User talk:Jmath666

From Wikipedia, the free encyclopedia

[edit] Welcome

Welcome!

Hello, Jmath666, and welcome to Wikipedia! Thank you for your contributions. I hope you like the place and decide to stay. Here are some pages that you might find helpful:

I hope you enjoy editing here and being a Wikipedian! Please sign your name on talk pages using four tildes (~~~~); this will automatically produce your name and the date. If you need help, check out Wikipedia:Questions, ask me on my talk page, or place {{helpme}} on your talk page and ask your question there. Again, welcome! Oleg Alexandrov (talk) 06:24, 17 February 2007 (UTC)

[edit] Regarding edits made to Domain decomposition method

Thank you for contributing to Wikipedia, Jmath666! However, your edit here was reverted by an automated bot that attempts to remove spam from Wikipedia. If you were trying to insert a good link, please accept my creator's apologies, but note that the link you added, matching rule \bexample\.com, is on my list of links to remove and probably shouldn't be included in Wikipedia. Please read Wikipedia's external links policy for more information. If the link was to an image, please read Wikipedia's image tutorial on how to use a more appropriate method to insert the image into an article. If your link was intended to promote a site you own, are affiliated with, or will make money from inclusion in Wikipedia, please note that inserting spam into Wikipedia is against policy. For more information about me, see my FAQ page. Thanks! Shadowbot 06:02, 17 February 2007 (UTC)

This was the link to the standard community webpage and a basic reference book on the subject. Reverted. Reply posted here for reference only. Jmath666 17:22, 17 February 2007 (UTC)

[edit] On revising the "Domain decomposition method" article

I replied on Talk:Domain decomposition method. Oleg Alexandrov (talk) 06:24, 17 February 2007 (UTC)

I replied on Talk:Domain decomposition method. Jmath666 17:22, 17 February 2007 (UTC)

I replied on my talk on LaTeX to Wiki translator. Oleg Alexandrov (talk) 19:12, 17 February 2007 (UTC)

[edit] Math not rendering correctly

Article with math is missing many formulas when rendered by browser. Jmath666 22:23, 21 February 2007 (UTC)

Try purging or bypassing your cache. John Reaves (talk) 23:58, 21 February 2007 (UTC)

Possibly timing issue, long wait for PNG formulas to appear. Upgrade to Firefox 2 and patience resolved this. Maybe Wikipedia servers overloaded. Jmath666 04:20, 23 February 2007 (UTC)

[edit] Articles you might like to edit, from SuggestBot

SuggestBot predicts that you will enjoy editing some of these articles. Have fun!

Stubs: Skew-T log-P diagram; Common Data Format; ANCOVA; Tephigram; Equivalent potential temperature; Minimum mean-square error; Conjugacy class sum; Convective inhibition; Contraction mapping; Technical rescue; General covariance; Mathematical statistics; Mathematical Reviews; Humidex; Random matrix; Bibliography; Convective available potential energy; Great dodecahedron; Multivariate random variable

Cleanup: Fast Kalman filter; Helmert-Wolf blocking; Country Fire Authority
Merge: Point estimation; Eigenvalue algorithm; Bambi bucket
Add Sources: Monotonic function; Random matrix theory; Continuous probability distribution
Wikify: Kálmán Tisza; Cedar Fire; Doctor of Pharmacy
Expand: Baker-Campbell-Hausdorff formula; Bandwidth; Memory hierarchy

SuggestBot picks articles in a number of ways based on other articles you've edited, including straight text similarity, following wikilinks, and matching your editing patterns against those of other Wikipedians. It tries to recommend only articles that other Wikipedians have marked as needing work. Your contributions make Wikipedia better -- thanks for helping.

If you have feedback on how to make SuggestBot better, please tell me on SuggestBot's talk page. Thanks from ForteTuba, SuggestBot's caretaker.

P.S. You received these suggestions because your name was listed on the SuggestBot request page. If this was in error, sorry about the confusion. -- SuggestBot 15:26, 15 March 2007 (UTC)

[edit] Fachinformationszentrum Karlsruhe

A tag has been placed on Fachinformationszentrum Karlsruhe, requesting that it be speedily deleted from Wikipedia. This has been done under the criteria for speedy deletion, because it is a very short article providing little or no context to the reader. Please see Wikipedia:Stub for our minimum information standards for short articles. Also please note that articles must be on notable subjects and should provide references to reliable sources that verify their content.

