Talk:Wavelength/Archive 5
![]() | This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | ← | Archive 3 | Archive 4 | Archive 5 |
Reference to Fourier series
Comment is sought as to whether a reference to Fourier series is appropriate under the heading general periodic waveforms. Some editors appear to find connecting Fourier series to wavelength is a digression. The text in question is provided below. Brews ohare (talk) 20:09, 20 April 2012 (UTC)
Text referring to Fourier series
This text is to appear following the equation defining the wavelength of a periodic function in the section More general waveforms, just before the header Envelope waves:
The wavelength, say λ, of a general spatially periodic waveform is the spatial interval in which one cycle of the function repeats itself. Sinusoidal waves with wavelengths related to λ can superimpose to create this spatially periodic waveform. Such a superposition of sinusoids is mathematically described as a Fourier series, and is simply a summation of the sinusoidally varying component waves: |
.. "Fourier's theorem states that a function f(x) of spatial period λ, can be synthesized as a sum of harmonic functions whose wavelengths are integral submultiples of λ (i.e. λ, λ/2, λ/3, etc.)."[Note 1] |
References
|
Author's remarks
- This text states the specific connection between wavelength and Fourier series as applied to spatial periodicity. It alerts readers to this connection between wavelength and some very important mathematics.
- Some discussion appears on this Talk page regarding a previous suggestion involving a more elaborate discussion of Fourier series presented here. That discussion is not pertinent in the present case because what we have here is a much more curtailed description. Brews ohare (talk) 13:47, 22 April 2012 (UTC)
Comments
"It is noteworthy that the wavelength of a general periodic waveform is related mathematically to its Fourier series expression..." is very odd. First of all, the source does not support a claim that "it is noteworthy"; that just sets off BS detectors. Second, the relationship of the wavelength (or the period) of a general periodic function to it Fourier series is simply that you need to know the former to compute the latter. How does this help explain or understand wavelength? And if he had used the word "period" like everyone else, would we even be discussing it? No. The only reason this guy cares about Fourier analysis of things with wavelengths is that he's doing X-Ray crystallography, and I don't think it's appropriate to get into that here; maybe a link from the crystals section would be OK. Dicklyon (talk) 01:57, 21 April 2012 (UTC)
- Dick, the source is cited as the origin of the quotation. I've moved the footnote to make this clearer.
- This author's use of "wavelength" in the quotation is entirely appropriate and accurate and often used. Ordinarily (although not invariably) period is taken to refer to periodicity in time or with respect to some general variable, say ξ or θ, while wavelength invariably refers to periodicity in space. Of course, periodicity can be expressed in many different ways, but wavelength is the subject here, and wavelength is used to define a periodic function in the subsection More general waveforms where this text referring to Fourier series is proposed to be placed. The mathematical definition is stated in the article as f(x−vt+λ)=f(x−vt). So I don't accept your stance that "wavelength" is being awkwardly squeezed into the discussion by using an uncommon usage of terminology.
- As for the importance of this topic, the entire subject of waveforms periodic in space with a certain wavelength is inextricably connected to Fourier series. Perhaps you can suggest some more innocuous wording to replace "noteworthy" that you would find acceptable? Brews ohare (talk) 15:25, 21 April 2012 (UTC)
- I'll wait until we get other comments. This is supposed to be a RFC, not us arguing some more. Dicklyon (talk) 15:46, 21 April 2012 (UTC)
- I removed "noteworthy" as unnecessary. Brews ohare (talk) 16:07, 21 April 2012 (UTC)
- I'll wait until we get other comments. This is supposed to be a RFC, not us arguing some more. Dicklyon (talk) 15:46, 21 April 2012 (UTC)
- I agree with Dicklyon. Fourier analysis is a tool for mathematical study of periodic functions. It has nothing in particular to do with wavelength. You can use Fourier analysis on a purely time-based phenomenon or an abstract function, neither of which has a wavelength. The Fourier series does not help readers understand waves or wavelength. Quite the opposite. And the source for the quote is a narrow technical one inappropriate for a mathematical article (and a commercial link is entirely inappropriate for a reference – let readers find that via the ISBN if they want to).--JohnBlackburnewordsdeeds 19:56, 21 April 2012 (UTC)
- John, you are making a statement logically equivalent to this: A knife is a cutting implement so it is irrelevant to point out under "weapon" that a knife can be used this way. Likewise, Fourier analysis can be viewed as a general discussion of periodicity in any variable. So you claim in particular, and as is the topic here, it is irrelevant to point out its use to analyze spatial variation. Brews ohare (talk) 20:44, 21 April 2012 (UTC)
- You can reinterpret my statement as you like; my point is I agree with Dicklyon and now IRWolfie- that it doesn't belong.--JohnBlackburnewordsdeeds 21:08, 21 April 2012 (UTC)
- John, no, the point is you "agree" with these editors but offer no reason for doing so, as I have pointed out. Brews ohare (talk) 00:57, 22 April 2012 (UTC)
- My reasons are here as you seem to have trouble finding them.--JohnBlackburnewordsdeeds 11:04, 22 April 2012 (UTC)
- Right, John. As to your comments that you have linked with caustic observation about myself, I see again that you propose as a "reason" for your supposed "agreement" with DickLyon, that because Fourier series have wide application to periodicity in general, therefore my proposed text stating the specific connection between wavelength and Fourier series as applied to spatial periodicity is neither useful nor appropriate. I think you can see your position is neither logical nor sensitive to the key role of WP in broadening a reader's grasp of the context of a topic they are reading about. Brews ohare (talk) 13:39, 22 April 2012 (UTC)
- You can reinterpret my statement as you like; my point is I agree with Dicklyon and now IRWolfie- that it doesn't belong.--JohnBlackburnewordsdeeds 21:08, 21 April 2012 (UTC)
- John, you are making a statement logically equivalent to this: A knife is a cutting implement so it is irrelevant to point out under "weapon" that a knife can be used this way. Likewise, Fourier analysis can be viewed as a general discussion of periodicity in any variable. So you claim in particular, and as is the topic here, it is irrelevant to point out its use to analyze spatial variation. Brews ohare (talk) 20:44, 21 April 2012 (UTC)
- Oppose It has nothing in particular to add to the topic of the article and thus strays off topic. IRWolfie- (talk) 20:24, 21 April 2012 (UTC)
- IRWolfie: How exactly is it straying off topic to point out the connection between wavelength and Fourier series for spatially periodic functions? Are you claiming the connection is not there, or that the role of wavelength does not enter the construction of a Fourier series for a spatially varying function? Either claim would directly contradict the cited sources (or in fact, any sources). Brews ohare (talk) 20:44, 21 April 2012 (UTC)
- The point is that wavelength is defined in general in terms of a general, spatially periodic waveform, and such waveforms are described by Fourier series having a particular relation to the wavelength of the waveform. That is a nontrivial observation and connects the topic of wavelength to the very important context of Fourier series. A sentence or two about this is not a serious expansion of the article and is a small addition to provide this important connection for the topic of wavelength. Brews ohare (talk) 01:35, 22 April 2012 (UTC)
- Regarding your first paragraph: the fact that wavelength plays a role in the construction of Fourier series does not help your argument. The problem is that the connection goes the wrong way. Wavelength is certainly relevant to Fourier series, but that does not imply that Fourier series are relevant to an article on wavelength. Relevance is not always bidirectional.
- Regarding the second paragraph: It is not clear to me that this observation is "nontrivial". --Srleffler (talk) 16:00, 22 April 2012 (UTC)
- I agree that backing Fourier series into a discussion of wavelength goes the wrong way around. A "See also" entry, or an inline wikilink such as the one in the section on wave packets, should suffice. __ Just plain Bill (talk) 16:17, 22 April 2012 (UTC)
- Hi Srleffler: I am happy that you see wavelength plays a role in Fourier series for spatially periodic functions. Your general approach to rejection of this observation in Wavelength is that somehow the importance of wavelength to Fourier series does not support its mention here because Fourier series is "not relevant to wavelength". That stance seems strange to me. You point out that this type of objection is not universally applicable, so for instance a pumpkin might be referred to in an article on vegetables, and vice versa. Perhaps you could explain further why the connection of wavelength to the sub-domain of Fourier series for spatially varying functions is of no interest to those interested in wavelength? Aren't readers of Wavelength entitled to know about this connection to a truly gigantic area of mathematics? Brews ohare (talk) 16:25, 22 April 2012 (UTC)
- This is a general issue that comes up now and again in discussions of appropriate content for Wikipedia articles. The fact that topic A is of great importance to field B justifies mentioning A in the article on B. It does not follow, however, that topic B deserves a mention in the article on A. Usually, B will be worth mentioning but for other reasons, not merely because A is important to B. Cases where there is strong one-way relevance are not uncommon, but are certainly not the majority. As a result, when they arise it may provoke a talk page discussion such as we have been having.
