User talk:Glrx

Welcome!

Hello, Glrx, 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:

• The five pillars of Wikipedia
• Tutorial
• How to edit a page
• How to write a great article
• Manual of Style

I hope you enjoy editing here and being a Wikipedian! Please sign your messages on discussion pages using four tildes (~~~~); this will automatically insert your username and the date. If you need help, check out Wikipedia:Questions, ask me on my talk page, or ask your question on this page and then place {{helpme}} before the question. Again, welcome! Ray^Talk 19:29, 2 August 2009 (UTC)

Reed-Solomon Codes

Hi Glrx, and thanks for your efforts on Reed-Solomon codes. However, I want to point out that you removed a concise description of how RS codes essentially work, namely by oversampling a polynomial. Even though that statement could have been expanded, it was clear. The text you added describes RS codes from the point of cyclic codes. Furthermore, what you essentially say is that an error can be detected if the received code word is not divisible by the generator polynomial, which is... trivial from a coding point of view, but does not provide the casual reader with any insight. Furthermore you lead the reader to believe in a tight connection with CRC codes, while the actual connection is with cyclic codes. Last but not least, it is actually true that RS codes were not implemented in the early 1960s because of their complexity—it _might_ have been possible to actually implement on some hardware, but nobody did it back then. As far as history tells, RS codes were not implemented until Berlekamp came up with his efficient decoding algorithm together with Massey, after which they were implemented in the Voyager 2 space probe. To summarize, the description that you have given is better placed at cyclic codes, and mathematical descriptions, if added, are better placed in the Mathematical formulation section. Cheers, and keep up the work! Nageh (talk) 18:19, 28 March 2010 (UTC)

Just checked your latest edits. :) So let me follow up to my previous comment. The very basic idea of RS codes is to construct a polynomial over the k message symbols and evaluate it at n > k points such that the polynomial defined by the n points is overdefined. The receiver, after finding out erroneous received points, can then use interpolation to recover them. I think so far it applies to both original and classic constructions. How exactly these points are created, and how decoding is done, can be explained later on in the article. But in both cases sample points are transmitted, and certainly not coefficients (at least that can't be true for the remaining n-k values). As far as I can see the different lies therein that in what you describe the classic view the source symbols are taken as coefficients in the polynomial, and then this polynomial is evaluated at n points (created by a generator of a finite field) and their values transmitted. I'd have to look up the details, which I'm too lazy to, but so much just as a hint for the moment. :) Nageh (talk) 21:01, 28 March 2010 (UTC)

reply

Hi Glrx, and thanks for your efforts on Reed-Solomon codes. However, I want to point out that you removed a concise description of how RS codes essentially work, namely by oversampling a polynomial. Even though that statement could have been expanded, it was clear.

I disagree that it was clear. Although RS arrived at their code from a an oversampled polynomial viewpoint, that statement is not clear but rather terse. Furthermore, the oversampled view doesn't comport with modern usage. The modern g(x) viewpoint makes s(x) disappear and lets the error correction focus on just n-k syndromes rather than interpolating polynomials. I reworked the introduction to follow the RS development after your comment, and now I'm unhappy with it -- it lead me into the same trap that I was trying to fix: describing stuff that distracts. I fell into restating the history. The goal should be to explain the code and give insight into how it works. The modern implementation is the BCH viewpoint and transmits coefficients and not values.Glrx (talk) 21:19, 28 March 2010 (UTC)

The text you added describes RS codes from the point of cyclic codes. Furthermore, what you essentially say is that an error can be detected if the received code word is not divisible by the generator polynomial, which is... trivial from a coding point of view, but does not provide the casual reader with any insight. Furthermore you lead the reader to believe in a tight connection with CRC codes, while the actual connection is with cyclic codes.

I mentioned the CRC processing to build an analogy. I deleted it, and now I'm sorry I did. It also gives context for error correction algorithm using the roots of g(x).Glrx (talk) 21:19, 28 March 2010 (UTC)

Last but not least, it is actually true that RS codes were not implemented in the early 1960s because of their complexity—it _might_ have been possible to actually implement on some hardware, but nobody did it back then. As far as history tells, RS codes were not implemented until Berlekamp came up with his efficient decoding algorithm together with Massey, after which they were implemented in the Voyager 2 space probe.

