has so many editors with intransigent positions of their own. And the questions raised, whether meaningful to Martin or not, are certainly meaningful to a naive reader and should be dealt with. Brews ohare (talk) 13:29, 28 July 2010 (UTC)

"most used reference length scale"

The beginning of the following sentence seems awkward and unnecessary:

"Since the most used reference length scale in modern experiments (the SI metre) is determined by the speed of light, the value of c is fixed when measured in metres per second."

Perhaps it should be changed to:

"Since the most used reference length scale in modern experiments (the SI metre) is determined by the speed of light, the value of c is fixed when measured in metres per second."

--Bob K31416 (talk) 20:29, 24 July 2010 (UTC)

Agree. Physchim62 (talk) 21:14, 24 July 2010 (UTC)

 Done. A. di M. (formerly Army1987) (talk) 10:54, 25 July 2010 (UTC)

Excessive detail in "Increased accuracy and redefinition of the metre" section

IMHO, the "Increased accuracy and redefinition of the metre" section now has an excessive amount of detail on the realization of the metre. This article is about the speed of light, not the metre. The article as it is, is already very long for its limited scope. The "see also" link to metre is there for a reason.

The info relevant to this article is that the metre was redefined and that the with the new definition the value of c in SI (based) units is exact. Any more information about the mise en practique of the realisation of the metre, is fluff to this article, better discussed in the metre article.

Effectively, the part starting with "The difficulty with..." and ending with mise en practique can be omitted here. (With some additional rephrasing of the surrounding text to reflect this rephrasing.) At least that is my humble opinion. TimothyRias (talk) 09:46, 27 July 2010 (UTC)

I agree.Martin Hogbin (talk) 14:21, 27 July 2010 (UTC)

I agree that it could use some trimming, but perhaps we need to be careful. For example, the part that you suggest omitting contains the sentence "In 1975, the 15th Conférence Générale des Poids et Mesures (CGPM) recommended using 299792458 metres per second for "the speed of propagation of electromagnetic waves in vacuum".", which seems essential, unless it's somewhere else in the article that I didn't notice. Maybe the section title is misleading, since it might appear that the section is mainly about the metre, rather than mainly about the increased accuracy of the speed of light, if I understand the purpose of the section correctly. --Bob K31416 (talk) 14:53, 27 July 2010 (UTC)

The 1975 decision by the 15th CGPM, simply was to establish what value should be accepted for the 1972 result in face of the ambiguity of the definition of the metre. The decision is again more a matter of the definition of the metre rather than an increase in the accuracy of the knowledge of the speed of light. This is also that the 1975 decision was not included in the table. But I do agree it is not simply a matter of deleting the suggested part the surrounding material material will need some modification as well. TimothyRias (talk) 16:38, 27 July 2010 (UTC)

I agree with Timothy and Martin.—Finell 06:39, 28 July 2010 (UTC)

This subsection is the only place in the article Speed of light that attempts to make clear the issues surrounding the introduction of a defined value for c. In view of the tremendous blather that has occurred on this Talk page for years, mainly in an attempt to install a quasi-religious belief in a magic number, it is evident that this topic requires explanation here. It should not be removed on the basis that it "belongs elsewhere"; it belongs here because it relates to the meaning of the SI version of c. It belongs here because it relates to the switch to 'time-of-flight' units which many editors on this page have yet to appreciate. It belongs here to clarify the distinction between a phenomenon, the measure of that phenomenon, and the measurement of that phenomenon. If this section is to be shortened, another entire article should be introduced on this subject and referred to here as the Main article on this subject. Brews ohare (talk) 13:40, 28 July 2010 (UTC)

Longer is not always clearer. By straying from the subject in to excessive detail (like the fact that the BIPM keeps a list of recommended sources for realising the metre) the main issues are obfuscated rather than clearified. As the section is written now the fact that the metre was redefined comes as an after thought, rather than as the main message of the relevant paragraph.TimothyRias (talk) 14:09, 28 July 2010 (UTC)

One can quarrel over details of the wording. For example, what is the purpose of mentioning a mise en pratique? Is it just useless baggage? Or, does it provide a useful contrast with the earlier method of keeping a hard-to-access specimen metre under standard conditions in Paris? Or, the specification of a more available single source, a krypton discharge lamp with various serous problems? Instead, the modern mise en pratique is very accessible, is continually updated and refined, provides numerous sources of standards, and provides the associated uncertainties (although the notion of uncertainties may be an anathema to some). This development of context has some interest, I'd say, and points out that although c has a defined value, its realization is only approximate. Brews ohare (talk) 14:32, 28 July 2010 (UTC)

I'd is not the realisation of c that is approximate, but the realization of the metre. As interesting as this subject is. It is about the metre and not the speed of light. TimothyRias (talk) 15:34, 28 July 2010 (UTC)

A review of the wording doesn't require drastic changes in this section. This subsection is the only place in the article Speed of light that attempts to make clear the issues surrounding the introduction of a defined value for c. Brews ohare (talk) 15:16, 28 July 2010 (UTC)

Timothy: As I have pointed out, the subject of the defined value for the speed of light is contingent upon the change of definition of the metre, as you well know. To say the topic is therefore about the metre and not about the speed of light is disingenuous, and suggests an agenda is being forced for reasons unstated. Brews ohare (talk) 16:10, 28 July 2010 (UTC)

Brews, my only agenda here is to make this article as clear as possible without unnecessary digressions. (btw, your suggestion that I'm trying to push some other agenda is: 1)extremely insulting 2)a violation of WP:AGF 3)a violation of your general probation, I suggest you try harder to keep this discussion civil.) Your saying that the mention of the the current mise en practique is a useful contrast against the practice for the realisation of the metre before that, but since the practice of the realisation of the metre before 1983 is also not discussed in this article (because it is about the metre not c), the sentecne about the mise en practique for reasilisation of the current definition is also out of place.

Again, my opinion here is that the only points truely relevant to this article are, that the metre was redefined and that the result of this redefinition was that the value of c in SI units is now a defined exact quantity. (Like the freezing point of water is a defined quantity on the Celsius scale (or at least was, I'm not sure how degrees centigrade are currently defined)).TimothyRias (talk) 17:10, 28 July 2010 (UTC)

Indeed. The maximum density of water at 1 atm used to be 1 kg/L exactly by definition of the litre, but an extensive discussion of that doesn't belong to the water article; the same applies to whichever other way to define a unit of measurement. How c would be different is beyond my grasp. A. di M. (formerly Army1987) (talk) 17:31, 28 July 2010 (UTC)

A di M: You question: “How c would be different is beyond my grasp.” Now, what is c in this observation? Is c the value 299, 792, 457.3 m/s? I'd guess you have little trouble imagining this number could be chosen to be any number whatsoever, although making it 1 m/s might cause some expense as all existing metre sticks went to the dump. On the other hand, is c the maximum speed of information transfer viewed as a limit, without particular regard to its expression in units? I can see where you'd have trouble imagining it to depend upon a choice of units.

What Timothy and yourself appear to me to overlook is that these two things are readily confused by a reader, and that a discussion of the switch to "time-of-flight" units in this subsection is a necessary attempt to straighten these matters out for the reader. Brews ohare (talk) 17:51, 28 July 2010 (UTC)

The same thing applies with any other way of defining anything. The temperature at which liquid water, water ice and water steam can coexist doesn't depend upon a choice of units, but it exactly equals 273.15 K by definition of the kelvin; but the water article spends no more than a paragraph to discuss this point, which is (rightly) more fully discussed at kelvin. In which way the speed of light differs from the triple point of water, and the metre differs from the kelvin? A. di M. (formerly Army1987) (talk) 18:20, 28 July 2010 (UTC)

BTW, I would rather remove the mention of mise en pratique, so that it just read "Based on this recommendation, the 17th CGPM in 1983 redefined the metre as ..."; more detail should go at Metre (which right now doesn't even make clear that there are several radiations used for that purpose). A. di M. (formerly Army1987) (talk) 18:20, 28 July 2010 (UTC)

Can you explain to me why the subject of how the "time-of-flight" definition of the metre has been implemented is irrelevant and a gross distraction here? It is plain that the implementation of this change underlies the use of the defined value for the speed of light, and so is pertinent to a reader's normal curiosity about how this change (that so fundamentally alters the system of units) is accomplished. It also links the reader to the mise en pratique which is a valuable source for further explanation. It also makes clear that the realization of the standard speed of light is subject to uncertainty by virtue of the uncertainty in the source radiations, and the notion that such uncertainty is present is a valuable antidote to the mystique associated with an "exact" speed of light. Can you respond to each of these points? I simply don't see that there is either (i) any pressure due to lack of space in the article, (ii) any distraction from the article as it is a subsection clearly devoted to this topic. Brews ohare (talk) 18:34, 28 July 2010 (UTC)

Any change in the definition of a unit does the same, but that is normally discussed in the article about the unit, not in that about the quantity used to measure it. The article water doesn't discuss arcane details of the redefinition of the kelvin which are discussed at Kelvin, the article Isotopes of caesium doesn't discuss arcane details of the redefinition of the second which are discussed at Second, so why should arcane details of the redefinition of the metre be discussed here rather than at Metre? A. di M. (formerly Army1987) (talk) 19:29, 28 July 2010 (UTC)