- The question we need to settle here is whether Fourier series have some relevance in the discussion of wavelength, regardless of the latter's importance to the former. --Srleffler (talk) 18:13, 22 April 2012 (UTC)
- Just plain Bill: The suggested text is not " backing Fourier series into a discussion of wavelength". The idea of wavelength is introduced in the article via the periodicity condition f(x−vt+λ)=f(x−vt). That is exactly and inescapably the subject of Fourier series. Pointing out that connection in a sourced sentence is hardly a large cost to this article in terms of added space. Brews ohare (talk) 16:25, 22 April 2012 (UTC)
- Hi Srleffler: I am happy that you see wavelength plays a role in Fourier series for spatially periodic functions. Your general approach to rejection of this observation in Wavelength is that somehow the importance of wavelength to Fourier series does not support its mention here because Fourier series is "not relevant to wavelength". That stance seems strange to me. You point out that this type of objection is not universally applicable, so for instance a pumpkin might be referred to in an article on vegetables, and vice versa. Perhaps you could explain further why the connection of wavelength to the sub-domain of Fourier series for spatially varying functions is of no interest to those interested in wavelength? Aren't readers of Wavelength entitled to know about this connection to a truly gigantic area of mathematics? Brews ohare (talk) 16:25, 22 April 2012 (UTC)
- I agree that backing Fourier series into a discussion of wavelength goes the wrong way around. A "See also" entry, or an inline wikilink such as the one in the section on wave packets, should suffice. __ Just plain Bill (talk) 16:17, 22 April 2012 (UTC)
The suggested text has been through at least three iterations since that green box appeared above. To a reader passing by, it is no longer apparent which version was being discussed at various times. I would prefer to see new suggestions given their own space where they were introduced, without obliterating the previous one(s).
Wavelength may be helpful in understanding some applications of Fourier series. Fourier analysis is not necessarily helpful in understanding wavelength. That's the asymmetry I meant by "backing into." Space on the page is one thing, but introducing tangential material is, IMO, a distraction to the reader. Sourced, you say... it is our job to be selective about which sources are relevant to the subject. __ Just plain Bill (talk) 17:29, 22 April 2012 (UTC)
- That's my concern too. The introduction of Fourier series appears neither to make the concept of wavelength clearer nor to provide a better fundamental definition of wavelength of a general periodic wave. The current proposed text is admittedly better than the many attempts prior to this RFC, in that it doesn't belabour the issue and focuses most directly on the connection between the two topics.
- I do see the appeal in trying to replace the definition of wavelength of a non-sinusoidal wave in terms of the wave's period of repetition with a definition that is tied directly to sinusoidal waves. I'm partial to this for the same reason that I was originally opposed to applying the term "wavelength" to non-sinuoidal waves at all. It's not clear to me that there is a non-negligible set of readers for whom this treatment would be beneficial, however.--Srleffler (talk) 18:13, 22 April 2012 (UTC)
- It's hard to use Fourier series to define wavelength, just as it's hard to use it to define pitch (music), since there may not by a sinusoidal component with wavelength matching the wavelength that you're trying to define, and because you need to know the wavelength before you can do a Fourier series. But we could perhaps connect it in by noting that non-sinusoidal waves are sometimes analyzed in terms of superimposed sinusoids, each of which has its own well-defined wavelength. The current "The wavelength of a general periodic waveform is related mathematically to its Fourier series..." completely misses the point, and provides no motivation for the claimed relationship, and no real role for the series, or the decomposition that it induces. The X-ray crystallography source isn't clear on this either, and in that field it would be much more fruitful to use wavevectors, since the wavelength formulation doesn't extend to the multi-dimensional case very naturally. If we had a source about decomposing waves into sinusoidal components for the purpose of easier analysis, that might make sense somewhere here; but it wouldn't necessarily be tied to perioidic waves, since its applicability is broader than that; so it might need to be Fourier transforms instead of series. Like Brews, many authors have forgotten what's special about sine waves that motivates such decompositions (for example, this one: "Sinusoidal waves are important because they occur in many physical situations..." -- seriously?). Dicklyon (talk) 18:37, 22 April 2012 (UTC)
- I am happy to see this conversation turning to real issues. It seems there are several points raised that I'd like to try to summarize and respond to, hopefully in a constructive manner:
- Fourier series for spatially periodic waveforms employ wavelength, but do not illuminate the concept of wavelength. That seems to me a valid observation. I'd argue, however, that because wavelength shows up prominently in such Fourier series, it is reasonable to point to this fact here. Gasoline may not be illuminated by its role in the internal combustion engine, but it can be of interest in an article on gasoline that it is used in internal combustion engines, and vice versa.
- DickLyon's remark: "you need to know the wavelength before you can do a Fourier series". This comment relates to using Fourier series to determine wavelength. That is an interesting exercise, but it is not the proposed reason for inserting this brief reference to Fourier series here. It seems to me that "using Fourier series" to determine the wavelength of a complicated repeating waveform is most likely to be a practical undertaking in the theory of separating a signal from background, where one might fit some experimental data with λ as a variable parameter and adjust it for a best fit. However, that or other practical instances where one would use Fourier series to determine wavelength, seems beyond the intended scope of simply pointing out that Fourier series use λ.
- DickLyon's remark: "it would be much more fruitful to use wavevectors, since the wavelength formulation doesn't extend to the multi-dimensional case very naturally." This article is about wavelength, so it is a digression to ask whether Fourier series are more easily generalized using different concepts.
- DickLyon's remark: "its applicability is broader than that; so it might need to be Fourier transforms instead of series." If one wishes to discuss non-periodic waveforms, the wavelength does not come up because its underlying definition is f(x−vt+λ)=f(x−vt), which does not apply except to periodic functions. Later in the article wave packets are described, but the text about Fourier series suggested here is proposed for insertion in the section on general periodic waveforms.
- DickLyon's remark: "many authors have forgotten what's special about sine waves that motivates such decompositions". If there is something particular that should be said about λ in this regard, that is an argument for adding to the proposed text, not for eliminating it. For example, one could digress to suggest that different musical instruments playing the "same" note have different voices because their Fourier series have different terms.
- Brews ohare (talk) 21:47, 22 April 2012 (UTC)
- It's hard to use Fourier series to define wavelength, just as it's hard to use it to define pitch (music), since there may not by a sinusoidal component with wavelength matching the wavelength that you're trying to define, and because you need to know the wavelength before you can do a Fourier series. But we could perhaps connect it in by noting that non-sinusoidal waves are sometimes analyzed in terms of superimposed sinusoids, each of which has its own well-defined wavelength. The current "The wavelength of a general periodic waveform is related mathematically to its Fourier series..." completely misses the point, and provides no motivation for the claimed relationship, and no real role for the series, or the decomposition that it induces. The X-ray crystallography source isn't clear on this either, and in that field it would be much more fruitful to use wavevectors, since the wavelength formulation doesn't extend to the multi-dimensional case very naturally. If we had a source about decomposing waves into sinusoidal components for the purpose of easier analysis, that might make sense somewhere here; but it wouldn't necessarily be tied to perioidic waves, since its applicability is broader than that; so it might need to be Fourier transforms instead of series. Like Brews, many authors have forgotten what's special about sine waves that motivates such decompositions (for example, this one: "Sinusoidal waves are important because they occur in many physical situations..." -- seriously?). Dicklyon (talk) 18:37, 22 April 2012 (UTC)
Weak Support I support mentioning Fourier series in passing, since I think it's a related concept. However I agree that it's useless for explaining what "wavelength" is to someone that doesn't know. If you understand Fourier analysis, you know what wavelength is. But the converse isn't true, so it's useful to have a mention in this article so that someone learning about wavelength here can go on and learn about Fourier. Waleswatcher (talk) 18:20, 22 April 2012 (UTC)
- Oppose: Just because wavelength and Fourier series relate to waves, doesn't imply there is any important connection, even mentioning the Fourier series really isn't essential at all. This conversation on including something non-essential to an article is also fairly silly, for such a petty issue. Why include it? Adding it to the "see also" section is plenty if you're that desperate to mention it in passing. That bunch of maths given right at the top is needless for understanding wavelength.