I don't understand this comment at all. I deleted a clause that claimed the digital hardware was not advanced enough at the time and left the clause about no practical decoder. The reason the codes were not implemented is because the decoding algorithm was impractical (even on modern hardware) for a large number of errors. If there were a practical decoding algorithm in 1960, there was hardware to do it. Your statement agrees with that assessment, so what does it want? Does it want to keep the inadequate digital technology clause because it may have been possible to implement impractical algorithms in 1960 hardware?Glrx (talk) 21:19, 28 March 2010 (UTC)

To summarize, the description that you have given is better placed at cyclic codes, and mathematical descriptions, if added, are better placed in the Mathematical formulation section. Cheers, and keep up the work! Nageh (talk) 18:19, 28 March 2010 (UTC)

reply (2)

Consider my first post canceled, I wrote it spontaneously. However, my second post should have been more clear.

Ok, first, it is still a common introduction to RS coding that it works by oversampling a polynomial, and this is true no matter whether you use classic or original encoding... how it is implemented in practice is another topic.

Second, I have already explained that it doesn't make sense if you say "coefficients are transmitted". You cannot sensibly transmit more than k coefficients of a polynomial of degree k.

Regarding the CRC analogy, the CRC is a cyclic code, and a correct analogy would refer to cyclic code upon which BCH/RS codes are based.

Please ignore my comment on insufficient hardware capability - I was nit-picking.

Last but not least, would you please refrain from implying that it was my fault that you wrote the text in the for you unsatisfying way you did?

Nageh (talk) 21:57, 28 March 2010 (UTC)

reply to reply

Ok, I checked the references, obviously I was only recalling the original understanding of RS codes. I apologize.

What I suggest is to move the introductory text on original vs. classic RS codes further down, into section Definition, which could be renamed to Introduction. Subsection Mathematical formulation and Reed–Solomon codes as BCH codes could be integrated into the two views, possibly covered in subsections Original view and Classic view. Remarks should go as the last subsection. Please note that Reed–Solomon codes as BCH codes denotes f_i as the coefficients in BCH code notation. By a slight reformulation, some statements could be generally given. Note [4] applies to RS as BCH codes, and I don't see why this is a requirement for the original view. Note [5] is correct if you replace f(x_i) by f_i, as defined in Reed–Solomon codes as BCH codes, and in fact can only be valid for such constructions (i.e., the classic view).

In the lead section, I would propose a very simple introduction, such as:

In Reed-Solomon coding, source symbols are viewed as coefficients of a polynomial P(x) over a finite field. Whereas the original idea was to create n code symbols from k source symbols by oversampling P(x) at n>k distinct points, and at the receiver use Lagrange interpolation for each possible k-size subset of the received n values to recover the closest match, this does not lead to an efficient decoding algorithm. The concept was soon replaced by viewing the RS code as a cyclic code, and instead deriving the code symbols from the coefficients of a polynomial constructed by multiplying P(x) with a cyclic generator polynomial. Wesley Peterson devised an early decoding algorithm for this classic view of RS codes. A practical solution was eventually found in 1969 by Elwyn Berlekamp and James Massey, and is since known as the Berlekamp-Massey decoding algorithm.

Sorry again, and keep up the good work! :) --Nageh (talk) 07:37, 31 March 2010 (UTC)

Harvard references

Hi, you added references to Berlekamp 1968 and Massey 1969 to the article. Could you, please, put the complete descriptions of what they refer to to the section “References”, so that it is clear what they mean? Thanks. Svick (talk) 20:40, 29 March 2010 (UTC)

Thanks. Svick (talk) 19:12, 30 March 2010 (UTC)

OFX and QIF Pages

Hi - you removed my edits the pages above, however, the links that I placed in the external links section are relevant to the titles. I am not sure why you believe it is a spamming effort on my part. Nepaluz (talk) 16:35, 3 May 2010 (UTC)