A di M: It is not an "arcane detail about the definition of the metre" that changes the speed of light from a measured quantity with an associated experimental error bar to an exact number beyond the reach of measurement. The effect of the definition of the metre upon the value for the speed of light is profound, affects methodology, measurement practice, and reduces the standard units of time and distance to a standard for time alone. There is no parallel with the other units you mention. Brews ohare (talk) 02:23, 29 July 2010 (UTC)

Brews, that is simply not true for reasons repeated to you a thousand times. If you want to be nitpicky then you can say the decision changes the standard units of time and distance to standard units of time and velocity, which is equivalent. This is analagous to what happend with the introduction of the international prototype for the kilo, (change from fundamental density to a fundamental mass), or the type of change that will happen with possible future redefinitions of the Coulomb, Ampere and kilogram. Nor the the changes is methodology and measurement practice as profound as you claim. TimothyRias (talk) 08:24, 29 July 2010 (UTC)

Indeed, in what way does it change methodology and measurement practice at all? Before 1983, precise lengths were measured by interferometry; since 1983, precise lengths are measured by interferometry. Big change, huh? Physchim62 (talk) 08:36, 29 July 2010 (UTC)

I don't object to stating that "changes the speed of light from a measured quantity with an associated experimental error bar to an exact number beyond the reach of measurement"; I object to use several paragraphs to discuss that including obscure details about interferometry and metrology which will interest about 1% of the readers (who could always read them at interferometry and metre anyway). That'd be akin to expand the section Water#Water as a scientific standard by a factor of two by discussing how densities and temperatures are measured and what the mise in pratique of the kelvin and the kilogram are. A. di M. (formerly Army1987) (talk) 10:49, 29 July 2010 (UTC)

Well folks, I'd say what all this amounts to is that you want to compress this section. Not for reasons of clarity or usefulness of the article, but because you don't want this material here. As space is hardly at a premium, and as the topic is difficult, I don't see any advantage in curtailing the discussion other than to please you all. Why that would please you is unclear. Brews ohare (talk) 15:20, 29 July 2010 (UTC) The real issue with this section is that it is poorly organized and so appears to wander about. It is not its content that is the problem. Brews ohare (talk) 15:34, 29 July 2010 (UTC)

(edit conflict)The reason is the article is very long, too long to be easily read by many readers, so it is indeed about clarity. See e.g. WP:LENGTH, a link to which appears whenever the article is edited, especially WP:SIZERULE.--JohnBlackburne^words_deeds 15:39, 29 July 2010 (UTC)

This argument is sounding a lot like the one that led to the arbitration. As I said above, rehashing this argument is not going to persuade those who hold the opposite view, is not contributing to the article, and is a waste of time and energy. Enough is enough.—Finell 16:28, 29 July 2010 (UTC)

Finell: You exhibit an unfortunate exasperation with the proceedings here. In fact, unlike the previous episode that you unnecessarily draw attention to, there has been a largely constructive revision of the article in this round of editing. I am confident that improvements will continue to be made. Please do not despair, and please do not resort to gloom and doom pronouncements suggesting that the dark ages are at hand. Brews ohare (talk) 16:44, 29 July 2010 (UTC)

Section title

Suggest changing the title of the section from

Increased accuracy and redefinition of the metre

to

Second half of the 20th century.

--Bob K31416 (talk) 15:21, 28 July 2010 (UTC)

Object, since the first is much more descriptive and to the point. TimothyRias (talk) 15:35, 28 July 2010 (UTC)

Thanks. BTW, does "Increased accuracy" refer to c, the metre, or both? --Bob K31416 (talk) 19:23, 28 July 2010 (UTC)

"increased accuracy is supposed to refer to c. (At least that was my intention when a chose the heading.) TimothyRias (talk) 22:46, 28 July 2010 (UTC)

Maybe add "of c"? --Bob K31416 (talk) 01:04, 29 July 2010 (UTC)

Gravity and gravitational waves

The sentence in the introduction about the relation of the speed of light to gravity and gravitational waves appears to go beyond established ideas on this subject to adopt a particular viewpoint. See Fabio Cardone, Roberto Mignani. I believe this sentence distorts the record. Brews ohare (talk) 15:27, 28 July 2010 (UTC)

"Established ideas"? You mean like "... u_grav >= 2.5 10^10 c, in astonishing agreement with the astronomical estimate by van Flandern"? Duh. DVdm (talk) 15:38, 28 July 2010 (UTC)

It says c is believed to be the speed of gravitational waves. That much is true. A. di M. (formerly Army1987) (talk) 15:40, 28 July 2010 (UTC)

See also: Part 3: On the speed of gravity controversy. Brews ohare (talk) 15:47, 28 July 2010 (UTC) BTW, a belief is something held without adequate justification, and probably signifies that a different belief is entertained by some. WP is supposed to be even-handed in such matters, and not present only one side. Brews ohare (talk) 15:49, 28 July 2010 (UTC)

Yes, but then we have to give a detailed overview of this issue to present the correct context. If we just mention that there are alternative points of view without explaining in any depth what the standard point of view is, that would amount to giving a huge weight to a contrarian point of view. Count Iblis (talk) 15:54, 28 July 2010 (UTC)

So, where some honest difference of opinion occurs among experts, a reader of WP should assume that one view will be presented and others ignored with a warning flag (if you're lucky) that the stated position is "believed" (by the editors in charge of the WP article at this moment in time)? Brews ohare (talk) 16:05, 28 July 2010 (UTC)

A preferable approach would be to state the preferred position with some sources, and say that there is some controversy over the point and cite those sources. If the matter is so subtle that only an extended discussion will cover the matter, and I believe that is the case, the reader is left with places to look. That methodology is preferable to introducing the tactic of "beliefs". Lord help us if the "belief" approach makes it to your favorite topic of global warming. Brews ohare (talk) 16:17, 28 July 2010 (UTC)

Lord help us if Van Flandern is considered to be an "expert" in this matter and cited as a source. DVdm (talk) 16:31, 28 July 2010 (UTC)

It appears that Schafer and Brugmann think the subject is up for grabs experimentally, and Cardone & Mignani feel the same way. Brews ohare (talk) 16:48, 28 July 2010 (UTC)

General relativity is by far the most widely accepted theory of gravitation among physicists (to the point that discussing alternatives would be WP:UNDUE except in very specialized articles), and it predicts that gravity travels at c. The reason for saying "is believed to be" rather than just "is" is that no-one has yet detected gravitational waves, though there's a lot of indirect evidence that they actually exist. A. di M. (formerly Army1987) (talk) 16:54, 28 July 2010 (UTC)

If the situation is ‘theoretically’ as stated, the position should be labeled that way. A statement that is open to the interpretation that it has experimental support where there is none is unacceptable. And to claim theoretical support is a ‘belief’ is a smack-down of theoretical arguments, which are after all much better established than some credo. Brews ohare (talk) 16:59, 28 July 2010 (UTC)

I could only read Schafer and Brugmann book (presumably because my IP address is different) and they make clear that the issue is purely one of experimental verification. They point out that Kopeikin's argument that you could measure the speed of gravity using the Shapiro delay effects when Jupiter passes in the direction of a quasar is flawed, as the effect does not depend on c_g. Count Iblis (talk) 17:05, 28 July 2010 (UTC)

That is exactly correct, and applies to the other source as well. The point is that the WP article should make it clear that the predictions of theory in this regard are unsupported, and that this is not a question of general "belief", which is a most unfortunate choice of wording. There are some apparently contradictory experimental observations, whose apparent conflict with theory has yet to be resolved, and some observations that appeared to support the theory that have been found to be erroneous calculations.

It is a bit mystifying to me why it is thought that what is proposed here is an alternative theory. Perhaps I phrased things incorrectly at the outset. Brews ohare (talk) 17:31, 28 July 2010 (UTC)

While on this subject, do we need to say the speed of "gravity and of gravitational waves"? Just "gravitational waves" would be simpler and maybe more accureate imho (speaking with the store of knowledge available to one who is just progressing from year 1 to year 2). Abtract (talk) 17:55, 28 July 2010 (UTC).

It appears from the sources that a distinction is drawn; Fabio Cardone, Roberto Mignani p. 222 say: “Gravitational radiation does admit retarded potential solutions of electromagnetic type. It therefore describes propagation of the perturbations of a static or near static gravitational potential field. ...On the contrary, how much is the propagation speed of the gravitational force? It answers the question of how much time a target body will take to respond to the acceleration of the source mass.” Brews ohare (talk) 18:24, 28 July 2010 (UTC)

For me, the changes by A di M fix the problem. Brews ohare (talk) 18:46, 28 July 2010 (UTC)

minimize the change in length of the metre

this change by MC Price has introduced the sentence:

"The CGPM chose this value to minimise any change in the length of the metre."