- A much simpler mathematical approach to understanding how the wavelength is associated with the phase of a wave is the fractions of a wave cycle:
- where x is a spatial length, is a phase angle, is a time lag/lead. This formula is not in the article, though would be more instructive than a Fourier series becuae it states the angular and time analogues to wavelength. Even so this is not essential for inclusion. Another thing not in the article is that wavelength can be calculated by:
- where N is the number of wave cycles passing through two fixed points of seperation L. More intuitive than the Fourier series, becuase this formula actually says what wavelength is?? Obviously not used in practice, its only for theoretical interest and definitely not essential. F = q(E+v×B) ⇄ ∑ici 23:38, 22 April 2012 (UTC)
- Comment. The proposed text seems very strange. The wording of "The wavelength of a general periodic waveform is related mathematically to its Fourier series expression as a summation of sinusoidally varying waves" is too mysterious. ("related mathematically"? how?) As others have noted, it also emphasizes sinusoidal basis functions, which have nothing at all to do with the wavelength. I think a better alternative would be simply something like "The process of Fourier analysis allows any periodic waveform to be expanded as a superposition of given basis waveforms whose wavelengths divide that of the original waveform." No real opinion on whether that's worth including though. Sławomir Biały (talk) 00:22, 23 April 2012 (UTC)
- To talk about such other decompositions, as Brews did, goes even further off topic. There's a reason why sine waves are special here, and to talk about the decomposition while not saying the reason just misses the boat. Dicklyon (talk) 02:39, 23 April 2012 (UTC)
- Sławomir Biały: Your "better alternative" is a paraphrase of the quote provided. Is it worth including? Of course it is: the whole subject of the mathematics of periodic waveforms is exactly the subject of Fourier series, and periodic waveforms form the definition of wavelength itself. Why wouldn't this connection be pointed out? What is the cost here? Is it that it takes too much room? Nonsense. Is it that the reader already knows about the connection? Maybe in some cases, but certainly not all. Adding the proposed sentence simply is a pointer to wider horizons. There is no cost to including the proposed text. Brews ohare (talk) 04:08, 23 April 2012 (UTC)
- To talk about such other decompositions, as Brews did, goes even further off topic. There's a reason why sine waves are special here, and to talk about the decomposition while not saying the reason just misses the boat. Dicklyon (talk) 02:39, 23 April 2012 (UTC)
- DickLyon: There is no reference to "other decompositions". There was originally, because you raised this issue, but it is there no more. And what is "the reason" for Fourier series that you allude to but never state explicitly? Brews ohare (talk) 04:08, 23 April 2012 (UTC)
- I raised it because what you said about sinusoids would have been equally true and applicable with other decompositions. The point was to illustrate how irrelevant the Fourier series was in your statement connecting it to wavelength, not to encourage you to generalize it. Dicklyon (talk) 06:18, 23 April 2012 (UTC)
- (@Dicklyon's original reply) I think your reply nicely summarizes some of my original reservations as well, that I was not able to clearly articulate. Sławomir Biały (talk) 11:53, 23 April 2012 (UTC)
- Sławomir Biały: It is unfortunate that you prefer to accolade the misconceptions about Fourier series entertained by Dick Lyon rather than address them. "Other decompositions" are not able to represent a periodic function throughput its domain, so Fourier series are special. Moreover, this topic of uniqueness really doesn't matter to the discussion, as the real issue is making this helpful connection between wavelength and the famous, powerful, and seminal apparatus of the Fourier series. Brews ohare (talk) 13:08, 23 April 2012 (UTC)
- "Other decompositions" are not able to represent a periodic function throughput its domain, so Fourier series are special. ← That's just wrong. Sławomir Biały (talk) 16:43, 23 April 2012 (UTC)
- Sławomir Biały: It is unfortunate that you prefer to accolade the misconceptions about Fourier series entertained by Dick Lyon rather than address them. "Other decompositions" are not able to represent a periodic function throughput its domain, so Fourier series are special. Moreover, this topic of uniqueness really doesn't matter to the discussion, as the real issue is making this helpful connection between wavelength and the famous, powerful, and seminal apparatus of the Fourier series. Brews ohare (talk) 13:08, 23 April 2012 (UTC)
- DickLyon: There is no reference to "other decompositions". There was originally, because you raised this issue, but it is there no more. And what is "the reason" for Fourier series that you allude to but never state explicitly? Brews ohare (talk) 04:08, 23 April 2012 (UTC)
- Oppose. I agree with most of the comment above. FWIW, I might phrase the alternative this way: "Any periodic function, with repetition interval λ, can be represented by a mathematically equivalent sum of sinusoids with repetition intervals λ/k, for k=1,2,...,∞. (See Fourier series)" But it would still seem incongruous in the proposed location. A footnote would be a little better. --Bob K (talk) 01:52, 23 April 2012 (UTC)
- I would object to putting odd content into footnotes. Brews has a long history of doing that, and I'm going to object if anyone tries it here. Dicklyon (talk) 02:39, 23 April 2012 (UTC)
- Bob K: Why is it "incongruous" or as DIck says "odd content" to point out the mathematical apparatus known as Fourier series that is the entire formal treatment of arbitrary periodic waveforms, when it is placed in a subsection about arbitrary periodic waveforms? Shouldn't the unknowing reader be made aware of this mathematical apparatus, and its connection to wavelength? What is the cost of including such a pointer? I see none at all: it is all upside and no downside. Brews ohare (talk) 04:08, 23 April 2012 (UTC)
- The incongruity is that the statement is an abrupt leap into the realm of frequency distributions without an explanation of why that might be useful. We need a seque. For instance: "Two periodic functions with the same wavelength are often compared in terms of their other frequency content." That opens another issue, which is Fourier series or continuous Fourier transform? Personally, I would rather see the continuous transform of one cycle of the waveform. The Fourier series provides only discrete samples of that transform. But if I am now straying off topic, I think that only strengthens my original point about incongruity. --Bob K (talk) 12:29, 23 April 2012 (UTC)
- Bob K: A segue, or maybe a new subsection could lead to an elaboration of the significance of the terms in a Fourier series and what they tell us about the complex waveform they represent. I'd have no problem with that discussion, which could be an elaboration about how wavelength connects to harmonics, and what they mean physically, or how they distinguish an oboe from a violin. However, DickLyon would never accept this as a valuable subtopic.
- For the proposed text, however, a much more limited purpose is entertained. It is simply to point out that wavelength is a fundamental concept in the powerful, widely used, and seminal apparatus of Fourier series. The insertion is a heads up, not a discussion. It connects Fourier series to wavelength simply and directly through a quotation. The insertion is proposed to appear in the subsection on arbitrary periodic waveforms, which is the subject of Fourier series, so further introduction seems unnecessary to me. Brews ohare (talk) 13:33, 23 April 2012 (UTC)
- I just don't find that approach interesting enough to get my vote. A more interesting approach is to start with a problem like analyzing the difference between an oboe and a violin, playing the same chord (if that's the right word... I'm not a musician). Point out that they would have the same wavelength, but different waveforms. That leads to the subject of harmonics and harmonic distributions and to Fourier series, if you want to take it that far. That said, this might not be the most appropriate place in Wikipedia for that information. And it might already exist someplace else that can simply be Wikilinked. --Bob K (talk) 00:56, 24 April 2012 (UTC)
- No meaningful connection between wavelength and sinusoidal Fourier series components has yet been mentioned. Dicklyon (talk) 04:14, 23 April 2012 (UTC)
- Dick: It is pointed out that the sinusoids entering the Fourier series have wavelengths that are integral fractions of the wavelength of the periodic waveform. That is mathematically meaningful. If you were to allow further digression, the subject of harmonics could be discussed, which would be physically meaningful as well. The topic of determining wavelength for a waveform in a noisy background by fitting it to the best-wavelength Fourier series could be examined. Perhaps you have some additional insights that you would be happy to see added? Brews ohare (talk) 05:15, 23 April 2012 (UTC)
- No meaningful connection between wavelength and sinusoidal Fourier series components has yet been mentioned. Dicklyon (talk) 04:14, 23 April 2012 (UTC)
Summary
To this point, objections to including the proposed text take on two forms: (i) Fourier series do not illuminate the concept of wavelength, a statement I agree with, and (ii) Fourier series is a "strange", "incongruous" and "odd" addition to an article on wavelength.
No-one seems brave enough to tackle the point here: the formal apparatus for dealing with arbitrary waveforms periodic in space is the Fourier series, no less. Indicating this point may make some readers of this subsection on arbitrary periodic waveforms aware of this hugely significant mathematical framework.
Addressing the connection to Fourier series involves a minor addition of a sourced one-line quotation used to associate Fourier series with wavelength, the topic of the article. It is a no-cost addition to the article. Its accuracy is not in question. Its value to some is not in question. It is of very minor length. So what is the problem here? Brews ohare (talk) 04:35, 23 April 2012 (UTC)
- The problem is that you haven't yet made a sensible connection. If your assertion is true that the formal apparatus for dealing with arbitrary waveforms periodic in space is the Fourier series, no less, then show us that with a source that says why, and maybe we'll be getting close to a useful connection. Actually, it's not true in general, but if you find the domain within which it is sort of true, you'll be on a good path to understanding how sinusoids relate to wavelength, and then maybe be able to say something sensible. Dicklyon (talk) 06:16, 23 April 2012 (UTC)
- DickLyon: Your claim that "no sensible connection" has been made between Fourier series and wavelength seems to ignore the quote provided:
- "Fourier analysis is a mathematical method of expressing any periodic function with wavelength λ as a sum of sinusoidal functions whose wavelengths are integral fractions of λ (i.e. λ, λ/2, λ/3, etc.)"