Help

I noticed you undid my contributions, Why? If you look at "Export Hacks for QuickBooks: exporting to QIF" above, they have the same kind of solution and a link to "XL2QIF Excel macro" And in References you have all links to similar websites: ▪ Official Specification ▪ How to parse online bank QIF with sed ▪ Python script to convert QIF to RDF. Look for def extract(path). Script provided by SWAP. ▪ Perl module to parse and create QIF files. ▪ GnuCash QIF format notes ▪ Online QIF to CSV Converter ▪ .NET QIF API What is wrong with adding my solution? Besides I am giving it for free to help people with same problem —Preceding unsigned comment added by Marcellovani (talk • contribs) 19:22, 6 April 2010 (UTC)

Your additions in Quicken Interchange Format were an advertisement for your code. Your contributions do not explain the Quicken Interchange Format -- which is the topic. The references that you cite have not insinuated themselves into the main article. Many of those references offer detailed information about QIF and could easily have been used as source material for the article. Someone seeking more information about the format could find more information. I haven't checked them all, but following a few links offers shows that they offer information and even source code. Your link does not offer information; it offers a service. It doesn't matter that it is a free service, it is still advertising rather than content. I'm pretty sure you understand that point. The Online QI to CSV Converter is a questionable link that probably should be removed.Glrx (talk) 19:59, 6 April 2010 (UTC)

Some thoughts about Wikipedia and original research

Thanks for your comments on my thinking! I'll post a reply within the next days, currently I'm a little bit busy. Nageh (talk) 15:45, 7 June 2010 (UTC)

I have marked you as a reviewer

I have added the "reviewers" property to your user account. This property is related to the Pending changes system that is currently being tried. This system loosens page protection by allowing anonymous users to make "pending" changes which don't become "live" until they're "reviewed". However, logged-in users always see the very latest version of each page with no delay. A good explanation of the system is given in this image. The system is only being used for pages that would otherwise be protected from editing.

If there are "pending" (unreviewed) edits for a page, they will be apparent in a page's history screen; you do not have to go looking for them. There is, however, a list of all articles with changes awaiting review at Special:OldReviewedPages. Because there are so few pages in the trial so far, the latter list is almost always empty. The list of all pages in the pending review system is at Special:StablePages.

To use the system, you can simply edit the page as you normally would, but you should also mark the latest revision as "reviewed" if you have looked at it to ensure it isn't problematic. Edits should generally be accepted if you wouldn't undo them in normal editing: they don't have obvious vandalism, personal attacks, etc. If an edit is problematic, you can fix it by editing or undoing it, just like normal. You are permitted to mark your own changes as reviewed.

The "reviewers" property does not obligate you to do any additional work, and if you like you can simply ignore it. The expectation is that many users will have this property, so that they can review pending revisions in the course of normal editing. However, if you explicitly want to decline the "reviewer" property, you may ask any administrator to remove it for you at any time. — Carl (CBM · talk) 12:33, 18 June 2010 (UTC) — Carl (CBM · talk) 13:22, 18 June 2010 (UTC)

What are you doing?!

You reverted one of my edits on 20 June 2010 to Demurrer. Are you even a lawyer?! Any competent attorney is aware that in most tort actions, the plaintiff generally wants to get in front of a jury and razzle-dazzle the jury into awarding gigantic punitive damages, while the defendant wants to keep as FAR away from the jury as possible. Only in certain commercial transactions does it make sense for a plaintiff to move for summary judgment in their favor, in which case they need to use a demurrer to knock out defenses in the answer and thereby limit the defendant's ability to manufacture a triable issue of fact. --Coolcaesar (talk) 03:14, 27 June 2010 (UTC)

Litz wire

Fix Litz wire.

Explanation?

You reverted all my edits to B-tree? Why? Your edit summary said, "Intervening edits well meaning but with technical errors about rotation of entries." I'm not sure what you are reffring to with the phrase "rotation of entries", though I can believe my improvements in the sections "Deletion from a leaf node" and "Deletion from an internal node" might have included some technical errors. After all, my edit summary for the former was: "Rewrite (based on my understanding of the cryptic prior version & the article overall. I think this is right, but if I made any errors, I apologize; please correct.)" However, "please correct" is not the same as, "please throw away all my work if this isn't quite right in all details." I was kind of hoping if someone found errors, they would /fix/ them, not revert to the old version which others wrote long ago on the talk page was practically incomprehensible.