Please do not remove the speedy deletion tag yourself. If you plan to expand the article, you can request that administrators wait a while for you to add contextual material. To do this, affix the template {{hangon}} to the page and state your intention on the article's talk page. Feel free to leave a note on my talk page if you have any questions about this. Realkyhick 05:11, 19 March 2007 (UTC)

That's good enough for me. I must say that it might be a bit too technical for a lot of readers, so if you can include something that gives a little more context and explanation, that would be good and would put this on the way to becoming a featured article. In the meantime, I've removed the speedy-delete tag. Bear in mind that another editor may put up a {{prod}} tag, but there's enough citations and the like now to make that unlikely. Realkyhick 06:15, 19 March 2007 (UTC)

[edit] (Ir)relevance of elegance

Mathematical elegance links directly to mathematical beauty.

That's exactly what I was affraid of.

Now, whatever that may be, "mathematical beauty" is, if not completely subjective, a niche concept, belonging to the academical culture of mathematicians, falling outside the actual scientific inquiry (considering mathematics as a formal science) and thus without relevance in an objetive exposition pertaining mathematics itself (and not the academical culture of mathematicians).

Such a thing belongs in an article about the culture, values and tradition of mathematicians, not mathematics. Perceived beauty is not requisite for mathematics, that's a personal/cultural assessment, not the least relevant for the thing itself (bear in mind I'm not talking about mathematicians motivations, inspirations and whatnot, but only mathematics in itself).

That would be like assessing how beautiful, awesome, cool, etc., this or that scientific theory is or is not.

Fred Hoyle thought the Big Bang was a ghastly theory. So what? Should we put in the article about the Big Bang a remark saying that it is the best theory EVEN THOUGH it's not that "elegant"?

How about an universe without aether? "How 'unelegant' that is!" That's certainly what many proiminent 19th century scientists would say.

What you tell me about Hardy, who thought that only useless mathematics were beautiful? Should we label every applied math article with "but this is not as 'elegant' as pure math"?

Really, this directly assessment of "elegance" sounds POVy to me. Wikipedia has no place in assessing the "elegance" of this or that. This word is too heavily loaded with personal/cultural aesthetical conotations, as the redirect of mathematical elegance to mathematical beauty itself proves.

I mean, at least write "mathematicians usually consider (whatever) to be elegant/have elegant properties", for crying out loud. Or, if you only mean "simplicity", write "simplicity" already, since that's the proper word for that, otherwise "elegance" is just slang for "simplicity".

But I really think they should be removed altogether.

Remember you're not writing these articles for the mathematics academical niche, but for the whole world to read. Loading it with cultural niche value assessment certainly will not help understand an already complicated subject as mathematics. —The preceding unsigned comment was added by 200.164.220.194 (talk) 20:08, 19 March 2007 (UTC).

I agree with you when you say that Wikipedia has no place in assessing the "elegance" of this or that and that the use of this word is being overused by some editors and that it should be be used on Wikipedia much less if at all. In addition, I also think that the word "elegant" might smell a bit of original research if the math in Wikipedia is something else than what can be found in cited sources. Many of us have favorite elegant proofs of this and that that we use in teaching and that cannot be find anywhere. I find the overuse of the word elegant on Wikipedia annoying but it does not bother me enough to go into a crusade about it. If you want to do anything about it the way to go might be to build a consensus at Wikipedia:WikiProject Mathematics. Then the civil minded editors might comply. For this it would be useful to make an account and get ready for protracted and very diplomatic discussion. I would suport you to the extent it corresponds with my own views (see below) but I do not really care enough to take lead in this.

My personal view: I do believe in clear effective communication. I do not stand in awe in front of self-serving mathematical beauty. I think that all good academicians should practice clear communication and distinguishing what is important and what not and that mathematical elegance is a very important part of clear communication. Maybe your experience is with bad academicians. When writing for practitioners the clarity of expression is even more important. An article competes for attention with many other articles and messy stuff gets ignored.

Thank you for your thoughts and good luck oif you choose to build that consensus. Jmath666 20:36, 19 March 2007 (UTC)

Regarding the simple=elegant, I wrote elsewhere that simple is not all. From Mathematical elegance the following attributes are important for good communication, in writing a Wikipedia article or anything else. My comments in boldface

A proof that uses a minimum of additional assumptions or previous results. So that it is more applicable...