- Doesn't this statement make the connection between Fourier series and wavelength? Of course it does. Brews ohare (talk) 12:55, 23 April 2012 (UTC)
- DickLyon: Your claim that "no sensible connection" has been made between Fourier series and wavelength seems to ignore the quote provided:
- Dick, apparently you doubt the assertion that the formal apparatus for dealing with arbitrary waveforms periodic in space is the Fourier series. It is completely true for arbitrary spatially periodic waveforms. I know you like to think of it as a subset of various methods for series expansions, which it is, but it is the subset that deals with arbitrary spatially periodic waveforms: none of the other generalized Fourier series deals with an arbitrary spatially periodic waveform over its entire domain. If you wish to take issue with this observation, please let me know. And please avoid commenting upon my abilities and good sense. Brews ohare (talk) 12:33, 23 April 2012 (UTC)
- To repeat what was said before:
- Generalizations of Fourier series include, for example, Fourier-Bessel series and Fourier-Legendre series.[Item 1] However, most often these generalized series represent the function over a finite interval, say one wavelength, and do not represent the periodic function throughout its entire domain.[Item 2] The length of the chosen interval appears in the analysis, but the concept of wavelength, or spatial periodicity, is not fundamental to these generalized Fourier series.
- Items
- ^ Gerald B Folland (2009). "Contents". Fourier Analysis and Its Applications (reprint of Wadsworth & Brooks/Cole 1992 ed.). American Mathematical Society. pp. ix ff. ISBN 0821847902.
- ^ Generalized Fourier series expressions most often are based upon orthogonal functions defined over a finite interval, for example, solutions to the Sturm-Liouville equation in an interval [a, b], so these expansions do not represent the periodic function outside this selected interval.
The point
The question of uniqueness of Fourier series in describing arbitrary spatially periodic waveforms is a nicety. Uniqueness establishes that other decompositions of a waveform do not employ the notion of wavelength, or spatial periodicity of the arbitrary but periodic waveform. The real point, however, is that Fourier series is a huge topic, a powerful apparatus, and it uses wavelength. The reader should be alerted to this connection. That "heads-up" is the point of the proposed text. Brews ohare (talk) 14:22, 23 April 2012 (UTC)
- (edit conflict)And Fourier series also uses trigonometry. And calculus. And infinite series. And basic arithmetic. And real numbers. Does that mean (as it's "a huge topic, a powerful apparatus") that Fourier series should be discussed in all those articles? No. But they are much more important to an essentially mathematical topic, not least as Fourier series do not depend on wavelength at all.
- And please stop editing and re-editing your own comments. Use the 'Show preview' button to get it right first time, write what you mean to write once and wait for replies. It creates work for other editors if they find the post they are replying to have changed (often multiple times) in the few minutes after they were posted.--JohnBlackburnewordsdeeds 14:55, 23 April 2012 (UTC)
- Blackburne, what you fail to recognize is that wavelength and spatial periodicity basically are the same concept, and also the fundamental concept underlying Fourier series. That makes it useful to point out this connection, while your topics do not qualify. So sorry that my editing to get the wording right has proved so distracting to your thought. Brews ohare (talk) 15:26, 23 April 2012 (UTC)
- BTW, Blackburne, your observation that "Fourier series do not depend on wavelength at all" is a flat contradiction of the quotation in the proposed text.
- Perhaps you could point out which part of Fourier series#Definition uses the wavelength? Or where in that article it is mentioned? It isn't mentioned because Fourier series do not depend at all on wavelength.--JohnBlackburnewordsdeeds 15:47, 23 April 2012 (UTC)
- Blackburne, what you still fail to recognize is that the topic here is periodicity in space, as that is where wavelength has a role. It happens that Fourier series apply to a function periodic in any variable. However, when applied to spatial periodicity, the Fourier series directly incorporates wavelength as pointed out in the quotation in proposed text. Please try to adopt the appropriate context. Brews ohare (talk) 15:54, 23 April 2012 (UTC)
- Perhaps you could point out which part of Fourier series#Definition uses the wavelength? Or where in that article it is mentioned? It isn't mentioned because Fourier series do not depend at all on wavelength.--JohnBlackburnewordsdeeds 15:47, 23 April 2012 (UTC)
- What about time period and Frequency of wave oscillations? And about wave phase? Why not include the Fourier series there also for the hell of it, because waves are also periodic in time and phase angle???
- If you actually think about the reader (more likely to be someone less interested in maths, but would like to know about wavelength), rather than assuming they will switch on to the Fourier series when they read "a superposition of waves each with wavelength (really!) can be superimposed to form a wave with periodicity in space, can be mathematically analysed as a Fourier series", isn't it more likely that they would like to find out what the wavelength is and what its for and why its measured (etc)?? will they not be thinking "so a Fourier series describes periodic functions and waves can also be described by periodic functions... isn't that a description of superposition?" (or words to that effect) ?? It serves the reader no purpose - again adding a see also link or template will indicate "for further descriptions of (spatial) periodicity and wavelength see the main articles Fourier series" (or words to effect).
- How many editors have said the same thing? Isn't it obvious? Stop accusing other editors of "failing to understand" - you are the one who seems to fail to understand the blatant consensus against adding a fluttery statement on the Fourier series, and are too ignorant of the reasons they provide. F = q(E+v×B) ⇄ ∑ici 16:42, 23 April 2012 (UTC)
- F = q(E+v×B): You bring up the topic of including a heads-up to Fourier series in other articles. That may be appropriate in some cases and in others not. As with these other topics, its appropriateness in the context of wavelength should be judged on its own merits.
- You also suggest that not all readers will find a heads-up to Fourier series is of interest, but of course, some will. There is benefit in presenting this heads-up for those that do find value in it. All it costs is a few lines that the disinterested can skip, if they like.
- I'd agree that many objections to this insertion have been registered. Some objections are based upon misconceptions about the nature of Fourier series, and can be discounted. The remaining objection is the one you raise: Fourier series is a "strange", "incongruous" and "odd" addition to an article on wavelength. You call it a "fluttery" statement, which I can only assume means something similar.
- All the proposed text is, actually, is a heads-up pointing to Fourier series for spatially periodic waveforms. It's brief, it's accurate, and it's helpful to some. There is no downside to its inclusion. Brews ohare (talk) 17:11, 23 April 2012 (UTC)
- Perhaps we can compromise on this: part of your current proposal
- "summation of sinusoidally varying waves with wavelengths related to the wavelength of the periodic waveform itself:"
- is quite a mouthfull and not much help, a typical reader could get lost. Perhaps change the statement to:
- "Sinusiodal waves can superimpose and create a spatially periodic waveform, which can be mathematically described by a Fourier series. The wavelength in this case corresponds to the (spatial) period in the Fourier series, i.e. the spatial interval for which one cycle of the function repeats itself."
- Perhaps we can compromise on this: part of your current proposal
- from a physical perspective (or words to that effect). That way the series is mentioned in passing with no obfuscation. Perhaps the best mathematical perspective (if preferred by others) would be Sławomir Biały's version above. I would simply prefer to add
- {{Hatnote|For further descriptions of periodic functions in a more general context to wavelength and phase, see also [[Fourier series]].}}
- at the top of the Mathematical representation section though, since this is more of a sidenote with no interruption of continuity, but will not do so. Not fussed about the reference... F = q(E+v×B) ⇄ ∑ici 17:55, 23 April 2012 (UTC)
- F = q(E+v×B): I have taken a whack at rewording along these lines. Brews ohare (talk) 19:28, 23 April 2012 (UTC)
- at the top of the Mathematical representation section though, since this is more of a sidenote with no interruption of continuity, but will not do so. Not fussed about the reference... F = q(E+v×B) ⇄ ∑ici 17:55, 23 April 2012 (UTC)
- Hrm, your version gives the explanation in 3 sentences, mine in 2, and seems more wordy than necessary. All we need to include Fourier's series is state superposition and representation by the series. Also adding two (very much) identical quotations will not amplify any meaning, I don't think there is any need for the quotations but you may as well keep the references for inline citations.
- Anyway I have made my points, and will no longer participate in this (can't contribute much more anyway - exams). Feel free to compromise further with the others... F = q(E+v×B) ⇄ ∑ici 20:31, 23 April 2012 (UTC)
- Where are these new versions you guys are comparing? And do they say why it would be more useful to use sine waves than, say, triangle waves, for such a decomposition? If not, then it's still quite pointless, isn't it? Dicklyon (talk) 23:22, 23 April 2012 (UTC)
- Dick: I rewrote the contents of the "green box".
- What are the properties of your "triangle waves"? Are your "triangle waves" defined over all space and do they exhibit periodicity with a certain wavelength themselves? Are series expansions using your triangle waves automatically periodic with wavelength λ? I imagine you could construct such a basis set, but it would be a fringe topic compared to Fourier series, and I wouldn't be surprised if to prove their completeness you'd have to reduce them to superpositions of sinsuoids and resort to Fourier's theorem. Brews ohare (talk) 00:35, 24 April 2012 (UTC)
- Triangles work in all respects you've mentioned just as well as sines do. It's not about what's mainstream and what's fringe, it's about what makes sine waves special, and I've told you at least four time in this talk page already. Here, read up: some clues: see "nothing arbitrary" in this Britannica article by Lord Rayleigh, [1], [2], [3], [4], [5]. Dicklyon (talk) 01:23, 24 April 2012 (UTC)
- Where are these new versions you guys are comparing? And do they say why it would be more useful to use sine waves than, say, triangle waves, for such a decomposition? If not, then it's still quite pointless, isn't it? Dicklyon (talk) 23:22, 23 April 2012 (UTC)
DickLyon's sources
Dick Lyon has undertaken to illustrate "what makes sine waves special" by referring to several sources. I have provided a bit more about these sources below. They fall into two groups: those that repeat what is already in the proposed text, and those that require a considerable expansion of this proposal to go into nuances not so far entertained as a possible insert.