Anyway, I really would like to know what is wrong, because I initially came to this page to learn about B-trees (having plenty of knowledge about other kinds of trees, linked lists, sorting algorithms, file systems, etc.)

But more than that, I am flabergasted that you appear to have reverted all my edits without reviewing them separately, trashing even the little grammatical ones that couldn't possibly have been degradements of the article in any way! Can it be so? Maybe you'd like to look back and make a more selective and more conservative rejection of my changes.

I think I'll reinstate some of them myself, now. I'll leave "Deletion from a leaf node" and "Deletion from an internal node" in the sorry state I found them in (twice now) until some consensus comes to the matter.

141.158.233.114 (talk) 11:46, 23 July 2010 (UTC)

Yes, your edits contained technical errors. I don't think it is appropriate for any editor to submit material that he is unsure of and then ask others to correct it.

I looked at each of your edits, and I did not think that they were appropriate. That's why I reverted all. I've left some of your reinsertions, but I'm unhappy about them. A filesytem is a database, so the in addition comment is technically wrong. Grammatically, I prefer the active voice. I also prefer conciseness.

Glrx (talk) 16:23, 23 July 2010 (UTC)

Power Factor

Glrx: You undid my addition of a link to the Power factor page and gave no explanation as to why you did this. Wtshymanski then removed the entire section I added. I have added to my user page but Wtshymanski has not been willing to respond. I am new to contributing to Wikipedia pages and am trying to learn the ropes quickly. I appreciate the spirit and goals of Wikipedia and do not want to degrade their mission. I am a physicist who has probably spent more hours over the last 30 years designing and experimenting with active power factor compensation than anyone else. I have founded 3 companies that design, manufacture and market this equipment and am currently the CEO of one of these companies so conflict of interest is definitely a possibility here. I am a recognized expert and innovator on this subject and was recently asked to give a presentation at EPRI's annual conference which took place in Quebec last June. My work is mostly a labor of love as I am genuinely trying to contribute to solving the world's electrical power problems so that we evolve into something that is more sustainable. To do this, my companies have to become very successful financially and that has certainly been the case with Heart Interface, Trace and Xantrex (now Schneider) which all just reeks of conflict of interest. However, the world has changed and power factor, which used to be a more esoteric engineering subject, has become something that is effecting most people's lives in ways they don't understand. Understanding this, by people with no engineering background, is a legitimate function of Wikipedia. Although the grid is a complex mix of loads, each load, by itself, consumes some small amount of power and creates some small amount of transmission loss in the entire system, including the windings of the massive grid generators and transformers. The grid has to pay for all of this energy and the losses so you can see that low power factor loads cost the grid more per kWh to provide than do high power factor loads. With the massive migration to CFL lighting and the great abundance of computer power supplies the grid is having to absorb increased costs which they ultimately pass on to the customer. There are some articles about this that appear in EDN and EETIMES and other engineering publications but the general public is generally not aware that this is even an issue yet they are the ones who are buying and installing these low power factor devices. Since they are ignorant of the issue they do not create any pressure to address it. So, if they somehow hear that it might be an issue and don't even know what "power factor" means they might go to Wikipedia. If all they get is a very technical, engineering and very correct and precise article, they will most likely not even read it. So, what might seem redundant to someone educated in electronics, might be the only thing that a non-engineer even reads. What are the guidelines to adding external links and references at the end of the article? Please respond. Heart141 (talk) 13:41, 27 July 2010 (UTC)

This is same material as Talk:Power factor#Power Factor for the Modern World. Glrx (talk) 05:11, 28 July 2010 (UTC)

Genetic algorithm

Hi

In the interests of averting an edit war - There has been mucho discussion on the talk page already so how do you see the matter as progressing ?

Chaosdruid (talk) 05:18, 28 July 2010 (UTC)

At the current point in time, I don't see a consensus. Do you see one?

My position (and I believe it is also OF's position) is that KB's work is not prominent enough yet. Quite simply, KB is engaging in self promotion. This case is not about a neutral editor adding the GFH to the article.