This last sentence is a bit obscure, as it leads to the natural query: Why didn't they make the change in the metre zero? That would be about as minimal a minimum as one could get.

The missed point is that metre was known only to be within a certain range of values (stated as 299,792,456.2±1.1 m/s), and the issue was to choose a value, ostensibly within this range. The chosen value 299,792,458 m/s actually appears to be outside this range (the upper limit is 299, 792, 457.3 m/s), leaving the reader wondering how a "minimal" change criterion could lead to a value outside the measurement error. Brews ohare (talk) 17:24, 28 July 2010 (UTC)

Mmmm, I wondered about that too. Abtract (talk) 17:32, 28 July 2010 (UTC)

The answer is given in a footnote. If you take the definition of the metre to apply to the weighted mid-point of the krypton line, NIST found a speed of light of 299,792,456.2(11) m/s; if instead you take the definition to be based on the maximum-intensity point of the line, the speed of light would be 299,792,458.7(11) m/s. The CGPM decided chose a value in between the two, but nearer to the maximum-intensity point. The quoted uncertainties are given as one estimated standard deviation, so you have to double them (at least) to get a 95% confidence range. Physchim62 (talk) 17:57, 28 July 2010 (UTC)

This explanation is a bit unsatisfactory inasmuch as the CGPM was at liberty to choose any definition, but chose this one. Why? Or, is it simply: Why not? In other words, there is no "minimal change" criterion, it is just six of one, half a dozen of the other. Brews ohare (talk) 18:12, 28 July 2010 (UTC)

Of course, given that there were an infinite number of idiotic redefinitions available to the CGPM, one wonders how they managed to chose an non-idiotic redefinition, does one? Physchim62 (talk) 02:05, 29 July 2010 (UTC)

The original choice of the 299,792,458 in 1975 is at http://www.bipm.org/en/CGPM/db/15/2/ and it seems to suggest that several such measurements were taken in account; does anyone have access to the references cited in it? (In any event, I have removed the "minimise any change" wording.) A. di M. (formerly Army1987) (talk) 19:36, 28 July 2010 (UTC)

I doubt those two references would shed much light on the matter in any case. Physchim62 (talk) 02:05, 29 July 2010 (UTC)

Physchim62: I wasn't suggesting that CGPM made a random choice between all possible options, but that they had a number of equally practical choices and so could just flip a coin. The removal of the wording about minimizing change fixes things anyway. Brews ohare (talk) 03:05, 29 July 2010 (UTC)

Indeed, I managed to get a PDF of the Metrologia article (good ol' ssh, wget and scp!) and it just has the text of the resolutions. Anyway, I'm trying to say in the article that several such measurements were available in 1975. A. di M. (formerly Army1987) (talk) 11:05, 29 July 2010 (UTC)

"To resolve the ambiguity in the definition of the metre"

In the section Increased accuracy and redefinition of the metre, suggest deleting the first part of the first sentence of the third paragraph:

To resolve the ambiguity in the definition of the metre, t The 17th CGPM in 1983 decided on a new definition of the metre: "the length of the path travelled by light in vacuum during a time interval of 1⁄299792458 of a second".^[120]

120. "Resolution 1 of the 17th CGPM". Conférence Générale des Poids et Mesures. BIPM. 1983. Retrieved 2009-08-23.

The reference doesn't appear to list that as a reason. --Bob K31416 (talk) 01:46, 29 July 2010 (UTC)

It's not an "ambiguity", its an "uncertainty in the realization" (first reason given in the resolution). Physchim62 (talk) 02:07, 29 July 2010 (UTC)

It's an "insufficiently precise" realization, a wording that doesn't suggest the problem is cured, just ameliorated. Brews ohare (talk) 03:09, 29 July 2010 (UTC)

Of course the "problem" isn't "cured"! You can never have an exact realization of a unit because, by definition, a realization relies on a real-world measurement. The new definition gives lower uncertainties than the old one, I thought we were agreed on all that. Physchim62 (talk) 03:51, 29 July 2010 (UTC)

What would you like to see done with the phrase in question? --Bob K31416 (talk) 14:39, 29 July 2010 (UTC)

BobK: Certainly the words struck out are worthy of deletion. However, their role is to suggest a motivation for the change, and striking them out leaves no motivation. So instead, one could place:

“To obtain a more precise realization of the metre, the 17th CGPM in 1983 decided on a new definition: "the length of the path travelled by light in vacuum during a time interval of 1⁄299792458 of a second".^[120]”

Of course, that raises the immediate question: Why did that result in a more precise realization? The answer isn't addressed at this point in the subsection. A rearrangement of the text in this subsection to deal with this question directly would be desirable. At the moment this subsection is disjoint and its object (the answer to this question) is not made its focal point. Brews ohare (talk) 15:32, 29 July 2010 (UTC) The new wording simply avoids this question and its answer. Brews ohare (talk) 15:42, 29 July 2010 (UTC)

Update: The beginning of the 3rd paragraph has been changed by another editor to:

"In 1975, considering that similar measurements of c agreed with each other and their uncertainty was comparable to that in the realization of the metre, the 15th Conférence Générale des Poids et Mesures (CGPM) recommended using 299792458 m/s for "the speed of propagation of electromagnetic waves in vacuum". In 1983, the 17th CGPM decided on a new definition of the metre: "the length of the path travelled by light in vacuum during a time interval of 1⁄299792458 of a second"."

so that the new version of the subject sentence in the 3rd paragraph seems OK now.[2] --Bob K31416 (talk) 16:20, 29 July 2010 (UTC)

BobK: Please see my suggestions in the following subsection of this page. Brews ohare (talk) 16:46, 29 July 2010 (UTC)

Thanks. I looked and commented, albeit on another editor's change, which I thought addressed your main point. Things are moving kind of fast and furious here! I'll try to keep up. : ) --Bob K31416 (talk) 17:13, 29 July 2010 (UTC)

Increased accuracy and redefinition of the metre

The subsection Speed of light#Increased accuracy and redefinition of the metre now has a discussion of increased accuracy in measuring the speed of light, which is related to the title of the subsection on "increased accuracy". It also has a paragraph on the redefinition of the metre. What is does not have is any attempt to connect these two matters. That is, it does not answer the focal question: Why did this change result in an “increased precision in the realization of the metre”, which was the espoused purpose of the change in definition as stated by the CGPM.

In an unseemly haste to rid this section of anything about the metre as being somehow too far afield in an article on the speed of light, this subsection whose entire motivation is to connect these two explains nothing, and has become a simple laundry list. Brews ohare (talk) 15:51, 29 July 2010 (UTC)

The explanation of the increased precision in realization has two parts:

First, the replacement of the krypton discharge lamp as a source. The subsection does detail some of the failings of this source related to line shape. In addition it lacked sufficient coherence requiring a two-step measurement. This source was replaced not with a single better laser source but with a methodology that allowed a variety of sources, anticipating that the "best" source would evolve over time and would differ depending upon the measurement length needed for a particular task. Clearly, these steps improved the precision of realization. They did not, however, require a change in the definition of the metre.

Second: The change in definition of the metre to a "time-of-flight" definition. The purpose behind this change was that the introduction of many sources required a methodology for comparison. The wavelength of the sources could be compared, but due to errors like the failure to obtain a plane wave and errors of compensation in balancing the two paths in an interferometer, using interferometry to measure wavelengths was error prone. On the other hand the measurement of frequencies had become very good. As all believed that frequency and wavelength were related by the speed of propagation, it was evident that a comparison of the frequencies of sources was tantamount to a comparison of wavelengths, but more accurate. The only issue was to insure that the speed of propagation was the same in all such comparisons, which led to the adoption of the speed of light in vacuum as the standard of speed of propagation. Here the improvement in realization of the metre was related to flexibility in choice of source to match the measurement requirements at hand.

The remaining uncertainties in the realization of the metre were now (i) the uncertainties in the radiations, now tabulated and kept current in the mis en pratique; (ii) the uncertainties in the realization of "vacuum" (or, equivalently, uncertainties in the corrections implemented to account for the medium employed) and (iii) the uncertainty in interferometry in establishing a wavelength. This last is no better than use of a better source and the old definition of the metre as a number of wavelengths of a specific transition. However, the flexibility in choice of source and the ability to let the standard evolve with improved precision of the sources are the determining factors in making the change in definition.

Something along these lines is what is needed in this section. Brews ohare (talk) 16:35, 29 July 2010 (UTC)

I've added " in order to allow a more precise realization of it" to the sentence about the 1983 redefinition. A. di M. (formerly Army1987) (talk) 16:47, 29 July 2010 (UTC)

That's OK, but how about tweaking to this version, in order to explicitly mention the motivation of CGPM, which is what was stated in the reference?

In 1983, because the definition of the metre did not allow a sufficiently precise realization of it for all requirements, the 17th CGPM decided on a new definition: "the length of the path travelled by light in vacuum during a time interval of 1⁄299792458 of a second".