1. Rayleigh: Quote: "there is nothing arbitrary in the use of a circular function to represent the waves. As a general rule this is the only kind of wave that can be propagated without a change of form; and even in the exceptional cases where the velocity is independent of wavelength, no generality is really lost by this procedure, because in accordance with Fourier's theorem any kind of periodic wave may be regarded as compounded of a series of such as (1), with wavelengths in harmonical progression."
The contribution of this source to the discussion of the proposed text is only to reinforce what is said in the two quotations. It suggests that in addition to this, more could be said about the propagation of waves in dispersive media, where only a wave of a given wavelength propagates unchanged. If the proposed text is to be expanded upon, that point could be added.
2. Rogalski Quote: "any periodic function can be represented as the superposition of harmonic terms of frequencies f1, f2 = 2 f1, f3=3f1... where f1=1/λ"
The contribution of this source to the discussion of the proposed text is only to repeat what is said there already.
3. Beerends Quote: "By superposition of the fundamental solutions one can usually construct a formal solution... The fundamental solutions describe in relation to one or several variables, sinusoidal functions with frequencies that are an integer multiple of a fundamental frequency. This fundamental frequency already emerges when one calculates the eigenvalues."
This general discussion is what was already alluded to in the above recapitulation of material regarding generalized Fourier series. The example provided later in this section is confined to the finite interval [0, L], and does not apply to a waveform periodic throughout all space. If we were allowed to expand the discussion, such matters could be explored.
4. Osborne This text describes propagation in nonlinear media. While this is an interesting subject, it is not germane to this discussion which concerns simply mentioning Fourier series and its connection to wavelength.
5. Gockenbach This text discusses Fourier series as one example of eigenfunction methods. Again, this matter has been dealt with in the above recapitulation of generalized Fourier series. If DickLyon's point is that Fourier series is one type of eigenfunction expansion, fine. But that point does not illustrate how special Fourier series are, but buries their individuality by suggesting they are merely one of the many finite-interval approaches. Fourier series representation are special because, in contrast, they apply to the entire domain of the periodic function , not just to some finite interval.
6. Smith This text describes the discrete Fourier transform, which is again a wonderful subject that has nothing to do with the present discussion.
In summary, this blizzard of sources adds no new considerations concerning addition of the proposed text, although they provide some avenues to expand upon it. Brews ohare (talk) 03:13, 24 April 2012 (UTC)
New section in article
I've rewritten the article section to try to motivate the sinusoidal decomposition; now there's a reason to mention Fourier series. Feel free to revert if anyone objects. Dicklyon (talk) 01:44, 24 April 2012 (UTC)
- Dick has identified the need to rewrite this section, and has made a first attempt to do so.
- From this revision it now is clear what Dick has been alluding to all along about the "special nature" of the sinusoidal wave: it propagates with unchanging waveform in a dispersive medium. This is the observation made by Lord Rayleigh. This fact is made a segue to Fourier series, introduced as a handy way to handle propagation of a disturbance propagating in a dispersive medium by treating it as a superposition of waves of a single frequency, so each component propagates unchanged, even though the shape of the propagating disturbance varies in time and in space, and has no associated wavelength. Dick uses the wording "linear medium", in preference to "dispersive medium", a poor choice as the topic "non-linear" media (also discussed) is not in contrast to "linear" medium in this sense of the word.
- Now, this property of sine waves in dispersive media is an interesting point to raise and it does introduce the notion of Fourier series, more properly, the Fourier transform because periodicity of the waveform is not part of this discussion.
- However, this singular property of sinusoids is a point contradicted later in Dick's revision when it is pointed out that periodic waveforms that are not sinusoids also can propagate in some media and they do have a wavelength. The discussion of wavelength for general waveforms has been emasculated by eliminating the generalized definition of wavelength for an arbitrary waveform f(x−vt+λ)=f(x−vt), the historical and mathematical foundation of Fourier series. The removal of the general definition eliminates this simple segue to the proposed text.
- The new version is a start. However, this section remains a disjointed assembly of disconnected topics. If the general definition of wavelength is not restored, it remains unclear how the proposed text might fit into the ultimate revision. Brews ohare (talk) 12:22, 24 April 2012 (UTC)
- I've rearranged Dick's revision, changed some wording, and inserted a version of the proposed text. I think the result is less disjointed. Brews ohare (talk) 13:58, 24 April 2012 (UTC)
- I believe you added several kilobytes and destroyed the point of it. I'll wait and see if anyone else cares to comment. Dicklyon (talk) 15:38, 24 April 2012 (UTC)
- Dick: I am alarmed by your statement that this version has destroyed the point of yours. IMO it is mainly a rearrangement of your own wording, plus the restoration of the periodicity condition that you removed. I added the historical remark of the end paragraph. Brews ohare (talk) 16:40, 24 April 2012 (UTC)
- The consensus above is against adding the digression on Fourier series, however worded, so that should not be there. Other than that it has destroyed the logic and structure of the section changing it from something that read well to a mess, for no good reason (at least no reason given in the overly terse edit summaries).--JohnBlackburnewordsdeeds 17:20, 24 April 2012 (UTC)
- Blackburne: I have made a conscientious effort in my rewrite as explained carefully in my critique of DickLyon's initial effort immediately above. The logic of the section is exactly as Dick wrote it. The reorganization mainly changes things to put each topic in one place instead of having them scattered about. Beyond that I restored the periodicity definition that was there originally, and changed or added a few sources. As for the connection to Fourier series: I believe that has to be assessed in the new context. Brews ohare (talk) 17:31, 24 April 2012 (UTC)
- The consensus above is against adding the digression on Fourier series, however worded, so that should not be there. Other than that it has destroyed the logic and structure of the section changing it from something that read well to a mess, for no good reason (at least no reason given in the overly terse edit summaries).--JohnBlackburnewordsdeeds 17:20, 24 April 2012 (UTC)
- Dick: I am alarmed by your statement that this version has destroyed the point of yours. IMO it is mainly a rearrangement of your own wording, plus the restoration of the periodicity condition that you removed. I added the historical remark of the end paragraph. Brews ohare (talk) 16:40, 24 April 2012 (UTC)
- I believe you added several kilobytes and destroyed the point of it. I'll wait and see if anyone else cares to comment. Dicklyon (talk) 15:38, 24 April 2012 (UTC)
Brews, thanks for trying, but it's hard to make a sensible rewrite when you still don't quite understand it and you have some favorite ideas you want to cram into it. I may take another stab at it. Some problems:
- 1. The opening generality "In general, the propagation of a disturbance takes a different form in different media, for example, media where the velocity of propagation depends upon wavelength (dispersive media) or upon the amplitude of the wave (non-linear media)" makes the next sentence "The wavelength is functionally related to frequency by the physics of the medium" meaningless or false.
- 2. The sentence "The same is true for a waveform that is composed as a combination of many sinusoids, all of the same wavelength but differing in amplitude, or phase, or both" has no useful role there, just a distractor.
- 3. "For example, sinusoids are simple solutions (eigensolutions or eigenfunctions)" got disconnected from "propagates with no shape change" which is essentially the definition of eigenfunctions
- 4. "Just which wavelengths can contribute to a waveform is decided by boundary conditions. (See the section on standing waves.)" is an unnecessary tangent
- 5. "Under special circumstances, waves other than sinusoids propagate with unchanging shape and constant velocity, called traveling waves" is a complete mixup of unrelated concepts.
- 6. "Periodic waves have a well-defined wavelength" is false in general; that's why you find this application of "wavelength" to arbitrary periodic shapes so rarely in source.
- 7. The sentence "For a fixed-shape waveform that also is periodic, its wavelength λ is defined mathematically in one spatial dimension x by the definition of periodicity" and following formula do not define the wavelength, since the formula would also be satisfied by all integer multiples of the wavelength.
- 8. The reintroduction of Fourier series as a way to analyze spatially periodic waves again completely misses the point of why one would want to decompose waves into sinusoids, and repeats a statement that would be no less true or relevant if done with triangle waves, and is backed up by a source that is equally clueless. If you apply Fourier series to time-periodic waves, though, it's easy to see how it is useful in linear media, and not just in the rare corner case of non-dispersive media; since you didn't get that, you destroyed it.
Dicklyon (talk) 21:41, 24 April 2012 (UTC)
- Dick: To avoid repeating each of your points, I have numbered them.
- 1. The dispersion relation definitely is related to the physics of the medium. A common example is the phonon in crystals where the propagation of various vibrational modes, for example, is determined by solving the problem of crystal vibrations (the medium). The dispersion relation for an acoustic mode in, say BaTiO3 is different in another medium, say GaAs. I think you are aware of this matter, so I guess your statement has some other meaning that you might try to explain further.