It is my understanding that KB is the sole proponent of his research. Although he gave citations to his work by others, those citations did not show acceptance (authors are not adherents let alone prominent adherents). I found no other references that use the term GFH that KB coined. I have not found any secondary sources that address his work. His GFH has not yet gained acceptance or controversy. He does not have significant stature to support the view.

KB's viewpoint is not currently held by a significant minority. It is a viewpoint held by an extremely small minority, so it does not belong in Wikipedia. It does not have weight. I disagree with your notion that the dissertation committee should be considered adherents.

The discussion, with ample pushing by KB, has wandered away from what an encyclopedia should say to a narrow, almost legal, interpretation of isolated sentences. There are various WP policies concerning coverage of a topic. KB reads those policies narrowly to argue for inclusion. His basic argument boils down to a PhD thesis is peer reviewed and published and therefore WP:V so it should get a reference. OF's talk page went into that argument. A commenter raised the more apt WP:UNDUE. I do not see KB's argument as persuasive at all. Theses do not get the same peer review as refereed journals. There haven't been secondary source reviews.

I fear that many editors did not comment on the RfC because they were scared off by KB's argumentative posture. GA has 192 watchers but few commented. Look at the length of the discussion on OF's talk page. KB won't let go. That discussion made me think long and hard about making any comment. I've been ignoring KB's entreaties because they are just openings for argument. That's not what a RfC should be about. It should be other editors commenting freely about the proposal.

WP:SCOIC states that "Once you have presented your case, it may be best to take a back seat and minimise your participation in the discussion, to avoid any perception or accusation of undue advocacy or pushing an agenda." KB's "sharp" advocacy speaks for itself.

KB has a significant COI that should not be ignored. I believe he is too close to his work to offer a reasonable perspective.

KB's edit history of the main article does not show neutrality.

The proposed edit is premature and should not be made.

What do you think should happen?

Glrx (talk) 18:44, 28 July 2010 (UTC)

Ok first of all lets just say that attacking me is not really constructive. Your comment "I disagree with your notion that the dissertation committee should be considered adherents." is way off the mark and quite honestly a little ridiculous. I never said they were adherents - I stated quite clearly that "The people that are included on the dissertation committee are pretty notable though"

I will respond to your comments when a suitable period of time has passed - maybe an hour or two so that you can re–read the posts and make sure that you can respond in a less aggresive manner as this seems to be in a knee-jerk rather than constructive mood. Chaosdruid (talk) 18:53, 28 July 2010 (UTC)

Sorry for the delay but a broken van at work left me stranded there overnight.

The COI is not really an issue. I think that it should be included but more as (the parts in "" are for wording to be decided later):

"An explanation of the paradox of "these problems of unexplained" results has been put forwards by the theory of GFH by Burjorjee. "brief explanation of no more than 20 words but it remains to be seen if this is accepted by the "GA community""

I think that would give no more undue weight than it deserves.

Chaosdruid (talk) 16:07, 30 July 2010 (UTC)

I disagree. COI is an issue, and it seeps into KB's edits. Without a secondary source evaluation of the GFH, the material does not belong in WP. Who are the prominent adherents of GFH? For a viewpoint to be included in WP, it should be easy to identify them. If the GFH is not being accepted or at least evaluated by others in the field, then why should it be in an encyclopedia? Glrx (talk) 17:08, 30 July 2010 (UTC)

Glrx, in the interest of resolving this issue, are you willing to enter into mediation (WP:RFM)? Please reply. Keki Burjorjee (talk) 19:48, 29 August 2010 (UTC)

Hello, Glrx. This message is being sent to inform you that there currently is a discussion at Wikipedia:Wikiquette alerts regarding an issue with which you may have been involved. Thank you.

