Talk:Udwadia–Kalaba formulation

Overblown significance

This page feels like advertisement. Almost all the references are written by Udwadia. The "Background" section looks especially like advertisement. The equation is not nearly as significant as it claims to be. pony in a strange land (talk) 19:29, 24 March 2018 (UTC)

pony in a strange land has removed the "Background" section (and rightly so).  Done --RainerBlome (talk) 00:20, 26 April 2021 (UTC)

I somewhat agree, specifically with use of the opening phrase 'In Theoretical Physics', ... contrasted to its closing 'Mechanical Systems'. The authors of the [U&K] formulation were working explicitly in dynamics of constrained mechanical systems, not in theoretical physics. My first encounter of the work was one under only Udwadia's authorship (Firdaus Udwadia. Fundamental Principles of Lagrangian Dynamics: Mechanical Systems with Nonideal,Holonomic, and Nonholonomic Constraints. Journal of Mathematical Analysis and Applications,Elsevier, 2000, 251 (1), pp.341 - 355. �10.1006/jmaa.2000.7050�. �hal-01395965] as posted at HAL archives-ouvertes.fr), and the first part of this WP page reads almost verbatim with it up to the point where constraints are described (holonomic,scleronomic,...). Udwadia is also with an Aerospace unit at the university and his other citations coauthor with many in the sensor applications field.

His co-author, the late Kalaba is/was more grounded in Dynamics, and his Response to Bucy (1996 also @JSTOR), dismisses Bucy's quibbles convincingly in my opinion. But the [U&K] construction is not Theoretical Physics per se, but more Dynamics under Contrived equality Constraints; which are useful in their own right without unduly associating to Theoretical Physics.

I personally found it useful in better understanding hand-held device accelerometer readings, constrained by hand-holding as it were. Then the fundamental acceleration gets partitioned by frame of reference, into the Inertial and the Rotational ( comprising what I loosely term the Euler Torques, The Coriolis Slippages and the Huygens Centrifugal force tendencies), well documented somewhere here in Wikipedia), and then [U&K] add to that an elegant Constraint Forces factor/frame of reference into it.

One can also track the evolution of this [U&K] formulation from their early Generalised Inverse form to the recent Metric which would in some way be addressing Bucy's Existence quibbles. My personal quibble if at all that, would have been the requirement that a constraint be explicitly time dependent. One empirically finds constraints topologically, in some force field, absent time as paramater (t); but their approach changes attack to the problem completely by trivialising the finding of a solution space but agonizing over how and when it can ever exist. (In fact the Elsevier, 2000, 251 (1), pp.341 - 355 article repeatedly places the onus of properly specifying ... force of constraint that is prescribed by the mechanician and not by physicists)--Mkhomo (talk) 13:54, 24 October 2020 (UTC)

I have replaced "In theoretical physics" with "In classical mechanics". --RainerBlome (talk) 00:20, 26 April 2021 (UTC)

Tagging for lacking NPOV

I agree that this page is written from a fan's perspective. For example, there are statements such as, "The fundamental equation is the simplest and most comprehensive equation so far discovered." Claims such as "simplest" or "most comprehensive" are not well-defined, and do not belong in an encyclopedic article. This is only one example, and the article has many more such instances of such issues. I am marking this page as violating Wikipedia's Neutral Point of View policy. - V madhu (talk) 16:43, 26 September 2018 (UTC)

The subjective phrases have all been removed or rephrased. Still most of the references are by Udwadia, but this in itself does not make the article NPOV. I suggest to remove the NPOV tag. --RainerBlome (talk) 14:51, 13 December 2019 (UTC)

I have removed the POV tag.  Done --RainerBlome (talk) 00:20, 26 April 2021 (UTC)

Non-notable and suggest deletion

This article is lacking in notability. In a peer-reviewed and published letter to the editor, Bucy presents the many ways in which the original article by Udwadia and Kalaba is flawed. To quote from the letter,

In a recent paper in this journal, Udwadia & Kalaba (1992) rediscovered the Gauss principle (Pars 1965), and combined it with the Moore-Penrose pseudo-inverse (Penrose 1955), to write the equations of motion for a constrained mechanics problem. These results are well described in the classical literature, the Gauss principle of least constraint (Hofrath & Gauss 1829), the Gibbs-Appell equations (Appell 1925), and are available in the survey of Arnold (1980), or in classical language in Pars (1965). Consequently the paper of Udwadia & Kalaba contains nothing new or important, and is simply a rearrangement of known facts.

Given the lack of notability of these equations and the apparent attempts at advertisement, I suggest moving this page towards PROD. However, a second expert in mechanics (I am one) should examine before moving to deletion.

-V madhu (talk) 12:14, 2 December 2019 (UTC)

TL;DR: I oppose.

In the "Expert needed|Physics" tag, you gave as a reason "The accuracy and utility of this topic is shown to be limited". Can you be more specific?

1. What has been shown to be inaccurate? Can we fix any inaccuracies in the article?

2. In what way is the utility of the method limited? Would it help if we described any limitations in the article?

3. Thank you for bringing that letter to our attention. Sadly, the letter is behind a paywall. The quote you gave is too general to make an informed judgment, for my taste. Does Bucy get more specific?