- 2. This "distractor" came directly following the self-same segue from the source you provided to Lord Rayleigh. If it didn't distract him, I can live with it.
- 3. The definition of an eigenfunction is that it satisfies an operator equation with an eigenvalue, as in O f = λ f. It is not defined in terms of propagation.
- 4. Perhaps so. It seemed to me to answer a question that pops up in this context, and it connects this section to the previous one. In any event, as a general matter, eigenvalues cannot be determined without boundary conditions, and if the boundary conditions change so do the eigenvalues and the eigenfunctions.
- 5. There is no mix-up here. It's not a mix-up, but a shared property. Following your lead (and Lord Rayleigh's) sine waves are interesting because they propagate unchanged in certain media. By the same token, then, traveling waves are interesting because they also propagate unchanged. It seems that cnoidal waves actually are superposed to form more complex traveling waves, a clumsy parallel to sines and cosines. However, I think this comparison is seen more accurately as a segue than as a deep parallel.
- 6. A periodic wave, as opposed to some approximation of a periodic wave, always has a wavelength as defined by f(x-vt+λ)=f(x-vt). If the wave doesn't satisfy this periodicity condition it is not a periodic wave. At least, this is the case in one dimension. Maybe you have something else in mind?
- 7. Fine, we'll make it the least value of λ for which this is the case.
- 8. This is a really interesting remark, and one we might focus upon as it seems to me to be the real cause of our failing to understand each other. To begin, I suppose we are talking about the last few lines beginning with a reference to Joseph Fourier. If that is the case, these words can be interpreted as no more than historical background, as they make no claims about the applicability of Fourier series. As an historical note, you might object that it is a digression, but that's about it.
- To pursue the matter further, you say: "If you apply Fourier series to time-periodic waves, though, it's easy to see how it is useful in linear media, and not just in the rare corner case of non-dispersive media; since you didn't get that, you destroyed it." Now, IMO this point was already belabored in the opening paragraphs where it was pointed out that sine waves propagate with fixed form enabling the application of the Fourier transform to the propagation of arbitrary waveforms. These words were cribbed from your first draft, except the more accurate Fourier transform is alluded to because the context is general waveforms and not spatially periodic ones. Are you clear that Fourier series represent periodic functions with the property f(ξ) = f(ξ+λ) for all values of the argument ξ and not restricted to representing f only for ξ sitting inside some selected interval [a, b]? In particular, we can use ξ=x-vt.
- I await your further comments, Dick. Brews ohare (talk) 23:32, 24 April 2012 (UTC)
- For reference, here is the section as I had it, which I think was much more correct, and not subject to the "problems" you are imagining:
The concept of wavelength is most often applied to sinusoidal, or nearly sinusoidal, waves, because in a linear system the sinusoid is the unique shape that propagates with no shape change – just a phase change and potentially an amplitude change. The wavelength (or alternatively wavenumber or wave vector) is a characterization of the wave in space, that is functionally related to its frequency, as constrained by the physics of the system. That is, sinusoids are the simplest solutions (eigensolutions or eigenfunctions) of linear partial differential equations that describe linear physical media, and the frequencies and wavelengths are related by the dispersion relations of the equations or those media. From these simple traveling wave solutions, more complex solutions can be built up by superposition.
In the special case of dispersion-free and uniform media, waves other than sinusoids propagate with unchanging shape and constant velocity. In certain circumstances, waves of unchanghing shape can also occur in nonlinear media; for example, the figure shows ocean waves in shallow water that have sharper crests and flatter troughs than those of a sinusoid. Large-amplitude ocean waves with certain shapes can propagate unchanged, because of properties of the nonlinear surface-wave medium.[1] An example is the cnoidal wave, a periodic traveling wave named because it is described by the Jacobi elliptic function of m-th order, usually denoted as cn(x; m).[2]
If a traveling wave has a fixed shape that repeats, it is a periodic wave.[3] Such waves may have a well-defined wavelength even though they are not sinusoidal. As shown in the figure, wavelength is measured between consecutive corresponding points on the waveform. Mathematically, the amplitude f of an unchanging waveform moving with a velocity v can be expressed as f(x − vt), with x = position and t = time. The amplitude at location x + Δx at time t + Δt is the same as that at location x at time t, if Δx and Δt are related by Δx = vΔt.
In more general linear media (that is, dispersive media), a wave that is periodic in time will not necessarily repeat in space, so may not have a well-defined wavelength. Such waves are typically analyzed into sinusoidal waves via the Fourier series, so that the different propagation speeds and wavelengths of their different frequency components can be separately handled. To an observer at a fixed location, the amplitude varies in time and repeats itself with a certain period, T. During every period, an integer number of each component wavelength of the wave passes the observer, but with different relative phases at different observer locations.
- I don't have time to try to explain it to you again. It may not be perfect, but what it says about Fourier series is much more sensible than anything you've come up with. Dicklyon (talk) 05:24, 25 April 2012 (UTC)
- Dick: You haven't tried to explain as far as I can see. You simply assert (based upon your personal opinion alone) that my rewrite has problems, which it does not, and that yours is better, which it is not. The way forward is to address the responses I've made to your points above, and come to agreement.
- Your paragraph repeated by you above contains nothing of value that is not in my rewrite, and my rewrite avoids several problems of wording in your paragraph, as noted earlier, but not responded to by you. You may be too busy to engage, and if so that is unfortunate. But simply reasserting the past formulation without addressing comments upon it is a waste of what little time you have to spend. Brews ohare (talk) 05:40, 25 April 2012 (UTC)
- It may be that a difficulty with getting across what you want to say stems from using the wrong subsection to explain it: instead of the general waveforms section the topic of dispersion and the special advantages of sine waves might go better in the general media section where dispersion is described. What do you think? Brews ohare (talk) 14:26, 25 April 2012 (UTC)
- I don't have time to try to explain it to you again. It may not be perfect, but what it says about Fourier series is much more sensible than anything you've come up with. Dicklyon (talk) 05:24, 25 April 2012 (UTC)
I have restored Dicklyon's version as a far better version of the section, as explained in detail above, and per the consensus in the RfC on the Fourier digression. I would note that the problem with the changes was much amplified by the lack of reason given for them, with edit summaries consisting often single words or entirely missing, and never giving a proper rationale for the changes, making it impossible to tell what the edits were intended to do. It was only possible to judge them by the actual result which was a far worse rendering of the section.--JohnBlackburnewordsdeeds 14:43, 25 April 2012 (UTC)
- That seems sensible. John, if you or anyone wants clarification of any of the issues raised in the various numbered points above, I'll be happy to discuss. I know from experience here and elsewhere that discussing them just with Brews will go nowhere useful. Dicklyon (talk) 15:01, 25 April 2012 (UTC)
- If you want to improve the section, I'd pull out everything about nonlinear into a final paragraph, as it does confuse when inappropriately mixed up with the linear and sinusoidal decomposition bits. Dicklyon (talk) 15:03, 25 April 2012 (UTC)
High-handed actions
The recent wholesale reversion by Blackburne removed the baby with the bathwater, deleting updates of references and some useful rewording of some completely uncontroversial matters. In addition, of course, and as accoladed by Dick Lyon, no attempt has been made to address substantive issues carefully laid out above and never responded to. Brews ohare (talk) 15:10, 25 April 2012 (UTC)
I've undertaken to repeat the edits that are non-controversial and correct some of the more obvious problems here. I have done this with individual edits so they can be examined one by one by Blackburne, instead of repeating his wholesale mindless reversion. Brews ohare (talk) 19:30, 25 April 2012 (UTC)
I've also restored two unrelated subsections Blackburne removed that were not involved in the discussion with DickLyon, possibly the unfortunate victims of an enthusiastic response too hastily implemented. Brews ohare (talk) 17:30, 26 April 2012 (UTC)
Dick Lyon's recent incorrect changes
Dick, you made these reversions that I think are incorrect.
- They include the following statement:
- "In more general linear media (that is, dispersive media), a wave that is periodic in time will not necessarily repeat in space, so may not have a well-defined wavelength. Such waves are typically analyzed into sinusoidal waves via the Fourier series, ..."
The one-line edit justification given is "period waves are analyzed by Fourier series".
Please notice the first part of the section states the waves under discussion may not be periodic. As you know, and as you have provided in your one-line edit rationale, Fourier series apply to waves that are periodic. Consequently, to cover the general case under consideration that includes non-periodic waveforms, one must resort to the Fourier transform, which treats waveforms of all kinds.
Consequently, Dick, your change from Fourier transform back to Fourier series has introduced an error. Probably Fourier integral would be better, but that is a redirect at the moment. Brews ohare (talk) 05:48, 26 April 2012 (UTC)
- In a second change, you have written:
- "If a traveling wave has a fixed shape that repeats, it is a periodic wave.[20] Such waves may have a well-defined wavelength.
