The following discussion is an archived debate of the proposed deletion of the article below. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.
It strains credulity that one would invent a new method for dealing with constrained motion in classical mechanics in 1992. Indeed, a peer-reviewed comment [1] (mentioned in the talk page) on the original paper says that they are just reinventing the wheel. It also does not seem notable. There are two references that are not by the authors themselves [2][3]; they are absolutely fawning, but I call into question their reliability. One is published in the European Journal of Physics, a borderline journal, and the other in the Chinese Journal of Mechanical Engineering. I'm not familiar with this latter journal, but it's not where I would expect a revolutionary new method in classical physics to be reviewed.
Withdrawn by nominator I would like to withdraw this nomination, as the source found by RainerBlome, together with the others already present, suffices to establish notability. It appears that I'm not allowed to do so, however, by WP:WDAFD, because other editors have supported the nomination. I would appreciate though if they would reconsider their !votes based on the new source "Flexible Multibody Dynamics". Tercer (talk) 10:26, 26 April 2021 (UTC)[reply]
Delete In the absence of better sources evaluating whether there is anything sufficiently novel here, we shouldn't have an article about it. It sure looks like the wheel has been reinvented; at the very least, we need better documentation to the effect that it hasn't. XOR'easter (talk) 16:21, 18 April 2021 (UTC)[reply]
Weak keep We have what looks to be three admissible references now: two papers (with overlapping author lists) and a section in a textbook. That might be enough to scrape by, though I'm not convinced that the available coverage gets us beyond the "section in an article on a broader topic" level. XOR'easter (talk) 16:57, 26 April 2021 (UTC)[reply]
I agree, Wikipedia would be better served having this material as a section in a larger article about constrained system. But (a) I searched for a bit, and couldn't find an article into which it would fit and (b) If the article doesn't violate Wikipedia policy and there's an editor interested in maintaining it, we should live and let live. Tercer (talk) 08:31, 27 April 2021 (UTC)[reply]
Comment the EJP is nowhere near borderline. That specific article may not rise to make it notable, but EJP isn't a journal for crazies. Headbomb {t · c · p · b}10:14, 19 April 2021 (UTC)[reply]
I have a vague recollection of seeing some pretty bad papers in EJP. But a vague recollection is not good enough, and I don't want to research further, so I'll defer to your judgment and strike out that comment. Note that it shares a couple of authors with the paper in the Chinese Journal of Mechanical Engineering. Tercer (talk) 11:07, 19 April 2021 (UTC)[reply]
There's bad papers in pretty much every journal. EJP isn't immune to that, but Nature also had bad papers in it. Anyway, the general point of this not being mainstream is still valid even if there's one paper claiming otherwise. Headbomb {t · c · p · b}20:27, 19 April 2021 (UTC)[reply]
I repeat point 8 from there: My understanding is that the defining novel element is the use of the Moore–Penrose pseudoinverse.
Regarding "reinventing the wheel": The letter said "a rearrangement of known facts". In my understanding, the method is not a "rearrangement", it is a combination of known methods, producing a new method with very different properties than the established methods. If the method is new indeed, I call this an advancement. It is not "reinventing the wheel", it is "inventing a new kind of wheel".
The method is notable in my view because it is a basic and broadly applicable technique in its field.
If you claim that the method is not novel, please provide a reference to where it was already established.
If a review is fawning, it may be because the review author is close to the inventor. Yet, it may also be because the reviewed item is actually better in some regards than what was available before. Has anyone actually read any of the reviews in their entirety and can actually judge their quality?
As an aside, note that Udwadia and Kalaba did not call the method after their names, others did that.
Yes, it surprises me a bit too, that something this apparently basic should have been devised only in the nineties. But keep in mind that the pseudoinverse is relatively new, and that application of the pseudoinverse has only become economical since the widespread availability of computers.
My guess is also that the established alternative methods of solving the equations of motion are firmly entrenched in the education pipeline. This has two effects: Little motivation to look for new methods (why invent a new one?) and making it difficult to establish new methods (why learn a new one?).
Also, for computation in general and specialized areas in particular, new mathematical methods are being invented all the time, and as far as I'm concerned this is a specialized area. In this regard, the method is in good company.
What matters are what the reliable sources are saying, you have given none. What we have here is support from a single research group, this is way too little, specially compared to the extraordinarity of the claim: extraordinary claims require extraordinary evidence.
Constrained motion in classical mechanics is taught at pretty much every university on the undergraduate level; heck I was taught that in my undergrad! If a great new method had been invented in 1992 shouldn't it be mentioned in a couple of textbooks by now? Or at least awakened more widespread interest? Or at least someone other than the authors would have bothered to write an Wikipedia article about it?