--Bob K31416 (talk) 17:01, 29 July 2010 (UTC)

Yes; I'm going to do that. A. di M. (formerly Army1987) (talk) 17:16, 29 July 2010 (UTC)

I have implemented some of the above suggestions of mine in the subsection. I hope that you all might agree with these changes, or some minor modifications of them. Brews ohare (talk) 18:29, 29 July 2010 (UTC)

Your changes seem to me completely unnecessary, as they say nothing about the speed of light, only about the metre, probably in far more detail than's appropriate even for that article, and are largely unsourced. As I noted above the article is already very long, so any extra content should only be added if closely related to the topic.--JohnBlackburne^words_deeds 19:33, 29 July 2010 (UTC)

I agree. Perhaps a separate article on the 1983 redefinition of the metre is the answer. Martin Hogbin (talk) 20:21, 29 July 2010 (UTC)

I agree, the detail brews is trying to add here should be added on the metre page. Here the discussion of this matter should be brief at the very most a couple of paragraphs in total, and refer to the metre page for further details on how the change result in a more precise realization of the metre, because that question is completely off-topic in this article. The only thing relevant here is the consequence of the change for the speed of light. To martin:maybe, but I'd vote for first trying to add some discussion of this to the metre page.TimothyRias (talk) 20:25, 29 July 2010 (UTC)

John, Martin and Timothy: Do we agree on the historical facts? I'd say they include these items: (i) There was a change in the definition of the metre (ii) This change resulted in the switch to a defined value for the speed of light. Then, given the foregoing, the change in the definition of the metre is pertinent to this article on the speed of light. One could merely note that the change occurred and was pertinent. However, that approach clearly leaves dangling the question of why that change was made. There is no doubt that this switch appears perplexing upon first exposure. (For example,the New Scientist stated “Despite powerful scientific arguments, the idea of expressing the metre as a fraction of a second may still seem a bit bizarre. ...do not expect to get away with complaining to your local timber merchant that he has been supplying you with wood that is 10^-9 seconds too short.” This reaction is why the discussion ought cover these details. Brews ohare (talk) 20:32, 29 July 2010 (UTC)

The facts were already in the article in more than enough detail, in that section and elsewhere in the article. In particular it was clear why the decision was made. Your changes have confused the matter by introducing irrelevant detail, almost entirely unsourced, that adds nothing to our understanding of the speed of light.--JohnBlackburne^words_deeds 20:49, 29 July 2010 (UTC)

My latest revision (copied and pasted from Bob K's proposal above) did address "the question of why that change was made". Please re-read it more carefully. What the coherence length of the krypton gas discharge is and similar stuff are way beyond the scope of this article. (Also, the source you cited about the coherence length does not mention at all the definitions of the metre and their realization and the Resolution 1 of the 17th meeting of the CGPM doesn't mention coherence lengths at all, so all that smells like WP:OR to me. A more detailed discussion of the redefinition of the metre should go at Metre, but even there please stick to what the CGPM actually said.) A. di M. (formerly Army1987) (talk) 21:57, 29 July 2010 (UTC)

I'd add to this that links to the decisions of the CGPM are pretty much mandatory, and the link to the 2002 mis en pratique provides a good summary. These documents cry out for some additional explanation such as that given here. The espoused purpose of the change in definition as stated by the CGPM is to improve the precision of realization of the metre, and that seems to make the metre a central issue in understanding what happened to the speed of light. Brews ohare (talk) 20:47, 29 July 2010 (UTC)

John, perhaps you could point out to me what specifically is said in the article that "made clear why the decision was made". If that assessment is accurate, I'd reconsider. Brews ohare (talk) 20:55, 29 July 2010 (UTC)

That section before your changes was clear to me. In addition the third paragraph of the lead sets out the facts of the definition, while the section immediately after it goes into more detail of the constancy of lightspeed.--JohnBlackburne^words_deeds 21:03, 29 July 2010 (UTC)

The section before changes, as I explained above, was a laundry list making no connections between the items listed. The facts of the definition are not at issue here. The constancy of light also is not at issue. None of these items at all supports the claim that the reasons for the decision have been made clear elsewhere in the article. Brews ohare (talk) 21:49, 29 July 2010 (UTC)

A di M: I doubt that you really disbelieve the coherence length difficulties, clearly implicated by the definition of the coherence length. I'll dig up sources from interferometry texts if you like. My intention however, was simply to indicate that the linewidth problem was only the tip of the iceberg. There is no need that the CGPM specifically state a coherence length issue. Brews ohare (talk) 22:44, 29 July 2010 (UTC)

The linewidth problem (strictly, the asymmetry of the line) was not "the tip of the iceberg", it was a fundamental problem with the pre-1983 definition of the metre. All the rest is secondary. Physchim62 (talk) 22:48, 29 July 2010 (UTC)

A di M: Not to be argumentative, but to describe my problems with your revision, here is my synopsis of it. The first two paragraphs describe difficulties in obtaining an accurate metre. That's fine, but the implication is only that some new method might be better. The last paragraph then makes a jump to the definition of the metre and the resulting defined value of the speed of light. It states that the motivation for the switch is that the old definition was insufficiently precise, but there is no clue here as to why the introduction of the new definition has any effect upon this problem. It is left to the reader to connect the dots in this non-sequitor. That is my difficulty with this version. Brews ohare (talk) 22:58, 29 July 2010 (UTC)

"Because the definition of the metre did not allow a sufficiently precise realization of it for all requirements, the 17th CGPM decided on a new definition" makes it quite obvious that the new definition allows for a more precise realization. If I told you that I replaced the strings I had on my guitar because they were too thick for my tastes, wouldn't you be able to understand that the new ones are thinner than the old ones, even though you might never have played a guitar? A. di M. (formerly Army1987) (talk) 23:18, 29 July 2010 (UTC)

BTW, the logical fix following the description of the first two paragraphs is simply to switch to a different transition that has a narrower and more symmetric line, or simply to state specifically what position on the Kr line to use in the measurement, instead of leaving it uncertain. These solutions were indeed considered, and yet the change in definition was made regardless of these more obvious solutions. Why? That has to be answered. Brews ohare (talk) 23:03, 29 July 2010 (UTC)

That's what they did. Before 1983 the metre was 1,650,763.73 krypton-86 transition wavelengths, after 1983 it is 9,192,631,770⁄299,792,458 caesium-133 transition wavelengths. A. di M. (formerly Army1987) (talk) 23:18, 29 July 2010 (UTC)

Brews, we have repeatedly told you "why", both last summer and this summer, yet you refuse to listen: "This way of defining the meter has proven to be particularly robust, since unlike a definition based on a standard such as the krypton lamp, length measurement can be continuously improved without resorting to a new definition." Physchim62 (talk) 23:41, 29 July 2010 (UTC)

A di M: Is your reply above to be interpreted as saying the responsibility of the article is to say what CGPM did, and not to provide any of the reasons behind it? That's what they did, period? Brews ohare (talk) 00:32, 30 July 2010 (UTC)

Are you dumb or what? Again, I don't oppose to stating in the article that the definition was changed in order to allow a more precise definition of the metre. What I oppose is giving too much detail about that, when 1) this article is already 100 KB long, 2) such detail would be more on-topic in the Metre article which right now is 26 KB long, and 3) some of the details were not even mentioned by the Resolution 1 of the 17th meeting of the CGPM. (It might well be obvious to you that the short coherence length was one of the problems with the old definition, but if it's not obvious to "any educated person without specialist knowledge" you need a source explicitly stating that. BTW, it's not obvious to me either: if that one was the issue, could you just use the krypton lamp to measure a length much smaller than the coherence length e.g. 2 cm and then measure the ratio of that length and a longer length by whatever means is the most precise?) A. di M. (formerly Army1987) (talk) 10:09, 30 July 2010 (UTC)

Sorry, A di M, I didn't get your point, which was that the Cs line was better. However, they did not institute the Cs line as a new replacement source for the Kr lamp. They went far further and defined the speed of light, opening the door to the multiplicity of radiations listed in the mise en pratique. Don't you agree? See the wording summarized in the mise en pratique Brews ohare (talk) 03:15, 30 July 2010 (UTC)

That's just a wording issue; see the seventh bullet of the resolution which says "these various forms [...] have been recognized as being equivalent". A. di M. (formerly Army1987) (talk) 10:09, 30 July 2010 (UTC)

Physchim62: Your quote above simply documents exactly what is said the addition I made to the subsection: “However, the flexibility in choice of source and the ability to let the standard evolve with improved precision of the sources are the determining factors in making the change in definition.” Brews ohare (talk) 00:58, 30 July 2010 (UTC)

Brews, although the question why this particular definition was chosen has indeed to be answered but not in this article. This should be discussed in some detail in the metre article. Here it is enough to state that it was changed and what the consequences of this were for the value of the speed of light. A reader wondering why this particular choice was made is bound to have enough wits to click on the wikilink to the metre article linked at the top of the section. Since that seems to have the preference of all but one of the editors here, I suggest we revert to a more concise version and move some of the detail to the metre article. TimothyRias (talk) 09:28, 30 July 2010 (UTC)