It is evident from the figure referred to in the paragraph containing this remark that spatial periodicity is meant, not time periodicity. Now, a spatially periodic wave always has a wavelength, by the definition of spatial periodicity, so the word "may" should be removed, returning this text to its form before your revision. Brews ohare (talk) 05:48, 26 April 2012 (UTC)
- Finally, in this reversion you moved these sentences:
- "To an observer at a fixed location, the amplitude varies in time and repeats itself with a certain period, T. During every period, an integer number of each component wavelength of the wave passes the observer, but with different relative phases at different observer locations."
out of the paragraph describing periodic waveforms to the paragraph describing general waveforms in dispersive media. Evidently in dispersive media all components of the waveform move at different speeds v=λf with f=f(λ) governed by the possibly (in general at least) arbitrary dispersion relation for the medium. There is no reason why each component should move an integer number of wavelengths in a given fixed time interval, including the period T. Consequently this change introduces an error. Brews ohare (talk) 05:57, 26 April 2012 (UTC)
- You seem to be very confused and wrong on each of these points. Please read them again. Dicklyon (talk) 14:54, 26 April 2012 (UTC)
- Dick: I have made very obvious points and you have not responded to them. IMO, you are in error, and with no specific indication from you why you think otherwise, I am inclined to think you have no sensible response to my points and, in fact, have not read them at all. Brews ohare (talk) 15:19, 26 April 2012 (UTC)
- Brews, I don't know if this is what Dick has in mind, but regarding the statement that general periodic waves may have a well-defined wavelength, note that the assignment of the term "wavelength" to the spatial period of non-sinusoidal waves is uncommon, and that there are other possible definitions of the term "wavelength" for such waves. The example that comes to mind is a modulated carrier wave. Such waves are of course not generally exactly periodic, but it is certainly possible to construct a perfectly periodic modulated signal. The wavelength of such a signal is, by conventional definition, the wavelength of the carrier wave, not the spatial period of the modulated signal. The statement about general periodic waves should probably be more strongly qualified, to indicate the rarity of this usage of the term.--Srleffler (talk) 17:31, 26 April 2012 (UTC)
- Srleffler: You discuss ambiguity of the wavelength assigned to a modulated wave, where "wavelength" could refer to the carrier wave rather than to the true periodicity of the waveform. A more extended discussion like that appears in the subsection on envelopes.
- The discussion of this particular paragraph is accompanied by a figure that shows the context intended for it. Also, the following math defining the periodicity condition in one dimension also makes clear what the subject is here.
- To use the word "may" in this paragraph to refer to a possible usage of "wavelength" markedly different from the topic of this paragraph places a perplexing obligation upon the single word "may". Brews ohare (talk) 17:54, 26 April 2012 (UTC)
- I've replaced the word "may" by a footnote referring to some exceptions, but not to different usages like carrier wavelengths, which I feel is covered in other subsections, notably the discussion of "local wavelength".
- I've changed Fourier series to Fourier integral, which I believe is the correct terminology.
- I've moved a modification of the statements about integral wavelengths passing an observer in one period to the section on spatially periodic waves, where it does apply. If Dick Lyon believes some version of this wording will work in the dispersive section, I suggest he put the appropriately worded sentence or two there, and not feel obligated to remove the discussion from the periodic section. Brews ohare (talk) 19:10, 26 April 2012 (UTC)
- The problem is that the periodic-in-space stuff is an odd corner case, in which Fourier series decomposition does nothing useful. But the periodic-in-time case, in more general linear media, leads to something worth saying. It seems that Brews hasn't yet been able to see that difference. As for wavelength of periodic waves being well defined, that's highly suspect on a number of levels; trying to introduce broad mathematical generalities like a least period naturally drags in all the weird excpetions, too, like functions that are periodic but have no minimum period. Dicklyon (talk) 20:20, 26 April 2012 (UTC)
- Hi Dick: You removed the following text, which has been in the article in one form or another for a long time:
- "Mathematically, a periodic wave in one spatial dimension x satisfies the periodicity condition: f(x−vt+λ) = f(x−vt). If this periodicity requirement is satisfied by λ, it also is satisfied by 2λ, 3λ, and so on. The least such λ is the wavelength of f.1
- Source
- 1. Eric Stade (2011). Fourier Analysis. John Wiley & Sons. p. 1. ISBN 1118165519.
- "Mathematically, a periodic wave in one spatial dimension x satisfies the periodicity condition: f(x−vt+λ) = f(x−vt). If this periodicity requirement is satisfied by λ, it also is satisfied by 2λ, 3λ, and so on. The least such λ is the wavelength of f.1
- Hi Dick: You removed the following text, which has been in the article in one form or another for a long time:
- I added the source (also removed), which is only one of hundreds that could be cited. This little piece of math says formally what is described in the figure accompanying this paragraph.
- Now, this definition goes all the way back to Joseph Fourier, and has been a part of all books on Fourier series for centuries. It states the subject that Fourier series describes. Its exploration by Fourier has huge historical and mathematical significance. To claim this basic statement is to be removed because it is only mathematical pretense strikes me as bizarre.
- By the same token, I cannot fathom why you insist upon introducing Fourier series as the basis for expansion of non-periodic waveforms, when it is entirely trivial to show that is not the case, quite apart from its basis in the periodicity condition you deleted. You can also read the WP article Fourier series that says the same thing. The Fourier integral, on the other hand is not so restricted, as I am sure you are aware. So why not use the correct term?
- Dick, what is the matter? Brews ohare (talk) 22:27, 26 April 2012 (UTC)
- The matter is that belaboring a spatially periodic wave is completely trivial, and adds undue weight to the rare application of the notion of wavelength to arbitrary spatially periodic functions. And the temporally periodic wave is what is typically decomposed by a Fourier series so that its sinusoidal components of different wavelengths and frequencies can be separately analyzed. I have not written anything about waves that are not periodic (in time). It seems that others understood; why haven't you? Dicklyon (talk) 03:42, 27 April 2012 (UTC)
Dick, you make confused explanations that fail to distinguish between time and spatial periodicity, and simply use "periodicity" to mean time periodicity even in paragraphs where spatial periodicity is the topic.
However that may be, the incorrect usage of "Fourier series" where "Fourier integral" clearly applies is simply wrong, and can't be justified by this confused usage. A "temporally periodic wave" can be expressed as a Fourier series if regarded only as a function of time, but the context (as very clearly set forth in the discussion) is the change in spatial form of a general disturbance as it propagates in space in a dispersive medium. That analysis cannot be based upon a Fourier series, but demands a Fourier integral. If you wish to expand this talk page discussion to explore this mathematical analysis beyond all possibility of slip-shod argument, perhaps progress can be made. Let me know. Brews ohare (talk) 06:24, 27 April 2012 (UTC)
- Brews, you should re-read the paragraph of the article, about which you are complaining. It explicitly specifies that it is talking about waves that are periodic in time. It is you who have failed to grasp the context. Your points below are irrelevant because you are thinking about the wrong special case. --Srleffler (talk) 17:10, 27 April 2012 (UTC)
Upon further reflection, even at a fixed location where a dispersing waveform is strictly a function of time, it will not be a periodic function of time, and therefore cannot be analyzed using a Fourier series in time because a Fourier series is necessarily periodic. The dispersing wave does not repeat in space, so at a fixed location one does not see a repeating pattern as the waveform travels past a fixed point. So the point remains that one needs a Fourier integral to capture the behavior, whether one seeks to capture it in space at a certain time or in time at a certain location. Brews ohare (talk) 15:26, 27 April 2012 (UTC)
- There are certainly special cases where the waveform can be a periodic function of time at a fixed location. Moreover, these are important special cases, since many waves are periodic in time at their source. One needs a Fourier integral for waves that are not periodic in time, but that is not what is being discussed in the paragraph in question.--Srleffler (talk) 17:10, 27 April 2012 (UTC)
Consequently, the present statement in the article about a wave in a dispersive medium:
- "To an observer at a fixed location, the amplitude varies in time and repeats itself with a certain period, T."
is an egregious misstatement. Brews ohare (talk) 16:22, 27 April 2012 (UTC)
- Again, you missed the crucial phrase in the first sentence of the paragraph: "a wave that is periodic in time". --Srleffler (talk) 17:10, 27 April 2012 (UTC)
- Note also that a wave that starts out periodic in time at one location will be periodic in time at other points as well, but not with the same period. One can decompose the initial wave into sinusoids (via Fourier series), which then propagate at different velocities. At any given location there will be a period for the wave as a whole, which is an integer multiple of the periods of each component wave (but not the same integer for each). As the article states, "During every period, an integer number of each component wavelength of the wave passes the observer..."--Srleffler (talk) 17:23, 27 April 2012 (UTC)
- Srleffler: Thanks again for attempting an explanation that DickLyon apparently could not attempt himself. To try to visualize what you are telling me, I imagine a driver producing a square pulse at regular intervals, so it is a time-periodic driver. At distances close to the driver, a square pulse is received. At greater separations, the faster components of the square pulse will move (let's say) faster than the rest, so the pulse begins to lose definition. The faster components provide a precursor and a rounded pulse arrives later. Depending upon the size of the separation of the observer from the driver, a precursor may even overtake an earlier rounded pulse, resulting in something quite different as the slow components of pulse leaving at time tj are overtaken by the fast components of pulses leaving at time tj+1, tj+2...