The Moore–Penrose pseudoinverse is nothing new: about a century old if you count by the original discovery, and about seven decades if you count by the rediscovery. It's also the obvious thing you would try if you wanted to generalise the inverse: keep the kernel as it is, and do the inverse in the range. Tercer (talk) 08:14, 22 April 2021 (UTC)[reply]
Would it improve Wikipedia if this article is deleted? I argue that deleting would harm WP.
Deletion is a last-resort measure, if the article is not salvageable.
The reliability of existing references has been explicitly questioned. That needs evidence, which I do not see. There are at least two non-U-K sources, "European Journal of Physics" and "Chinese Journal of Mechanical Engineering". The claim that EJP is "borderline" has been contested already.
The CJME journal is published by Springer. From my observation, in particular the Springer publications available to me, Springer publications are very reliable sources. So none of these sources has been shown to be unreliable.
If the article needs better references, let us add them.
Has anyone actually clicked on the links provided in the notice box "This article is being considered for deletion"?
Clicking on the JSTOR link yields nothing, because JSTOR is too dense to understand the em-dash in the link. Remove the em-dash and you get some results.
Clicking on Google scholar yields 583 results (I mean the number at the top of the result page). Those are all, I think, by *other* authors referring to U&K (since they themselves did not call the method after their names). By the way, searching for『Udwadia–Kalaba equation』would not yield better results, because citations use the words "equation", "approach", "control", "dynamic", "formulation", "method", "theory" exchangably.
So yes, within the scientific community the topic appears notable to me.
I just meant that as basic as the U-K method may seem, it could not have been discovered earlier than the pseudoinverse. But we are not talking about the pseudoinverse itself here, we are talking about a possibly novel use of it.
"We have carefully gone through the references cited in Bucy (1994) and nowhere have we found the close connection between generalized inverses and constrained motion as described in our paper."
> If a great new method had been invented in 1992 shouldn't it be mentioned in a couple of textbooks by now?
Yes, it should, but it does not have to be. Whether a method is "great" (your wording) depends on your requirements. I can imagine all kinds of reasons why this is not taught more widely.
"Can be found in a couple of textbooks" is is my view not a requirement for Wikipedia inclusion. Requiring this would instantly put a lot of valuable articles in jeopardy.
Note that the first reference in the article *is* a textbook, by U&K themselves (so it is not independent), "Analytical dynamics: a new approach", published by Cambridge University Press, which I consider reliable. According to Google Scholar, it has been cited 550 times. Not too shabby.
I know of no independent textbook.
It would not be the first time that people take their time adopting new ideas. Compare Grassmann algebra. Also see Clarke's three laws, especially "If an elderly but distinguished scientist says that something is possible, he is almost certainly right; but if he says that it is impossible, he is very probably wrong."
Delete - The subject is not sufficiently acknowledged in the literature. Not yet notable. Moreover, the claim that this is a new method needs strong evidence/documentation. --SimoneD89 (talk) 16:25, 24 April 2021 (UTC)[reply]
Regarding notability, see above.
Regarding novelty: Novelty can't be proven, only disproven. So no, it is the other way around: Claims of plagiarism and non-novelty need strong evidence. As far as I can see, there has only been a single attempt to disprove the novelty, the letter by Bucy. U&K have responded to this, see above. --RainerBlome (talk) 02:55, 25 April 2021 (UTC)[reply]
I confirm my vote because of the notability guidelines. I don't see enough secondary/tertiary, independent, and reliable (i.e. published in reasonable journals + have an acceptable number of citations) sources. If it is a recognized method, it should be simple to find sources or chapters in classical mechanics textbooks. It is not up to Wikipedia editors to decide whether the method is new or not. The argument was that since we are talking about a new method of an old and very established theory, one expects to see more evidence (i.e. reliable sources) that shows the acknowledgment in the literature. --SimoneD89 (talk) 06:06, 25 April 2021 (UTC)[reply]
I fully agree that it is not up to Wikipedia editors to decide whether the method is new or not. Please note that the deletion proposal brought this up, and to me it sounded as if you supported pursuing this. I'm glad if we can drop this aspect.
First, please note that "A New Perspective on Constrained Motion" has been cited over 400 times. Yes, that's a primary source, but I want you to be aware of this. I think this is an indication of notability, if not of reliability.
Second, the U&K textbook itself is a secondary source. It has been cited 550 times. Yes, that's not an independent source, but I want you to be aware that the number of citations does lend credibility to the work and some notability to the article topic discussed here, because the article topic is the main point of the work.
Third, above I have already linked to the Google Scholar page that lists over 500 citations of "Udwadia-Kalaba" by other authors. About one hundred of those sources have been cited a dozen times or more. Several have been cited more than a hundred times. This is strong indication that we are not in fringe land here.
To me, this is a secondary source. It cites the U&K textbook three times, and cites three of U&K's articles. According to Google scholar, this secondary source has been cited 256 times. Reliable enough?