A proof that is unusually short. Torturing the reader with pages of proofs is BAD...

A proof that derives a result in a surprising way (e.g. from an apparently unrelated theorem or collection of theorems.) More powerful tools....

A proof that is based on new and original insights. Hopefully clearer, but for effective communication not important...

A method of proof that can be easily generalised to solve a family of similar problems. These tend to be clearer, too, not burdened by specific details...

Regarding the fact that mathematical elegance is redirected to mathematical beauty, I agree also. I think these are different concepts. There should be a separate mathematical elegance article. It might be tough to find citations, remembering the no original research rule - a personal view is not good enough, Wikipedia is supposed to describe what exists not what we think should be, and with subjective concepts, one should simply discuss the different views while following a neutral point of view. It might be a good idea to outline the plan at the talk page of mathematical beauty and let it sit for a while first soliciting feedback. I think an article with proper attribution to sources would not be contested. But this all takes work... Jmath666 23:52, 19 March 2007 (UTC)

Hmm... I agree with you. Maybe I am overreacting a bit too. I mean, it's only a word afterall. I think an article about mathematical elegance (that does not link to mathematical beauty) would be a good middleground, if one can properly define elegance and make this definition available in Wikipedia itself (most of my opposition comes from that being so poorly-defined and subjective... once this is solved, I wouldn't bother so much). —The preceding unsigned comment was added by 200.164.220.194 (talk) 12:17, 20 March 2007 (UTC).

I have removed the statement re elegance at Estimation of covariance matrices after proposing this on its talk page and waiting for 5 days. Let's see what happens. Jmath666 23:07, 31 March 2007 (UTC)

[edit] Sobolev space

I would like to attempt to rewrite Sobolev space. This article, which is quite important, is written in a messy manner (in my opinion). Some points which I would like to stress are described in User:Igny/Sobolev space (they are somewhat mentioned in the article, but like I said it is a mess). In particular I would like to stress the connection to the Fourier transform of distributions, which, by the way, deserves a separate article in my opinion. I will appreciate any input from other editors, in particular a blessing to proceed. (Igny 19:17, 21 March 2007 (UTC))

Yes, the article could be better. If you undertake such project, could you please allow for multiple definitions of Sobolev spaces? Perhaps you could structure it as section "Definition of Sobolev spaces", with subsection(s) for definitions, so that more definitions can be added in future. Because:

different definitions do not always give equivalent spaces
simple definitions though maybe not as satisfactory have an important place in teaching and are very suitable for encyclopedic purposes. In order of accessibility:
definition by completion of a space of smooth function (requires only the concept of completion of metric space)
definition by $L p$ weak derivative (requires Lebesgue integral but neither Fourier transform nor distributions)
the distributions/Fourier transform way goes the whole mile but is the least accessible
the definition by Fourier series on an interval is a good example for teaching and sometimes a nice trick to know

And yes, distributions should have their own article. So should interpolation of spaces.

Also, it would be good to have at the top of the article something simple yet specific even if maybe not 100% accurate so that people without much background get the correct idea what the topic is (i.e. without knowing what

L p

and multiindex are and so on). Many math article are done this way. Maybe something like this: Sobolev space is a normed space of functions. The norm on Sobolev space of order n involves the value of the function as well as its derivatives of order up to n. The Lebesque spaces ... are a special case of Sobolev spaces of order zero. Negative order Sobolev spaces are defined as dual spaces to spaces of positive order, and Sobolev spaces of non-integer order are defined by interpolation of normed spaces (which is not the same as interpolation of function). The importance of Sobolev spaces lies in the fact that the smoothness of a function is measured by in which Sobolev space it is, and solutions of PDEs fall naturally in Sobolev spaces." Then the example of the most common space,

H 1 = W 2,2

, in 2D, with all partials written out, and saying that the derivatives are suitably generalized for this whole thing to work, then the TOC and then the messy technical stuff. Thanks for taking this up! Jmath666 01:12, 22 March 2007 (UTC)

I have improved the content and outline of User:Igny/Sobolev space. Please take a look and comment. I have been busy with other things, and I plan to add embedding, trace, extension, and references which I used sometime later. Probably, we should also add history. Unfortunately, I am not familiar with the space interpolation to contribute much there. My version of the article reflects my view on the subject, it is in no way a replacement of the current article, but I hope to start (cautiously) merging the content at some moment. (Igny 20:46, 31 March 2007 (UTC))

Thank you. I have copied our discussion from Wikipedia talk:WikiProject Mathematics above for reference and to avoid repeating myself.