- So I guess with some work I could figure out how this whole thing evolves. I can imagine that because a steady state is envisioned, whatever the form of the disturbance at a distance, it will have no obvious resemblance to the initiating pulse, and it will repeat at some interval in time that may be actually many periods of the driver, depending how many driver pulses contribute high speed precursors to the location of the slowly propagated, rounded form of the filtered square wave. So a Fourier series in time could represent this situation at a fixed location, contrary to my earlier intuition, albeit with a period set not solely by the driver, but also depending on the distance from the driver and the dispersion relation determining how speed varies with wavelength.
- So I guess that leaves me with a few questions:
- (i) Is it helpful to be told that at a fixed point in space the behavior can be represented as a Fourier series in time with a hard-to-find period? Shouldn't a Wavelength article focus on the spatial aspects?
- (ii) Is the discussion presented likely to be clear to a representative reader as it now stands?
- (iii) A question no-one can answer: why wouldn't Dick Lyon try to explain this matter, rather than sneering about it?
- In any event, thank you for your explanation. Brews ohare (talk) 21:53, 27 April 2012 (UTC)
- Every time I explained it, you pronounced it incorrect, and changed it to something wrong. Eventually I gave up. Dicklyon (talk) 05:18, 28 April 2012 (UTC)
Restoration of two subsections
In a reversion by Blackburne intended to replace a few paragraphs written by myself with DickLyon's version, two unrelated subsections not subject to dispute were removed with their accompanying figures and sources. I have restored these two subsections. Brews ohare (talk) 17:23, 26 April 2012 (UTC)
- As I and others told you before at Talk:Wavelength#phase_and_group_velocity the stuff on envelope waves is too odd to rate coverage here; especially with that source that says the envelope wave propagates at a velocity that depends on the envelope wavelength, which just shows that author to be clueless or careless. Dicklyon (talk) 03:53, 27 April 2012 (UTC)
- Dick: This material about envelopes was placed here long, long ago, after some discussion, and is not something recently introduced by myself. It serves a purpose, although you don't like it. Brews ohare (talk) 06:37, 27 April 2012 (UTC)
Brews, please stop
Brews, we've got 120 KB of new talk this month (April 2012) due to your attempts to add trivial and inappropriate content to this article, with zero support from any other editor, while you ignore everything we try to tell you and keep returning to your idiosyncratic ways. Are you trying to repeat your mess of June/July 2009, where we had over 200 KB of talking past each other before you went away? If you can find anyone at all who agrees with you, maybe there will be something to talk about. For now, there's not. Please stop. Dicklyon (talk) 05:19, 27 April 2012 (UTC)
- Dick: It's really silly to blame the extensive talk page discussion here on me, when it is almost all due to your unwillingness to engage in actual dialogue, ignore all comments, refuse to explain yourself, and make silly assertions based upon your own opinion and nothing else, and instead castigate my efforts, as this little addition of yours examples. Brews ohare (talk) 06:10, 27 April 2012 (UTC)
- Brews, it is you that started an RfC after you refused to accept Dicklyon's arguments, then have continued to argue after the consensus in the RfC was clearly against you. So yes, the extensive talk page discussions were all initiated by you. You have had all the arguments above. That you keep rewriting your arguments only makes things worse. It does not mean that every time you do so editors have to anew refute them point by point. You have had your chance to convince people, you've failed, repeatedly. Now please stop.--JohnBlackburnewordsdeeds 10:42, 27 April 2012 (UTC)
- Brews they have repeatedly explained things to you, which is engaging in dialogue, and responding what you say, but you ignore them: its always about your own points. In other words see WP:STICK. F = q(E+v×B) ⇄ ∑ici 11:09, 27 April 2012 (UTC)
- I plan to archive all up to section Mathematical representation, not much help reducing the byte count but better than nothing. F = q(E+v×B) ⇄ ∑ici 12:35, 27 April 2012 (UTC)
- F=q(E+v^B): I am discouraged to hear you think things have been repeatedly explained to me. Perhaps your observation is based upon what looks like a ping-pong match on this Talk page.
- Some of that was about including a brief quotation about the connection of wavelength to Fourier series,
- Fourier's theorem states that a function f(x) of spatial period λ, can be synthesized as a sum of harmonic functions whose wavelengths are integral submultiples of λ (i.e. λ, λ/2, λ/3, etc.)."[Schaum's outline]
- I plan to archive all up to section Mathematical representation, not much help reducing the byte count but better than nothing. F = q(E+v×B) ⇄ ∑ici 12:35, 27 April 2012 (UTC)
- an aside that was condemned by many, apparently on the grounds that if Fourier series did not illuminate wavelength, then wavelength as the fundamental parameter in Fourier series should not be mentioned. Such a stance is hardly WP policy, and on most pages an aside like this wouldn't raise an eyebrow.
- More recently, I raised a three-point critique of Dick's revision. Blackburne has contributed nothing beyond "I agree with Dick" and Dick has contributed little more than snide remarks.
- In any event, sourced and factual matters like the mathematical definition of wavelength f(x−vt+λ) = f(x−vt) are discarded as "trivial" and "mathematical pretense". Even a simple matter like substituting Fourier integral for Fourier series is refused based on a specious argument that an article on Wavelength (I'd say obviously a spatial concept) is really about analyzing time dependence of a dispersing waveform as a "temporally periodic wave".
- I'll leave this for a few weeks and try again later. Brews ohare (talk) 14:08, 27 April 2012 (UTC)
- The battles long ago with Dick Lyon over a reference to the envelope of a waveform (which it seems he continues on this page) led to my construction of Envelope (waves) which has attracted about 1100 hits so far. Perhaps an article on the analysis of propagating waveforms in a dispersive medium would be a helpful resource. Brews ohare (talk) 15:43, 27 April 2012 (UTC)
On "Such a stance is hardly WP policy" One core policy is consensus, and a formal mechanism for achieving it is a RfC. As that is over (you removed the tag) and the consensus was clearly against you your proposal failed. Stop now trying to re-argue the point as if you are right and every other editor is wrong. That's not how it works. You tried to convince other editors. You failed. Please don't try to start the same argument, now or in a few weeks. Drop it and move on.--JohnBlackburnewordsdeeds 16:26, 27 April 2012 (UTC)
- Blackburne: You mistake my purpose in my remarks. I wish to register dissatisfaction with the rejection of material that is sourced and relevant on the supposed basis that it is trivial and discursive, an imposition of subjective matters of taste that cannot be supported objectively. Dick Lyon refuses to accept even the most minor changes in this article based on his personal views, many mistaken or arbitrary matters of taste, and you facilitate his actions with your "me too" approach that contributes nothing substantive to the discussion. Brews ohare (talk) 16:46, 27 April 2012 (UTC)
- Let me close what I have to say: Brews I have nothing against you, and I'm not "taking sides" with Dicklyon and Blackburne. The earlier response was a perfectly neutral observation of the RfC (not a "ping pong match"). The RfC was what it was. Don't get me wrong - much appreciated you would like to improve the article but this isn't getting anywhere... F = q(E+v×B) ⇄ ∑ici 17:08, 27 April 2012 (UTC)
- F=q(E+v^B): I was not under the impression that you were siding with these two. The RfC figures large in Blackburne's mind because it involved some outside participants, making it look less like just a quarrel. However, most of the ping-pong activity on this page is just DickLyon and myself with Blackburne as the "me too" cheerleader.
- In his desire to appear as always right, DickLyon has introduced actual errors of fact into the article, and has reduced the article's value to WP by deleting useful and sourced material of general interest to readers, such as the mathematical definition of wavelength described by Dick as "trivial" and "mathematical pretense". These actions also appeared some time ago on the topic of "envelope", and led me to write Envelope (waves) just to avoid DickLyon, which may have turned out to be a plus for WP, but the circumstances leading to its birth are a testimony to his intransigence. Brews ohare (talk) 17:30, 27 April 2012 (UTC)
- Let me close what I have to say: Brews I have nothing against you, and I'm not "taking sides" with Dicklyon and Blackburne. The earlier response was a perfectly neutral observation of the RfC (not a "ping pong match"). The RfC was what it was. Don't get me wrong - much appreciated you would like to improve the article but this isn't getting anywhere... F = q(E+v×B) ⇄ ∑ici 17:08, 27 April 2012 (UTC)
This is an archive of past discussions. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page. |
Archive 1 | ← | Archive 3 | Archive 4 | Archive 5 |
- ^ Ken'ichi Okamoto (2001). Global environment remote sensing. IOS Press. p. 263. ISBN 9781586031015.
- ^ Roger Grimshaw (2007). "Solitary waves propagating over variable topography". In Anjan Kundu (ed.). Tsunami and Nonlinear Waves. Springer. pp. 52 ff. ISBN 364209032X.
- ^ Alexander McPherson (2009). "Waves and their properties". Introduction to Macromolecular Crystallography (2 ed.). Wiley. p. 77. ISBN 0470185902.