Citations

Hi, A general question, these fancy multi-line citations like in Beta encoder, what do they buy? It seems to me they make the text hard to edit. What do they buy for you guys - several people do them. Thanks. History2007 (talk) 18:42, 17 September 2010 (UTC)

Citation macros buy several things. First, they make the format of citations consistent across articles. Second, they help highlight missing information in the citation. The beta encoder article's citations are missing author, title, and volume information. What, exactly, is the Mathematical Review citation? Third is a chance for automation. If some text is labeled as a citation, then a bot can process it. Glrx (talk) 18:55, 17 September 2010 (UTC)

I like the 3rd advantage, namely bot automation. The 1st advantage does not do much for me, but I guess some people like that. I added the Googlebooks link to that anyway. I think the future is in Gbook links. Anyway, I wish the bots would make the thing on ONE line, not several, becaus eit makes teh text so hard to edit afterwards. History2007 (talk) 19:47, 17 September 2010 (UTC)

Wikipedia:Mediation Cabal/Cases/2010-08-29/Genetic algorithms

I have offered to mediate this. Please indicate your acceptance or lack on the mediation page. Thanks. Hipocrite (talk) 14:27, 4 October 2010 (UTC)

May I ask why I'm not acceptable to you? Thanks. Hipocrite (talk) 20:30, 5 October 2010 (UTC)

Sort page edits

I contacted the author about using these to augment the lack of implementations (in general languages) on the sort pages and he gave his permission as well as adding a notice to the original page. I would appreciate the edits being rolled back. 108.27.89.177 (talk) 15:37, 4 October 2010 (UTC)

I rolled the edits back because the source issue has been resolved. Please post back here if you have additional concerns. 108.27.89.177 (talk) 17:19, 4 October 2010 (UTC)

We don't need multiple versions of the same algorithm. I'm also tempted to roll back selection sort because I don't like the promotional aspects of the GPL. The algorithm is trivial. Glrx (talk) 17:34, 4 October 2010 (UTC)

The algorithm may be trivial however there is currently no actual code listing. Would you suggest another license other then GPL? Also you rolled back the insertion sort however there are two implementations there, one in Pascal and another in straight pseudocode. Shouldn't the Pascal one be pruned off then? 108.27.89.177 (talk) 18:09, 4 October 2010 (UTC)

I replaced the bad JS implementation of selection sort; it didn't follow the text and it was unstable. I'm not going to favor one programming language over another. One good, clear, implementation should be enough. The GPL requires attribution, and attribution for trivial algorithms sounds in promotion. Glrx (talk) 18:20, 4 October 2010 (UTC)

I'm not sure how it could be described as unstable. There are languages where the implementation will be more clear then others. 108.27.89.177 (talk) 18:45, 4 October 2010 (UTC)

The JS selsort imp was not a stable sort. Glrx (talk) 18:52, 4 October 2010 (UTC)

Continuing this, why are now all the sort page links removed? Is there a reason for this mass deletion of them? Are they not applicable? Josh Kehn (talk) 23:57, 9 October 2010 (UTC)

Toroids Pictures

I put the pix in the public domain. Thay can be used anywhere. —Preceding unsigned comment added by Constant314 (talk • contribs) 03:20, 20 October 2010 (UTC)

Toroidal inductors and transformers, Changing Citation Style

I'm just curious why you changed the citation style. I'm always looking for a better way to do things; that is why I am asking.

I hate generating a new reference that is the same as a preceding reference except on a different page; that is why I chose the style that did. But if I could accomplish the same thing with a different style I'd be glad to adopt it. Constant314 (talk) 17:32, 24 October 2010 (UTC)

Unfortunately, there isn't a great way to do citations; different methods have strengths and weaknesses.

I took your references and converted them to {{Citation ...}} macros. That enforces a uniform style on the citations; it also helps identify missing information. See WP:Citation templates.

For frequent citations to authorities at different page numbers, Harvard citations are useful. The Harvard citation macros (sort of) hyperlink to the Citation macros (they can also link to Cite macros if the ref field is done right). In technical articles, I use parenthetical inline citations with {{harv|Author|2010|p=13}}. To some readers, inline citations are a little annoying. IIRC, you were putting the citations in footnotes, so I kept your method and used the {{harvtxt ...}} variation to fill in the footnote text. See WP:Harvard citation template examples.

Multiple citations to the same page are also awkward. They can be handled by giving a

<ref name="unique name">

. That way only one reference line is generated for similar citations. The drawback is the user must maintain the linkage. If the reference with the information is deleted, then the info is lost for the other citations. That's a reason to avoid using a ref with name="identifier" when it is only used once - if an editor is going to delete a named ref, then he's supposed to fix its other uses. If the ref doesn't have a name, then the editor doesn't have to worry about it.