The essential thing about Sobolev space is use of the weaker

L p

norms instead of the stronger max norm; the rest, weak derivatives and all, follows logically.

Just few of the things I noticed: There are many norms here so one should always say in which norm unless the standard norm for the space is meant, and sometimes even then.

C k (Ω)

is complete, under its standard norm, which is implied by the notation. I still remember multiindices were a hurdle for me and I would never use them in introduction. I'd spell out some. Some of the theorems hold only for domains with certain boundary properties (smooth, Lipschitz,...) or bounded domains. Some things do not hold for $p=\infty$ .

W - k, p (Ω)

are not defined as dual spaces of $\left(W^{k,q}(\Omega)\right)^*$ but $\left(W_0^{k,q}(\Omega)\right)^*$ which you must define first. Bra-ket is not a standard notation here. Normally we use $\langle u, v\rangle$ or simply

(u, v)

for duality pairing and/or inner product (usage differs, but there is never the "|").

It is possible to define Sobolev spaces without interpolation, but it is clumsy. I might write something about interpolation later but I am not expert in that, just a user of it, and I am busy now.

What is your source? Some books have errors or the notation may mean something else than what meets the eye. It is a difficult dark art to be lucid, compact, intuitive, and not to say anything incorrect at the same time. Jmath666 22:54, 31 March 2007 (UTC)

One of the sources is [1] (you can also look at other chapters which are about applications to the finite elements method). Another source is Lecture notes on distributions and Sobolev spaces (in Russian) [2], they have Gelfand, Sobolev, Schwartz, Shilov listed in References. Yes I know that some theorems require additional conditions and I was sloppy at times, I will fix this (eventually).

My motivation was that I thought that Sobolev space is a math article which deserved at least a GA status. I am also working on this article in my spare time, so it may take a while. (Igny 00:50, 1 April 2007 (UTC))

Yes, I see that Ron has made a great job on the summary but there is little context. And the matter is not so simple and he is already so terse that omitting anything will give incorrect statements. Looks like some careful wording or ugly "suitable assumptions" will be in order. Now just to know what can be be swept under the carpet and what cannot be; it will require a lot of care to distinguish what is important and what is just hairsplitting... Jmath666 01:29, 1 April 2007 (UTC)

[edit] Re:Jmath666/Computational mathematics

Hi, thanks very much for watching out for vandalism. That was quick and it is very nice to see someone is doing that. In this case I put the {{delete}} myself but the login expired so it got logged under the IP. Jmath666 16:08, 7 April 2007 (UTC)

Oh, right. Sorry about that- I didn't want your page to be deleted without you knowing! Anyway, cheers- CattleGirl ^{talk | sign!} 02:27, 8 April 2007 (UTC)

[edit] Requestion

Hi. I've been on holiday so have not posted for a while on the Hilbert-Hermitian wavelet discussion page. I've just read the latest entires and wanted to write to you reagrding Requestion. I don't agree that it isn't a personal attack on Jon Harrop. I think it is very personal based on Requestion's past comments on the HH wavelet page and other pages.

This is my married name which I rarely use. I normally contribute using another name. I set up a new account under this name in order to avoid harassment from Requestion. I have met Jon Harrop persoanlly on more than one occasion. I can say that he is genuine guy with good intentions at heart. He has been pulled up before for 'spamming'. However, after the issue was discussed and resolved he publicly apologised to pjacobi for spamming and promised not to do it again (since then he has not posted any links, which could be regarded as spam, on the pages in question).

Jon Harrop has taken the time to make useful contributions to various wikipedia articles. However, Requestion's attitude is not acceptable. It appears to me that he wants to remove anything that Jon Harrop has written just because he believes that Jon may spam again. I get the feeling that Requestion wants to chase Jon Harrop off the wiki completely. I am not surprised that Jon Harrop is becoming a little defensive and thinks the comments from Requesion are personal. Requestion has been very inflammatory and I fond that he is rarely constructive. Looking into other situations involving Requestion and various other users Jon Harrop is not the only person that finds Requestion to be highly irritating and vindictive.