I'm not sure what's being discussed here. Is it this series of edits [3] [4] [5]? If so, I don't see why there is all this discussion. Just go by the reliable source for the sentence which doesn't support the additions in these diffs regarding Krypton etc. For that reason, I have again reverted it [6]. Please do not add similar material that is not supported by the source for the sentence without first getting consensus here. --Bob K31416 (talk) 01:19, 30 July 2010 (UTC)

BobK: I really don't care if this sentence makes sense, so I'll leave it alone. However, it is perfectly clear that the 1960 definition is based upon the Kr transition, and to revert this phrase as unsourced is ludicrous. Brews ohare (talk) 02:31, 30 July 2010 (UTC)

As for the earlier version referring to a Kr discharge lamp, look at This way of defining the meter has proven to be particularly robust, since unlike a definition based on a standard such as the krypton lamp, length measurement can be continuously improved without resorting to a new definition. Here's another one. Here's another one. BobK, you have to become more familiar with the subject, and not go by the Talk page debates. Brews ohare (talk) 02:55, 30 July 2010 (UTC)

I've just restored the version prior to your lengthy additions.There's little point wasting our time debating whether individual sentences were factually correct or properly sourced; as a lengthy discussion of the metre they were simply irrelevant to this article, which already covers the relationship of the metre definition and speed of light very well. The article is already too long and does not need content adding which is not on the topic, the speed of light.--JohnBlackburne^words_deeds 09:39, 30 July 2010 (UTC)

Restoration of the original form of this subsection without regard to anything that has transpired here is just arrogance. Once again, this form of the subsection fails to connect the topic of "increased accuracy" with the topic of "redefinition". This presentation is just a juxtaposition of disjoint topics. Brews ohare (talk) 14:46, 30 July 2010 (UTC)

Separate article

This is an interesting subject, maybe more interesting to some than others, but it is holding up progress towards FAR for this article. Like the One-way speed of light this subject contains an interesting mix of the practical and philosophical that could be debated endlessly. Nevertheless, it is an important part of physics that should be explained in WP somewhere. I propose moving all the detail to a new article where those interested can discuss it.

By the way, I suggest that there should be a link to the One-way speed of light from this article. Martin Hogbin (talk) 09:49, 30 July 2010 (UTC)

Don't give a fig about FAR, but I agree this subject needs no more than a line or two here. Stuff it in another article "Exact metric speed of light", link to it and be done.--Michael C. Price ^talk 09:57, 30 July 2010 (UTC)

I have started Redefinition_of_the_Metre_in_1983 by copying the section here to a new page. The refs need copying too. Martin Hogbin (talk) 10:34, 30 July 2010 (UTC)

I would try to expand the Metre article first (currently it's just 26 KB), and then to split it off should it become too large. A. di M. (formerly Army1987) (talk) 11:13, 30 July 2010 (UTC)

That is an idea, but let us move the detail from here. Martin Hogbin (talk) 11:50, 30 July 2010 (UTC)

Indeed. Redefinition_of_the_Metre_in_1983 riskes becoming a PoV fork for those who think it's important. Physchim62 (talk) 11:53, 30 July 2010 (UTC)

Agree that would be the usual "best practice".TimothyRias (talk) 12:07, 30 July 2010 (UTC)

But the "exactness of the SoL" does lie in the intersection of the SoL and the metre. That justifies its own article. --Michael C. Price ^talk 12:20, 30 July 2010 (UTC)

"That justifies its own article."??? Not if we don't have anything to say about it! The big problems here is that some editors are giving the redefinition of the metre far more attention than it merits. What is wrong with just merging things into Metre#Standard wavelength of helium-neon laser light? Physchim62 (talk) 12:31, 30 July 2010 (UTC)

So, do the "exactness of the International Prototype Kilogram", the "exactness of the period of the ground state hyperfine transition of caesium-133", the "exactness of the magnetic permeability of free space", the "exactness of the triple point temperature of water", the "exactness of the luminous efficacy of 540 THz light" and the "exactness of the molar mass of carbon-12" all justify their own articles because they do lie in the intersection of their respective constant and the relevant units? A. di M. (formerly Army1987) (talk) 12:38, 30 July 2010 (UTC)

The big problems here is that some editors are giving the redefinition of the metre far more attention than it merits. That isn't a problem if the subject has its own article. And the constancy of the SoL is more interesting than the other examples listed, so your analogies break down. --Michael C. Price ^talk 20:01, 30 July 2010 (UTC)

I might agree about the definition of the kilogram (after all, that's just some old piece of metal) or the candela (it's a semi-arbitrary function intended to model the frequency response of the human eye); but as for (e.g.) the ampere, I fail to see how μ₀ is any more or any less interesting than c: after all, of the constants μ₀, ε₀, Z₀ and c any two can be expressed in terms of the other two, so I'd say none is more fundamental than another. A. di M. (formerly Army1987) (talk) 00:08, 31 July 2010 (UTC)

From a physicist's POV you are correct, but the general public are more likely to find the constancy of the speed of light interesting than the constancy of μ₀. --Michael C. Price ^talk 06:18, 31 July 2010 (UTC)

However, the constancy of the speed of light has nothing to do with our choice of units, nor should we suggest that it has. The speed of light is not constant because it has a defined value in SI units, that's getting the argument the wrong way round! Physchim62 (talk) 06:40, 31 July 2010 (UTC)

An excellent point which must be emphasized. --Michael C. Price ^talk 07:10, 31 July 2010 (UTC)

And is emphasized in the article. The constancy of the speed of light is discussed in completely different section than the exactness of the value of c in SI units. (And to be nitpicky, just for shits and giggles: Strictly speaking because the metre defined in terms of an exact speed of light is a unit of proper length, anything that previously would be interpreted as a change in c now has to be interpreted as a change in the spacetime metric. In this sense the modern definition of the metre may actually be interpreted as defining c to be constant over spacetime. This, of course, is over interpretation of the definition since no such effect was intended and any evidence that c has changed over time (or space) is bound to lead to renewed evaluation of the current definition which assumes c is constant.) TimothyRias (talk) 09:40, 31 July 2010 (UTC)

AFAIK, general relativity says you can always define things in such a way that the local speed of light is c, so do you mean "any evidence that GR is broken"? BTW, the CIPM in 2002 explicitly stated that "in the context of general relativity, the metre is considered a unit of proper length. Its definition, therefore, applies only within a spatial extent sufficiently small that the effects of the non-uniformity of the gravitational field can be ignored (note that, at the surface of the Earth, this effect in the vertical direction is about 1 part in 1016 per metre). In this case, the effects to be taken into account are those of special relativity only." A. di M. (talk) 10:31, 31 July 2010 (UTC)

(outdent) I meant, it must be emphasized in the separate article - which it now is. --Michael C. Price ^talk 09:52, 31 July 2010 (UTC)

The move of a paragraph or two to a new venue without attempting to make a stand-alone article out of it is just stupid. To testify to that, the article is tagged with AfD before anything has been done about it, indicating that this move is no more than a ploy to avoid rewriting the section on this subject in Speed of light. Tsk tsk. Brews ohare (talk) 14:46, 30 July 2010 (UTC)

The consensus is that the section does not need rewriting in such a way. It certainly did not need the lengthy off-topic and largely unsourced additions that were removed earlier. But yes, just because the content does not belong here, that does not mean a new article is needed for it - hence the PROD and AfD.--JohnBlackburne^words_deeds 15:08, 30 July 2010 (UTC)

The Speed of light subsection is now a juxtaposition of topics, not an explanation of their interconnection. A new article would be great, but it should be more detailed and broader in scope. Brews ohare (talk) 15:24, 30 July 2010 (UTC)

No it is a simple statement of fact (with sources), which is very appropriate for an encyclopidia. TimothyRias (talk) 15:32, 30 July 2010 (UTC)

Timothy: The first two paragraphs describe difficulties in obtaining an accurate metre. That's fine, but the implication is only that some new method might be better. The last paragraph then makes a jump to the definition of the metre and the resulting defined value of the speed of light. It states that the motivation for the switch is that the old definition was insufficiently precise, but there is no clue here as to why the introduction of the new definition has any effect upon this problem. It is left to the reader to connect the dots in this non-sequitor. That is my difficulty with this version. Brews ohare (talk) 15:49, 30 July 2010 (UTC)

(ec)Brews, the first two paragraphs are not describing difficulties in obtaining an accurate metre. The first paragraphs relates improvements in the accuracy of measurements of the speed of light in the second half of the 20th century (the increased accuracy part of this subsection), up to the point that no further improvements could be made do to limitations in the way the metre was defined. The second paragraph then explains what the definition of the metre was at this time and what the trouble was with it. (This may already be digressing further from the subject of the speed of light than necessary.) The last section then relates that the metre was redefined to the modern definition and what the effect of this on the value of the speed of light was. This is a simple chronological statement of facts as is appropriate in the history section of an encyclopedic article. The fact that you read the first two paragraphs as being about the metre rather than the speed of light, explains a lot about our difference of opinion and may indicate that something in these paragraphs needs to be clarified. TimothyRias (talk) 16:15, 30 July 2010 (UTC)