Glrx (talk) 20:10, 24 October 2010 (UTC)

Hash tables

Could you explain how keys in a hash table are still subject to out-of-bounds errors? It would also be useful to define "well-dimensioned" (well-sized?).

138.16.3.129 (talk) 20:56, 25 October 2010 (UTC)

Re Hash table

In two ways. First, the index into the table must not exceed the bounds; often met with a modulo table size in the hash, but the programming consideration must still be there. Second, one cannot keep adding entries to a finite table -- at some point there's no more room (the logical equivalent of an overflow).

Well-dimensioned is not precise. When the table is full, there are lots of collisions, and that means the access does not take constant time. Generally, a hash table should perform well when it is half full. Some can perform well at 80%.

Glrx (talk) 22:41, 25 October 2010 (UTC)

Oh well I was thinking in the key-space; the hash function makes sure that the hash of a key can't be out of bounds in the array backing the hash-table. No? —Preceding unsigned comment added by 138.16.3.129 (talk) 02:42, 26 October 2010 (UTC)

"jk blog links"

Glrx- Is there a issue of sorts with this page?

Baltar, Gaius (talk) 18:12, 28 November 2010 (UTC)

See edit history of Sorting algorithm for 9-10 October 2010. Links to his blog were also removed on other sort pages. Discussion at User talk:Josh Kehn#Your sorting links. Glrx (talk) 06:09, 29 November 2010 (UTC)

The issue is the sort page is political? I just found it on stackoverflow.com. In order to include this link what do i have to do?

Baltar, Gaius (talk) 17:04, 29 November 2010 (UTC)

If you want to include the links, bring the subject up on the article's talk page. Supply links to the existing discussion. State why you think the links are appropriate for WP. I'd like to see your argument why WP should tolerate links that hang a user's browser or denigrate the use of IE. Glrx (talk) 17:45, 29 November 2010 (UTC)

I will do that.

Baltar, Gaius (talk) 21:45, 29 November 2010 (UTC)

Quadratic equation edits

I was just wondering what was wrong with adding the method to get from the quadratic equation to the quadratic formula in the section quadratic formula. I believe that, in spite of the fact that it's a derivation, the majority of people would still look in the section quadratic formula rather than derivations for this information. (I'm assuming here, though, that most people who need to look up the quadratic formula would not know it's a derivative of the quadratic equation. I'd also suggest adding in that section something along the lines of, "The quadratic formula is simply a derivative of the quadratic equation." —Preceding unsigned comment added by 67.86.109.242 (talk) 20:56, 2 December 2010 (UTC)

Such a discussion belongs on the article talk page -- not here. The material was already in the article, and the detail was not needed at that point in the article. Glrx (talk) 02:15, 5 December 2010 (UTC)

Growler37 and external links

They should have used an edit summary to explain what they were doing. It turns out that http://www.beaconworld.org.uk is nothing more than a redirect to http://www.beaconworld.fotopic.net/. Other than photographs they don't appear to have any information on NDB's NAVTEX or Differential GPS. On the other hand the http://www.ndblist.info/ site does seem to have information on all of those. Cheers. Enter CBW, waits for audience applause, not a sausage. 15:00, 4 December 2010 (UTC)

Apology for undoing your edit.

I apoligize for undoing your recent edit on impedance matching. I was in the middle of a huge edit; it kept getting bigger and bigger, and although I tried to merge propperly, I did not see your edit.

Please have at now.Constant314 (talk) 06:07, 7 December 2010 (UTC)

That's OK; I've been there.

Changed Z₀ to Z_S because the xline might be the source or the load.

BTW, naming refs for Harvard citations is usually not needed. For non-Harvard citations, naming the ref means all the details of the citation need not be repeated. For a Harvard citation, it's no big deal to repeat the author/year. Furthermore, the page number usually changes from citation to citation, so named ref often used just once. Harv citations are often going to the specific page (i.e., they point directly to the cited material rather than pointing to an entire chapter).