Yes, finally, that's what I've been asking for, a reliable source! Let's look at it, shall we? The author is a professor of mechanical engineering [4], it's not a coauthor of Udwadia and Kalaba, and the book has been published by Oxford University Press, so it does count as independent and reliable. The citation number you gave is not for the book, it's for a review; the book itself has 21 citations [5], but that doesn't matter, we don't use citation count to establish reliability. But the important thing is, what does it say about the Udwadia-Kalaba equation? Well, I got the book from the Russians, and the answer is, nothing. I doesn't even cite the paper where Udwadia-Kalaba introduced their method. The mentions in the text are all to the textbook, there are three: For further details, see books on computational/multibody dynamics; for example, Nikravesh (1988), Udwadia and Kalaba (1996).., also for Gauss' principle, in particular, see, for example, Udwadia and Kalaba (1996). and See, for example, Girtler (1928), Lilov and Lorer (1982), Lilov (1984), Vujanovic and Jones (1989, chap 7; this also contains a ‘‘complementary’’ formulation of Gauss' principle where the accelerations are kept fixed and the impressed forces are varied), and Udwadia and Kalaba (1996).
This is pretty damming for the notability of the Udwadia-Kalaba equation. Even a textbook that cites the authors completely ignores it. Tercer (talk) 19:07, 25 April 2021 (UTC)[reply]
My impression here is that "only the strictest standards are good enough". I disagree, see below. Whoever did it at the time, it was damaging to put those promotional passages in the article, but that issue has been at least mostly solved. I'd say the article content is neutral enough. Do you agree?
This experiment gives an impression of the level of page views for random articles, and by extension of the level of notability that these articles have. Yes, I think that we *can* (but don't have to) honor page views as an indication of notability. We do this here for *our* readers, not for the authors and readers of secondary sources. If an article is often viewed, it means that readers *want* a good article at that address. If we delete such an article, we are sending them away and telling them "What you are looking for was not good enough to let you read it" (or improve it).
The article deletion process says to also use common sense. Common sense says: Wikipedia is good the way it is. Yes, not every part of it is good, and yes, every part could be better, but it is already good as whole. In particular, Wikipedia does *not* primarily suffer from too many articles. It does suffer from bad articles, but this is not one of them.
As far as I can see, the point of the U&K textbook is to thoroughly expound the method. Do you agree with this? If an author cites a textbook, they *note* the textbook. They note that the book says something and they implicitly say that it's worth noting this. So yes, they found the textbook to be notable. They don't have to reiterate the content.
There may also be economy at work here: The primary sources are freely available, and there seems to be a dedicated textbook available. Both of these shrink the market for other secondary sources, which may explain why we have not found better others so far.
Section 11.2.8 compares the method to other ordinary differential equation (ODE) methods covered in earlier sections.
Quoting the comparison section: "Udwadia's and Kalaba's formulation presents a number of advantages over the other formulations. The Moore-Penrose generalized inverse ... always exists whereas the other formulations require a full rank constraint matrix. ... U–K ... is capable of dealing with ... rank deficient constraint matrix ... such as those involving redundant constraints."
This is analysis and evaluation of the primary sources, making it a secondary source.
I have not found an easy way to check if Bauchau is independent of U&K, but my impression is that he is. Any help appreciated.
Yep, this is a good source. It is a textbook, published by a reputable company, the guy is a professor of engineering [6], and I checked his list of publications, he is not a coauthor of Udwadia and Kalaba. This establishes that it is reliable and independent. Now, the crucial thing, it actually talks about the subject, unlike the other source you gave. It reviews the Udwadia-Kalaba formulation, compares it to another method, and says it's good stuff. There you have it, notability is established, I'll withdraw the nomination for deletion. The article should be renamed, though, as the subject is not an equation, and it is not referred to as such, but a method, or formulation. I propose "Udwadia-Kalaba formulation", following this source. Tercer (talk) 10:18, 26 April 2021 (UTC)[reply]
Thank you for checking the author. Yes, the word "equation" is less than ideal here, which is why I have systematically used other words. I agree that『Udwadia–Kalaba formulation』would be better. "Formalism" might be an alternative, but "formulation" seems to be the preferred word in this area. "Method" would be slightly more general, but I would avoid being less specific than we can. --RainerBlome (talk) 12:37, 26 April 2021 (UTC)[reply]
Wikipedia:Wikipedia is not for things made up one day cautions people against writing Wikipedia articles about their own inventions. Even if someone close to U&K edited here, the article is not "written by them". The article is now 12 years old and there were about 30 authors. The "is not" page lists many reasons. Many of those do not apply here, because there have been scores of people using and citing U&K, which is secondary coverage. Since you cited the rule page, which reasons do you mean specifically? --RainerBlome (talk) 21:30, 25 April 2021 (UTC)[reply]
The above discussion is preserved as an archive of the debate. Please do not modify it. Subsequent comments should be made on the appropriate discussion page (such as the article's talk page or in a deletion review). No further edits should be made to this page.