Do you really think that an interested reader can read the current version of the article and be unable to figure out that the 1983 definition allows a better realization of the metre than the 1960 definition does? A. di M. (formerly Army1987) (talk) 16:04, 30 July 2010 (UTC)

Now I've edited it to state "in order to allow a more precise realization of the metre". This is as clear as it gets; a reader who can't get the point now wouldn't be helped by obscure details of the interferometric techniques at all. A. di M. (formerly Army1987) (talk) 16:14, 30 July 2010 (UTC)

A di M: The addition of words like "in order to allow" serve to join sentences together, but they don't explain how the goal is accomplished. To suggest that the reader is well served by being forced to digest the entire article in order to understand the connections between a few paragraphs is just laziness in editing. Add the bridge. Brews ohare (talk) 16:27, 30 July 2010 (UTC)

Brews, editors are not stupid or lazy just because they disagree with changes you would like to make. The consensus is clearly against you on this, I suggest you stop trying to argue it now, especially given the tone of your recent comments.--JohnBlackburne^words_deeds 16:45, 30 July 2010 (UTC)

John, no comments about editors were made. It was suggested that the action of transferring a few paragraphs into a stand-alone article without taking the time to reshape them was stupid. It was said that avoidance of a few words of explanation connecting paragraphs was lazy editing. Perhaps if you wish to adopt full and sole responsibility for those actions or inactions, you can then take these comments personally. However, that was not my intention. Brews ohare (talk) 17:11, 30 July 2010 (UTC)

Please see my comments on this subject in the 'articles for deletion' page and the 'redefinition' talk page. Martin Hogbin (talk) 10:20, 31 July 2010 (UTC)

My objections to this article

This is best explained by writing up what I think the article fails to mention. I just wrote that up here (not 100% done yet, though, but you can see what my point is). Comments welcome on the talk page of that page, I don't want to derail the FAR of this article. Count Iblis (talk) 03:55, 1 August 2010 (UTC)

The main omission seems to be that the classical/Galilean limit may be obtained by expanding in 1/c. I agree that this should be mentioned somewhere, although briefly (not more than a paragraph) linking to another article with a more detailed explanation. (Classical limit would be a good candidate to add such an explanation. Any suggestion for such an introduction?TimothyRias (talk) 10:12, 1 August 2010 (UTC)

The material mentioned above seems to be Count Iblis's thoughts on the subject based on SR. There is nothing of that nature missing from here. Martin Hogbin (talk) 10:48, 1 August 2010 (UTC)

The role of c in QM/GR and the classical limits seems to be worthy of mention here IMO. As far as the FA goes, it'll fail on grounds of article stability, so don't worry about getting the star. Headbomb {talk / contribs / physics / books} 14:00, 1 August 2010 (UTC)

In 1983, in order to allow a more precise realization of the metre

In this section, this thought is inadequately followed through. Why did the definition accomplish this goal? To make it clear just why, the point must be made that a "time-of-flight" definition makes the metre have the same error as the measurement of time (or frequency). I believe that is mentioned somewhere, but it is needed here to connect the dots. Brews ohare (talk) 05:30, 1 August 2010 (UTC)

Re "the point must be made that a "time-of-flight" definition makes the metre have the same error as the measurement of time (or frequency)." - Seems like a digression. I looked at the Metre article and I didn't notice this point there. If this point appears in a reliable source, perhaps you may want to try to add it to the Metre article instead. Just a suggestion. --Bob K31416 (talk) 06:20, 1 August 2010 (UTC)

1. No further explanation is required. Wikipedia is not a textbook, but an encyclopedia and can suffice with sourced statements of fact, especially when the explanation would be a digression away from the main subject of the article. For this article it is simply not needed for the reader to understand why the change of definition allowed a more accurate realization, nor will a reader be interested in this, and if the reader is wondering about this he/she will look up the metre article.

2. The new definition does not automatically give the realization of the metre the same error as the realization of the second.TimothyRias (talk) 09:49, 1 August 2010 (UTC)

Now we have an article on the specific topic, I suggest that much of excessive detail can be removed. We want one short paragraph at the most, more or less the last paragraph as it is now. Martin Hogbin (talk) 10:44, 1 August 2010 (UTC)

Timothy: I've exaggerated the situation. The time-of-flight definition for length would have a lower bound of accuracy that is the same as that for the time interval involved. The interferometric method would have a larger error due to the added errors of interferometry.

The point you all choose to ignore, however, is that WP claims that the change in definition enabled higher accuracy in the realization of the metre. In the case of a time-of-flight measurement, that is possibly true; I don't know much about errors in time-of-flight measurements, and so far as I can tell they are seldom used in a laboratory. However in the normal lab implementation of part (b) using an interferometer, it was not the change in definition that allowed greater precision, but the change away from the krypton discharge lamp to better laser sources. From the perspective of a wider variety of applications, like surveying, the definition enabled better selection of a source for the purpose, which in a broad sense enabled greater precision of realization in practical pursuits. The requirements in these situations are less precise than a scientific context, of course.

So the bottom line is that the statement in WP is misleading, and largely wrong. The accuracy improvements are due to better sources, not due to the choice of definition, which has mainly the advantage of allowing accurate comparison between sources. That capacity enables evolution of the standard with evolution in sources. Brews ohare (talk) 00:32, 2 August 2010 (UTC)

The whole point of the new definition is that it is no longer linked to any particular source, but allows a realization of the metre with any radiation source of which the frequency is accurately known. So, yes the new definition did enable higher accuracy in the realization of the metre. A claim that the article doesn't even make, even the old text only claimed that the intent of the new definition was to allow for a more precise realization of the metre. TimothyRias (talk) 09:11, 2 August 2010 (UTC)

TImothy: The definition of the metre made it measurable in seconds. That change was completely unnecessary so far as improving the accuracy was concerned. The old definition would work just as well as the new one if a source better than the krypton lamp were selected and the old definition changed to refer to wavelengths of the new source. So accuracy and the new definition are unrelated. As you point out, what the new definition did was to allow selection of a variety of sources, which meant as a practical matter that the CGPM didn't have to introduce a new definition every time a better source came along. I suspect we both agree about this; the issue is just a bit better precision in expressing matters. Brews ohare (talk) 13:53, 2 August 2010 (UTC)

Well, the new definition did allow a better realisation didn't it? A discussion why this definition was chosen over other definitions that would have allowed a better realisation of the metre, seems out of place here since that requires much more detail. Such an explanation is much better suited for the new Redefinition of the metre in 1983 article or the metre article. At the present, the article simply gives the reason they chose a new definition at all, without going into the reasons why they chose that particular definition. This seems fine to me.

(As to your first sentence: the 1960 definition made the metre measurable in seconds times a ratio of energy levels of krypton and caesium so I don't see there be that much change except the elimination of the dependence on an unnecessary quantity.)TimothyRias (talk) 14:22, 2 August 2010 (UTC)

Timothy: Well, if you wish to avoid discussion of why the definition was adopted, fine. But if you wish to enter upon why it was adopted, it is best not to present a distorted version. The conversion of lengths to seconds is by far the most striking aspect of the new definition and the one that attracts the most curiosity. So to say by implication that the purpose of the new definition was greater accuracy is tantamount to saying that the switch to lengths as times led to greater accuracy. That interpretation would be incorrect. Because that interpretation is the likely outcome of the present WP wording, that wording should be changed. The full list of eight reasons for the switch is found here. Brews ohare (talk) 14:45, 2 August 2010 (UTC)

Seems like further details beyond what is already in the article, regarding why the definition of the metre was adopted, would be too much of a digression from the topic of the article Speed of light, IMO. --Bob K31416 (talk) 15:20, 2 August 2010 (UTC)

The wording is now: “In 1983, considering that the definition of the metre did not allow a sufficiently precise realization of it for all requirements,” a wording probably meaningless to the reader (although employed in the resolution), who cannot presume to understand what "all requirements" includes. So why not just link again to the CGPM Resolution that itemizes the eight reasons and let it drop: “In 1983, for various reasons detailed in Resolution 1,” eh? Brews ohare (talk) 17:53, 2 August 2010 (UTC)

If more detail is desirable, this sentence could be added to, for example, to point out the desire to avoid annual redefinitions of the metre based upon ever newer & better sources. Brews ohare (talk) 18:04, 2 August 2010 (UTC)

Compared to the first item that you were suggesting, I think it's more informative while still concise, and the source is still available in the present version for the reader to look up further details. The second item that you are suggesting to be added would be a needless digression IMO. --Bob K31416 (talk) 21:55, 2 August 2010 (UTC)

What else can be included before FAR

Scharnhorst effect. What other not widely known things or technical issues could we mention in the speed of light article before going to FAR. If not in the main article then as footnotes or see alsos. 89.240.130.248 (talk) 14:31, 1 August 2010 (UTC)

The Scharnhorst effect is already mentioned in this note. --Bob K31416 (talk) 14:43, 1 August 2010 (UTC)

What other not widely known things or technical issues could we mention in the speed of light article before going to FAR. If not in the main article then as footnotes or see alsos. 89.240.130.248 (talk) 14:31, 1 August 2010 (UTC) —Preceding unsigned comment added by 89.242.128.249 (talk)

Notation

It seems that the international standards organization symbol c₀ for the numerical value is not mentioned in the article. Brews ohare (talk) 01:09, 2 August 2010 (UTC) Sorry, I found it. Brews ohare (talk) 01:11, 2 August 2010 (UTC)

Constancy of the speed of light and units

On the Talk page above, it is said:

However, the constancy of the speed of light has nothing to do with our choice of units, nor should we suggest that it has. The speed of light is not constant because it has a defined value in SI units, that's getting the argument the wrong way round! Physchim62 (talk) 06:40, 31 July 2010 (UTC)

That statement is accurate. However, the ease and accessibility to realization of the unit of speed is certainly an important factor. That is why the speed of light in vacuum was chosen instead of (say) the speed of sound in water under standard conditions.