I have also read Jon Harrop's thesis (all of it) and I have read the papers written by groups using the wavelet. I do not agree with Requestion's conclusions.

I have said what I think about the HH wavelet article on the discussion page. I do not think it should be removed completely but I do not intend to waste time arguing with Requestion about it.

I hope the discussion reaches some sort of conclusion soon.Marie Mason 13:03, 30 April 2007 (UTC)

The sad person that is Requestion has taken it upon himself to delete the Hilbert-Hermitina wavelet article anyway. Even though the article has been changed and discussed. He is God apparently and if he wants it to be deleted then it shall be deleted. If you ask me Requestion has small man syndrome. I'm not going to bother with this wiki anymore (I've found more suitable biological wikis). This one is now full of weird, egotistical 'spam fighters'. If only they would divert some of their energies to helping good causes (such as child mortality) the world might be a better place. Instead they spend all day every day on this wiki - 'patrolling' it. It's very sad indeed. I don't want to worry about having my stuff deleted all the time. I (like you) have even felt the need to change my account to avoid persecution from Requestion et al. I don't want my previous contributions to be deleted.Petdoc 18:01, 12 May 2007 (UTC)

I actually agree with Requestion. I tried to find a middle ground at one point, but now I do not think it is possible. Sorry. This is a blatant case of self-promotion, which is not what Wikipedia is for. Otherwise it would be full of everyone's spam. There is a well recognized vehicle for information about own work: scientific publishing. Which Wikipedia is not. Jmath666 22:43, 12 May 2007 (UTC)

Okay - I'm trying to get a better understanding here . The work has been published in scientific papers but that's not adequate. Is that correct? Therefore, am I right in thinking that only work published in a hard back book is accepted as established knowledge. Therefore, once the work is published in a book with an ISBN it will be fine (as long as the wiki article is written by someone else). It isn't enough that it is already used commercially and by academics? I guess I don't get as wound up by spam as some. And I admit I don't see a lot of contributions on the wiki as spam when others feel strongly that they are. I think some of the OCaml stuff was self promotion and that should have been removed and was (a long with a lot of other peoples' self promotion). However, I really have to disagree that the HH wavelet article is blatant self promotion. If all links to Harrop's external websites are removed then the article contains useful information. However, if you and Requestion feel so strongly about it then I guess it will be deleted. I think it would have been easy to find a middle ground had Requestion (and Harrop) not been so active in the discussion. Although Harrop was usually rebutting a comment made by Requestion. They obviously upset each other and I can see that it must look very personal to Harrop. Harrop makes a lot of contributions throughout the web that aim to share knowledge and help others (contributions that have no self gain at all). I think that should be noted. Take care. Petdoc 01:08, 13 May 2007 (UTC)

Writing about one's own work is a treacherous ground as I tried to point out to jhd30. There is an avalanche of papers, books, and all, little of that is notable, and it takes more to be notable as Requestion pointed out referring to WP:SCIENCE. We all have stuff that is used. So what? You can tell the notable stuff by citations. There is an established vehicle for that also, Web of Science. To some extent also Google Scholar. No, I do not feel that strongly about it; I have responded to your entry on my talk page here but I did not reenter the discussion on the article talk page and I do not intend to do that. Jmath666 06:46, 13 May 2007 (UTC)

[edit] Mathematics CotW

Hey Jmath, I am writing you to let you know that the Mathematics Collaboration of the week(soon to "of the month") is getting an overhaul of sorts and I would encourage you to participate in whatever way you can, i.e. nominate an article, contribute to an article, or sign up to be part of the project. Any help would be greatly appreciated, thanks--Cronholm144 22:34, 13 May 2007 (UTC)

[edit] Dab pages

Please read the MoS concerning disambiguation pages. --Mel Etitis (Talk) 15:44, 28 May 2007 (UTC)

~~Please link to what you want me to read. What MoS?~~ Apparently you mean Wikipedia:Manual of Style (disambiguation pages). However all examples there have brief verbal descriptions as my version [3] of Realisation (and pretty much any other DAB page I've seen) not a bare list of links like your edit [4]. Please explain. Jmath666 17:57, 28 May 2007 (UTC)