4. You state that "Bucy presents many ways in which the article is flawed", yet name only one, that the article presents nothing new. What other criticisms does Bucy have?

5. As such, that letter is "just a letter to the editor", so it stands "article against letter", both peer-reviewed. Did the journal editors reply in any way?

6. Just for fun: Maybe Bucy was grumpy because he thought the U&K method so obvious that he thought "Damn, why didn't I think of this? Can't let them get away with this." :-)

7. How did you find that letter, by the way? I had also tried to judge the credibility of the article, and do not remember finding the letter.

8. The use of the pseudoinverse is a defining part of U&K. The pseudoinverse has been "widely known" since only about 1950, so I fail to see how earlier publications are relevant to Bucy's argument. Do earlier publications, for example the one Bucy mentioned from 1925, mention anything akin to the pseudoinverse?

9. The contribution that I see in Udwadia's and Kalaba's work ("U&K") is show-and-tell, at the very least. They say "look, this is how it's done cleverly, today". However, even though I have had some training in the method of Lagrange, and think that I understand the principles that U&K use, at the moment I do not understand both methods well enough to be able to soundly judge their merits. I keenly remember that Lagrange is difficult to apply in practice. U&K claim that their method is much simpler to work with, and that it works well enough for practical use, provided numerical horsepower is available. If what they say is true, I think that the method is very valuable. (The widespread availability of computers may have played a role in the U&K method being fully developed only in the 1990s.)

Their method is the only one that I know of that does the exact job that they claim it does (note that I do not claim that it *is* the only one). As far as I understand, they did not claim to have invented Gauss's principle. They just said, "Hey, look, now that we can easily compute pseudoinverses, we can use that to solve the equations numerically and be done with it." Bucy may call this "a rearrangement of known facts", but if the specific "rearrangement" is unpublished so far, I do see value in publishing it, and I do call it new. Whether it is important is another matter, and depends on the comparison to other methods.

--RainerBlome (talk) 15:05, 13 December 2019 (UTC)

OK, agreed that deletion may be too severe a response. Instead, a section needs to be added on the controversy related to its notability. They responded to Bucy, but only to one of his criticism: https://doi.org/10.1098/rspa.1996.0053.

--V madhu (talk) 16:29, 9 January 2020 (UTC)

Kudos to your search-fu, thanks for digging up that response. You write "They responded to Bucy, but only to one of his criticism[s]". Sounds like you did not read their entire response, the page that you linked to shows just an excerpt. The full response is behind a button "View PDF". In it, U&K reply, numbered, to five of Bucy's points. (Because I do not have access to Bucy's letter, I do not know how many points he made.) In particular, they responded to the issues raised in the quote you gave.

To paraphrase, you responded, but only to one of my questions, question 5. :-) For ease of reference, I have now numbered them all. 9. is not a question, just my opinion. Responding to two of my own questions:

6.: Searching for "Bucy" shows that he lectured at USC when he wrote his letter, where U&K worked when they published their article and were still working when they published their reply to Bucy. Isn't that interesting? Maybe my joke is not as far-fetched as I thought it to be.

8.: U&K state in their reply to Bucy that they checked the literature, and claim that they did not find literature that uses the Moore–Penrose inverse to solve for equations of motion using generalized coordinates.

--RainerBlome (talk) 00:56, 19 January 2020 (UTC)

Coming back to this after a long time. I think my issue is not in the nitty-gritty but that the U&K's contribution is yet another paper and yet another intepretation of classical mechanics. A new formulation such as this would have been a big deal 150 years ago. But in light of major developments, like Noether's theorem, it lacks novelty beyond being a specific technical paper that does not merit a Wikipedia page. On the basis of correctness alone, one could argue that most published papers should have a Wikipedia page. That would be fine if it met the criteria for Wikipedia:Journal to wiki publication, namely it was a review paper of note. The U&K formulation does not come close to meeting that bar, because it is simply yet another technical paper that provides a new interpretation of conservation laws under constraints. But just because it is a peer-reviewed paper, it does not make a Wiki article.

--Madhu (talk) 02:56, 11 February 2022 (UTC)

Formally: There has been a formal deletion discussion, and the decision was to keep the article, because notability was established. You will be able to find links to the discussion in the page history.

Informally: "U&K" isn't "just one paper among many", it's a method. The Wikipedia article isn't about a paper, it's about the method. The method's name varies. The authors themselves did not call it "U&K", that name was given to it by others referencing media about the method, for example the first paper. Last time I looked, Google Scholar counted 896 references mentioning "Udwadia–Kalaba". The paper you talk of is not the only source describing the method. There are textbooks by other authors that mention the method and at least one independent textbook fully describes and evaluates the method. And there is a whole textbook about the method by U&K themselves. What made you think that "it's just one paper"? U&K's claim is that applying the Moore–Penrose inverse to solve the problem is new. Quoting the response of U&K to the criticism by Bucy:

"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 you know better, please show us a publication that does this that predates U&K. --RainerBlome (talk) 01:08, 12 February 2022 (UTC)

References