This point is not made in the article. Probably it belongs in the redefinition section. Brews ohare (talk) 18:15, 2 August 2010 (UTC) I made an attempt at doing this. Brews ohare (talk) 19:12, 2 August 2010 (UTC)

You have got it the wrong way round Brews, as Physchim62 says. Everything we know about the universe tells us that the speed of light (in appropriate conditions) is a fundamental constant. It is therefore a good idea to incorporate it into our standards system. Martin Hogbin (talk) 19:21, 2 August 2010 (UTC)

Martin: I have agreed entirely with Physchim62, and made a quite separate point that agrees entirely with yours. Brews ohare (talk) 19:37, 2 August 2010 (UTC)

Sorry Brews, I am still puzzled over the exact point that you are trying to make in this article. Martin Hogbin (talk) 10:42, 3 August 2010 (UTC)

Let me explain: suppose we set up a standard speed of sound in water by specifying as carefully as need be the circumstances of the water, the exact pitch of the sound, its amplitude and a myriad of complex details that must be established by an experimenter to insure that the standard speed is realized. Call this speed s₀. Then the metre could be defined as the distance traveled by standard sound in a ms (say). The speed of light in vacuum would be measured to be some multiple of s₀. Everything would work the same way it does now in the SI units. For instance, s₀ would be a defined constant, s₀ = 1000 m/s, outside measurement. Improvements in measuring sound or refinements in defining the standard water medium would not change s₀, but would refine the metre.

What are the defects of this approach? Mainly they are the complexity of realizing the standard speed, not the least of which is the preparation of the standard water medium and insuring that it is in fact a good representation of the standard medium.

What are the advantages of choosing instead the speed of light? Viewed entirely from an empirical observational viewpoint, and not from the viewpoint of relativity (which only explains these observations), the advantages of choosing the speed of light are that it is readily realized because it is independent of almost every parameter one might think of: orientation, speed of the emitting source, frequency of the radiation, intensity of the radiation, polarization of the radiation, blah, blah, blah. The hooker remains that the standard medium, vacuum, still has to be obtained, but the preparation of vacuum is thought to be understood, especially if you steer clear of things like quantum field effects by sweeping them under the rug (or, if you like, by virtue of their being so very small that for measurement purposes they are way down the list of errors). Gravity is a bit of a problem.

So what this means for the article is to point out that there is a general choice available in choosing base units, namely to choose a standard of speed to replace the standard of length. As a standard of speed, the above properties of light recommend it above any other such standard one might propose. Those advantages are not based upon theory (although supported by theory, of course, as that is the function of theory) but based upon empirical observations. Brews ohare (talk) 13:11, 3 August 2010 (UTC)

When you say that the advantages of light are not based on theory you are wrong. It would be better to say that the advantages of using light are based on theory, which is itself based on empirical observations.

I think by theory you really mean theories that you do not like the look of. Pretty well every experiment and observation in physics has to be based on theory. Martin Hogbin (talk) 13:28, 3 August 2010 (UTC)

Martin: Spoken as a true theoretician. Of course, the theoretician believes experimentalists are a bunch of unimaginative lackeys who must be told where to look or they wouldn't ever observe the important things. However, the experimenters may object to this characterization, and point out that Planck came up with the quantum trying to explain black body data, and Maxwell stumbled upon the em origin of light by observing a coincidence in the ratio of units. In any event, your view of matters is debatable. And in this particular context, as a practical matter in lauding the speed of light as a standard of speed superior to all others, theory is not the issue: the issue is the observed advantages, which will not go away even if relativity goes down the drain. Brews ohare (talk)

You seem to have a dislike of some theories. The problem is that all physics is based on theory. It would be impossible to attach any meaning to the term speed of light without some underlying theory of what light was and what it did. How would you measure the speed of light using Ptolemy's emission theory? The question would have no meaning. Martin Hogbin (talk) 13:47, 3 August 2010 (UTC)

What is the advantage in pointing these matters out in the article? A major advantage is that it removes the mystique of the speed of light from the formulation of a standard of speed. By lifting this baggage, the adoption of a speed standard becomes a simple idea. It is not confounded with all the brouhaha surrounding the speed of light and its multiple roles in physics over the millenia: the aether, relativity, em propagation etc. The adoption of the speed of light as a standard is just a super convenience, not a magical event. Brews ohare (talk) 13:24, 3 August 2010 (UTC)

I think it is something between those two. No one claims that the speed of light has magic properties or even that it is a fundamental part of nature. How could we tell? It is, however, a fundamental part of our currently accepted and verified theories of physics and, as such, it is a much better basis on which to build a system of units than an arbitrary measurement such as the speed of sound in water. The trend in metrology is towards using fundamental (according to our current understanding) constants of physics. It would be better if we could use units based on our fundamental concepts of the future but unfortunately nobody know how to do this. Martin Hogbin (talk) 13:47, 3 August 2010 (UTC)

Martin, I am just being a devil's advocate here. Of course theory and experiment have both played a role. I don't agree with your assessment that the speed of light is a great basis for metrology because of its fundamental role in theory. For example, if it happened that the speed of light was very dependent upon polarization, field intensity, direction of propagation etc. etc. , then it would not be suitable for metrology, regardless of its fundamental importance. And if the propagation of sound in some 0 Kelvin superfluid had super reproducibility, it would be a great candidate, even if it was a negligible theoretical curiosity. The advantage of the speed of light for metrology is its super convenience, which comes from its easy realization regardless of almost any parameter you can think of, except gravity. That capacity of light is a fact, not a theory. Brews ohare (talk) 14:07, 3 August 2010 (UTC)

Apart from this religious digression, the above advantages for the article of first pointing out the general possibility of setting a standard of speed rather than length, and then followed by the advantages of light for this purpose, remain the point. Brews ohare (talk) 14:07, 3 August 2010 (UTC)

The point is that theory tells us that the speed of light does not depend on any of the things that you mention. Of course the theory could be disproved tomorrow but it is still better to build a measurement system on what we confidently expect to be the case rather than some arbitrary assumption or artifact.

I still do not think you have understood how much physics depends on theory of some kind. We have to start with a theory that tells us that light even exists. Martin Hogbin (talk) 14:20, 3 August 2010 (UTC)

Martin: Theory is not pertinent here. What matters for metrology is reproducibility and accuracy, and these are experimental matters, whatever their explanation. For the article, the point is that the logical possibility of using a standard speed exists, and the exemplary empirical reproducibility of the speed of light (whatever its explanation) eminently suits it for this role, and thereby weights the decision about units strongly in favor of adopting a standard speed as a basis for the SI system. That is what the article should point out. It should illuminate the metrology argument, not the importance of theory. Brews ohare (talk) 14:31, 3 August 2010 (UTC)

You still have not understood what theory is. Even the fact that light exists at all is a theory. Read this article! Martin Hogbin (talk) 14:38, 3 August 2010 (UTC)

Martin: It doesn't matter whether I understand what theory is, nor whether you do. Please address the metrology. Brews ohare (talk) 14:46, 3 August 2010 (UTC)

Metrology is about repeatability, reproducibility and accuracy. Brews ohare (talk) 14:56, 3 August 2010 (UTC)

If you do not have a theory that tells you light exists then you cannot measure it in any way at all, let alone one that is repeatable, reproducible and accurate. You are using the word 'theory' to mean stuff you do not like. Martin Hogbin (talk) 16:48, 3 August 2010 (UTC)

(ec, I also agree with Brews latest comment on metrology) I think that Brews has a point here about the reproducibility of the experiemntal conditions one uses to define the metre, but one has to add here that small deviations around this ideal state don't lead to significant errors, partially because they produce small effects and also because such effects are predictable to a high degree of accuracy and can thus can be compensated for. If you measure the same length over and over again, you get some distribution around an average value, and then what matters is that the spread around the average is smaller than what you would get using another method, otherwise you would have to think about using that other method to define the metre.