It's not the descriptions (which are optional, and are often not used), but the multi-linking within each entry, as well as the dubious extra entries. You also removed, without explanation, the Britis-English alternative spelling. --Mel Etitis (Talk) 18:08, 28 May 2007 (UTC)

OK. What British-English alternative spelling? I could not find that place. Jmath666 16:04, 29 May 2007 (UTC)

[edit] July 2007

Please refrain from making unconstructive edits to Wikipedia, as you did to Haha. Your edits appear to be vandalism and have been reverted. If you would like to experiment, please use the sandbox. Thank you. Gscshoyru 13:54, 13 July 2007 (UTC)

Sorry about that. Someone got hold of my laptop for a moment. Jmath666 15:03, 13 July 2007 (UTC)

No problem... be a little more careful I guess. Gscshoyru 15:16, 13 July 2007 (UTC)

[edit] Michael P. Drazin

Hi Jmath. Thank you for the new Michael P. Drazin article. I have a note. It is good that articles be wikified (relevant concepts should be linked) and that they are put in a relevant category (Category:Mathematics is too generic, and should be reserved to very fundamental articles). Also, the first occurrence of the article title should be bolded. Small things, but I thought I'd let you know. Thanks. You can reply here if you have comments. Oleg Alexandrov (talk) 20:07, 6 September 2007 (UTC)

Hi Oleg I meant to drop you a note to thank you that you have wikified along the lines you mention generalized inverse and Drazin inverse. I must confess that I hoped someone who remembers those things (I took a wikibreak and don't) will fix it and you did within few hours, each time. I just could not stand the mess that was there around pseudoinverse a week ago and so I wrote the skeleton anyway even if I did not have the time to re-learn those things... Well with Michael P. Drazin your patience has run out. If I start new articles often enough to remember those things I will mend my ways. Thanks again for the help. Jmath666 03:28, 7 September 2007 (UTC)

Heh. :) Oleg Alexandrov (talk) 15:57, 7 September 2007 (UTC)

[edit] Thanks

Thanks for adding the section on numerical computations to the null space article. It's well-written and quite informative. Jim 18:16, 16 September 2007 (UTC)

[edit] Wikiality.com

Please stop adding this external link to various articles claiming that it is operated by Stephen Colbert. It's not. —bbatsell ¿? ✍ 05:18, 18 October 2007 (UTC)

Two, in fact. Thank you for the information. Can you please correct/reword my entries. Jmath666 05:57, 18 October 2007 (UTC)

[edit] DYK

On November 1, 2007, Did you know? was updated with a fact from the article Bramble-Hilbert lemma, which you created or substantially expanded. If you know of another interesting fact from a recently created article, then please suggest it on the Did you know? talk page.

Kindly nominated by PFHLai. We always need more maths article so feel free to self nominated.Blnguyen (bananabucket) 07:37, 1 November 2007 (UTC)

[edit] INSPEC or Inspec

Hi. It appears that INSPEC should be Inspec, as on the Inspec page at IET. I just wanted to check that you are ok with us making that change, as you are the main contributor to the article. Nurg (talk) 05:23, 4 February 2008 (UTC)

Yes, you are right, thanks. I checked at the search page I normally use also. Jmath666 (talk) 08:40, 4 February 2008 (UTC)

Except the title still had the wrong capitalization and I could not move it because a redirect was there. So I had an admin swap the the redirect and the page, OK now. Jmath666 (talk) 09:14, 4 February 2008 (UTC)

Good on you. thanks a lot. Nurg (talk) 09:43, 4 February 2008 (UTC)

[edit] Gauss-Newton algorithm

I have reverted your recent edits of this article because of issues which have been discussed extensively on Talk:Gauss-Newton algorithm. There is a difficult problem in reconciling the use of GN for data fitting and for minimizing a sum of squared function values. While the two applications use the same updating formula, their properties are different because of the different assumptions made when deriving that formula. Therefore we have chosen to present the topic in two separate sections with notation specific to each application, rather than try to present GN as a single algorithm. In the past that was attempted, but it always came down to conflicting requirements.

Regarding local and global minima, this is an issue connected with multiple minima and not specifically with the Gauss-Newton algorithm. There is a section on multiple minima in non-linear least squares. The sections on convergence properties are intentionally vague about the nature of the minimum. This, too, is a case where data fitting and what you call general optimization are different. In data fitting multiple minima are rare except when fitting trigonometrical models. In general optimization they seem to be the rule rather than the exception. So here, too, a generalized discussion is inappropriate.

I hope you will appreciate that this is a more complex issue that would appear at first sight and that the reversion is justified at the present time as the principal protagonists in the discussion are not yet in full agreement. My home page Petergans (talk) 12:01, 15 March 2008 (UTC)

So, what exactly are those assumptions that guarantee single minimum and convergence to it in the data fitting case please? And can you please prove it or refer to a proof? If not, it should not be claimed as truth or alluded to. The issues are not complicated at all. You have local quadratic convergence in either case. The r's have a special form in the data fitting application. That's all there is to it. I think I have succeeded in making it a bit less confusing and vague though I did not go all the way. I would appreciate constructive edits instead of a revert. Thank you. Jmath666 (talk) 18:55, 15 March 2008 (UTC)

I apologise for the reversion; it seemed like the only option as your edits interfered with and in part negated an ongoing process of constructive edits.

I think that opening the convergence section with "not guaranteed" is too strong. Speaking pragmatically, in the vast majority of cases GN converges smoothly. I have plenty of experience of non-convergence in determination of equilibrium constants. They occur when the model is inconsistent with the data. In that application model selection is a major concern, so inappropriate models have to be tried in order to find the best one. The main cause of failure is "normal equation singular". In turn that means very high correlation between parameters. Thus, the root cause of non-convergence is a deficiency of the physical model in relation to the experimental data. With sound model/data systems convergence is "guaranteed" (in the common English sense). I have distributed our software worldwide on that basis for more than 30 years, with hundreds of resulting publications listed in citations index.

Some more background: I have throughly revised least squares, linear least squares and non-linear least squares with minor revisions to linear regression and nonlinear regression. There is an issue of consistency with those article that link to Gauss-Newton. I hate it when I see apparently contradictory things in linked WP articles. Petergans (talk) 09:52, 17 March 2008 (UTC)

Peter, no problem. I gave it the same treatment I would any other piece of technical writing and edited for clarity. Sure, go ahead and say first that the method converges in the vast majority of cases, and only later that it is actually not guaranteed to converge. Please understand we are coming in as mathematicians and our training dictates to make it clear when something is not a mathematical statement. This is also the level of accuracy expected in math part of Wikipedia. So, if a method fails to converge in just one instance, no matter how construed, and no matter how often it works otherwise, for a mathematician the statement "the method converges" is false. That's why mathematicians qualify their statements by loads of carefully worded assumptions. So we need to say something like "the method works in the vast majority of cases encoutered in practice, however convergence is not guaranteed", or "it may not converge for an arbitrary f", or something like that.

I understand that the discussion of the reasons why the methods converges and when is different in the two application areas. However, that does not make it two different algorithms. So, why not have one description of the algorithm followed by a comprehensive discussion? Jmath666 (talk) 16:01, 17 March 2008 (UTC)

Thank you for uploading Image:Robert Llewellyn 2004.jpg. However, it currently is missing information on its copyright status. Wikipedia takes copyright very seriously. It may be deleted soon, unless we can determine the license and the source of the image. If you know this information, then you can add a copyright tag to the image description page.

If you have any questions, please feel free to ask them at the media copyright questions page. Thanks again for your cooperation. NOTE: once you correct this, please remove the tag from the image's page. STBotI (talk) 01:18, 3 April 2008 (UTC)

[edit] Linear least squares

Jmath, in an edit to talk:linear least squares earlier this month, you are guilty of a commonplace confusion. Linear regression is linear regardless of whether the dependence on x is on the one hand linear or affine, or other other hand more complicated. The term "linear" in "linear regression" is not about the nature of the dependence on x. If you fit a parabola

$y = ax^2 + bx + c\,$

by ordinary least squares, then that's linear regression. The dependence of the vector of least-squares estimates of a, b, and c upon the vector of y-values is linear. This is linear regression regardless of whether you're fitting a line or a curve. That is standard terminology and there are good reasons for it. Michael Hardy (talk) 15:03, 20 April 2008 (UTC)

My point exactly in Talk:Linear_least_squares#Linear least squares is not simple linear regression, and in response Oleg has replaced the picture to reflect that. Please see also simple linear regression for context. Jmath666 (talk) 18:36, 20 April 2008 (UTC)