Another issue is the experimental limits on any deviations of the theory that describes the experimental setting that renders the metre. In general, it is not the case that if you can render some standard with an error of sigma, that therefore the theory that describes the experiemntal setting for this can have deviations that fall just within this area. In case of special relativity, this certainly is not the case. Any violation of Lorentz invariance that are not yet ruled out, cannot influence the metre standard, because these would produce effects that are many orders of magnitude smaller than the other known effects. Count Iblis (talk) 14:58, 3 August 2010 (UTC)

Regarding the comment by Brews ohare that "...the above advantages for the article of first pointing out the general possibility of setting a standard of speed rather than length, and then followed by the advantages of light for this purpose, remain the point." - In what reliable source do these ideas appear? --Bob K31416 (talk) 15:12, 3 August 2010 (UTC)

BobK: I don't think we need any additional sources to point out that in the SI units the speed of light is a defined constant. Likewise, no additional sources are needed to point out that length is now a derived unit based upon the defined speed of light. Perhaps it is unclear that any choice for a standard speed has the same effect, namely lengths are measured in seconds? Assuming that point of logic is understood, it follows that there exists any number of choices for that standard speed, and one can ask: Upon what basis is a selection made? The answer to that question, according to the precepts of metrology, is based upon repeatability, reproducibility and accuracy. Brews ohare (talk) 15:40, 3 August 2010 (UTC)

Re " I don't think we need any additional sources..." - Please give the excerpt from the reliable source that expresses the idea of "the general possibility of setting a standard of speed rather than length, and then followed by the advantages of light for this purpose". --Bob K31416 (talk) 15:58, 3 August 2010 (UTC)

BobK: The SI units have done exactly that: they choose a standard of speed rather than length. It hardly needs a source to point out the possibility when we have an existence proof in front of our eyes. As for the advantages of light as a choice, they have been described in detail. I could ask you for a counterexample that would be better; I don't think you can find one. Brews ohare (talk) 16:04, 3 August 2010 (UTC)

Please note that I will have to see the excerpt before I can support any addition to the article related to the subject ideas and consensus will be needed for any such additions. --Bob K31416 (talk) 16:25, 3 August 2010 (UTC)

OK, BobK: How's this for a paragraph outline:

• SI units use a defined c. [Source CGPM resolution]
• Lengths are measured in terms of times of transit, t by ℓ = c t. [Source CGPM resolution]
• If instead of c one chose some other speed s, the analogous thing would be to set ℓ = s t. [Obvious]
• What considerations would enter a choice for s other than the speed of light? [Source on metrological considerations of accuracy reproducibility ease of access etc.]
• Observation that the speed of light is independent from variation in a host of parameters [Source: texts on experimental verification of relativity]
• Conclusion: Speed of light satisfies the metrology requirements pretty well. [Source: Sydenham on accuracy]

Would a paragraph along these lines meet with your approval, or why not? Brews ohare (talk) 16:45, 3 August 2010 (UTC)

Before I can support this addition to the article I would have to decide whether it is consistent with WP:NOR, so please give the wording of the conclusion that you would like to put in the article and give the excerpt from the reliable source that makes that conclusion. --Bob K31416 (talk) 17:54, 3 August 2010 (UTC)

Hi BobK: I'd be happy to make an attempt at drafting a few lines, but Blackburne has decided to make a federal case out of my participation here and here, so I'm quitting for the moment. Brews ohare (talk) 19:19, 3 August 2010 (UTC)

Relation of c to fundamental units

Here is an excerpt describing the role of fundamental constants:

“In one common usage the fundamental constants are a set of quantities, the knowledge of which is sufficient to predict, using appropriate theory, all the properties of matter and radiation at both a macroscopic and a microscopic level. Candidates for such a set are G, the gravitational constant; h, the Planck constant; c, the speed of light; e, the electron charge; m_e, the mass of the electron; m_p, the mass of the proton; k, the Boltzmann constant; θ_w, the Weinberg angle, which relates the charged and neutral weak currents; G_F, the coupling constant for the weak nuclear interaction; θ_C, the Cabibbo angle, which relates the strangeness changing and nonstrangeness changing weak nuclear interactions; and Λ, the quantum chromodynamic coupling parameter, which characterizes the strong nuclear interaction.”

“[For example] quantum electrodynamics and the Dirac theory of the electron [enabled] the calculation in 1947 of the magnetic moment of the electron in terms of e, h, m_e and c. There is as yet [1983] no such precise calculation of the magnetic moment of the proton.”

Question: What should this article say about how a definition of c = 299,792,458 m/s unrelated to measurement affects its role in such a context? Shouldn't the article discuss the modification of the relationship of c to fundamental properties of nature when the definition of the units precludes alteration of c by measurements, no matter how refined, and no matter what path the evolution of theory may take? Brews ohare (talk) 15:29, 3 August 2010 (UTC)

I think the article's fine on this, and don't see that anything else needs adding. Of course there is more to it than is in the article, but for deeper theoretical and and mathematical underpinnings the reader is better served by e.g. the article on special relativity. As for 'what path the evolution of theory may take' that is simply an unknown about which little useful can be said until it actually happens and we know what new form theory takes.--JohnBlackburne^words_deeds 15:43, 3 August 2010 (UTC)

(edit conflict) Regarding the title of this section "Relation of c to fundamental units", it doesn't appear that the quoted excerpt discusses that but rather its topic is described in the sentence, "In one common usage the fundamental constants are a set of quantities, the knowledge of which is sufficient to predict, using appropriate theory, all the properties of matter and radiation at both a macroscopic and a microscopic level." --Bob K31416 (talk) 15:46, 3 August 2010 (UTC)

(edit conflict)Nothing really. The only thing that should be made clear as the redefinition of the metre does not affect c at all, in any way. The constant c does not care in what units you express it. This also means that its value does not have any predictive value whatsoever. The quote above is somewhat naive in this respect. It is widely accepted that only dimensionless constants can be fundamental in the sense described in that paragraph.TimothyRias (talk) 15:50, 3 August 2010 (UTC)

There are three responses above: (i) It's too complicated, let's forget about it. (ii) I don't understand the issue. (iii) It doesn't matter because I choose to define the term “fundamental constant” differently, and so I can duck the question. Brews ohare (talk) 15:55, 3 August 2010 (UTC)

Well, you can write down the complete theory with all these constants and then using scale transforms map it to a theory of the same form but with different values for c, hbar, k, and G. Any set of values for these constants can be obtained from any other via scale transformations. What does that tell you about these constants? Count Iblis (talk) 16:06, 3 August 2010 (UTC)

The speed of light as a conversion factor contains absolutely no information. The speed of light as a phenomena limiting the transfer of information has significance in separating our universe for other hypothetical ones. There is a difference here: info vs. no info. It may be that additional info is needed, but there is still some info here that is not present in a conversion factor set in a conference room debating easy and accurate lab procedures. I'd guess that the correct answer to the question is that various expressions involving c are not to be evaluated setting c = c₀ but by reforming all such formulas so c doesn't appear any more. Brews ohare (talk) 16:21, 3 August 2010 (UTC)

Or, by interpreting c as not a reference to the SI c₀ but to the real phenomena; that is, in principle, c ≠ c₀ but given by some formula based upon a theory of em propagation. Brews ohare (talk) 16:32, 3 August 2010 (UTC)

This is my favorite answer to the question: in such formulas involving c, where the real speed of light in the sense of something measurable is meant, c can be replaced in SI units by c₀/n where n is the refractive index of the medium, probably referring to a form of vacuum such as the quantum vacuum or the quantum chromodynamic vacuum. Brews ohare (talk) 17:22, 3 August 2010 (UTC)

c as a conversion factor is relative to some standard normalization. The conversion factor would not be needed in the first place had the units be chosen in a "natural way" w.r.t. the fundamental laws. Then because classical physics is invariant w.r.t. rescaling of time, there was no way we could have defined the units for distances and times relative to each other in the "correct way" in previous centuries. But the relativistic equations are not invariant w.r.t. to rescaling the time coordinate. Then because we still want to rescale the time coordinate, even when the laws are not invariant under such a rescaling (for hystorical reasons, for practical reasons or whatever), the constant c appears. It's role is merely to compensate for any rescaling of the time coordinate. Count Iblis (talk) 18:06, 3 August 2010 (UTC)

Does this article have space to attempt to illustrate why c is c ?

I have just been wondering this for sometime and have not found any full answers via whitepapers on the subject.

What is the fundamental reason the maximal speed of anything in the universe is 3x10^8 m/s ? I found an article on Scientific American here that hypothesized that it comes about due to objects residing in the dimensions of space and time needing a factor of conversion to truly be facets of the same coin (of 4 dimensional space-time). Is this argument completely off the wall or are there any respectable scientific journals that debate this concept and bring light(pardon the pun) to the origins of this upper limit of traversal of any mass-less object from one point in space to another...

It probably doesn't deserve a place in the article unless there are some papers discussing the origin of fundamental constants giving rise to this idea. But i'd like to see it explored if it has merit. Does anyone have any opinions or data they could add to this query?

TL:DR I understand that there couldn't exist a universe with no upper limit, but i don't understand what drives the value for the limit as it stands today. —Preceding unsigned comment added by 109.78.64.45 (talk) 17:19, 3 August 2010 (UTC)