Glrx (talk) 06:55, 7 December 2010 (UTC)

So instead of

<ref name="Feynman1515">{{Harvtxt|Feynman|1964|p=15_15}}</ref>

just use

<ref>{{Harvtxt|Feynman|1964|p=15_15}}</ref>

?Constant314 (talk) 17:11, 7 December 2010 (UTC)

Yes. Why write the page number twice. There's no good all round system. Harvard citations are effective when there will be many citations to the same reference but at different pages numbers. Moreover, there's no reason to add "name=<id>" to a ref if that citation is only used once. Glrx (talk) 18:31, 7 December 2010 (UTC)

I see you changed Z0 to Zs in the formula for the reflection coefficient, where Zs is understood to mean the source impedance. But in entire remainder of the section I use Z0 as the transmission line impedance. I'd rather use Zc for characteristic impedance, but the original article used Z0. So I propose to put it back like I had it and then add it back like you have it a little further down and call it a generalization. Constant314 (talk) 17:59, 7 December 2010 (UTC)

I think you're missing the point. The xline might be on the source end of the interface, or it might be on the load end (eg, a transmitter). Furthermore, when transmission lines are used for matching networks, there are often many different impedance levels. Z₀ is the usual notation for talking about just one impedance level (and more common that Z_c). To put a fine point on it, how would you describe the reflection coefficient when connecting a 50-ohm line to a 75-ohm line? Or vice-versa?

Although I applaud your efforts, I'm leaning toward reverting your last impedance matching edit.

Glrx (talk) 18:31, 7 December 2010 (UTC)

Let me try with Zc and see what you think.Constant314 (talk) 19:42, 7 December 2010 (UTC)

Yes, I am doing the specific example of the transmission line driving a load. I think it might get confusing if it is too general. But how about this: I make it clear that I am doing the transmission line to load interface and and a paragraph about the reflection coefficient at the source. Also, the first place reflection coefficient is introduced is a link to a more general discussion of reflection coefficient, so maybe I can stick to the single specific case? Regarding the 50/75 interface I would probably use Z1 and Z2, but I might use Z50 and Z75. Constant314 (talk) 19:55, 7 December 2010 (UTC)

I have intention of expanding the section I added about the source end, but I am burned out for today. Do whatever you think will make it better, if you wish. Constant314 (talk) 21:34, 7 December 2010 (UTC)

Your deletion of EL on IPv6

You reverted my EL contributions to the article on IPv6, perhaps you would be kind enough to explain why, since your brief note did not explain.

• IPv6 Act Now operated by RIPE
• Test your IPv6 connectivity operated by Jason Fesler, Senior Principal Architect & IPv6 Evangelist at Yahoo

Your only comment was Questionable links. Added content? Diagnostic appropriate?)

With all due respect, considering the standing of the sources, I fail to see why these would be questionable. Considering also the urgency of migration to Ipv6, this is a current issue requiring action on a broad front in the industry, I feel it is highly appropriate to include links to reputable sites that provide practical help on migration from IPv4 to IPv6. At the end of the day, readers want to know what, if any, changes they need to make to their systems, and these links are to resources to help them do just that.
Enquire (talk) 07:18, 14 January 2011 (UTC)

You are reversing the burden. It's your burden to justify your insertion of the links. There are guidelines for including an EL. Why do you think the ELs are on topic and appropriate? It is not my burden to explain why the links should be removed.

The IPv6 article is about the protocol suite. What IPv6 is, why it was developed, how it differs from IPv4, etc. The article has some detailed explanations and is supported with authoritative references.

The content on ipv6actnow.org is trivial. What vital information does the site supply about IPv6 that is not included in the article? (EL links are supposed to cover information that should go in the article but hasn't been included yet.) Furthermore, the thrust of the website is about converting from IPv4 to IPv6, and that is not the subject of the article. How to adopt IPv6? How to deploy IPv6? How to implement IPv6? How to plan for IPv6?

The test-ipv6.com link is similarly off topic. How to test IPv6 installations is not the subject of the article. The article is WP:NOTHOWTO guide about how to transition to IPv6. The EL is also an unneeded advertisement for Jason Fesler and Yahoo.

Don't confuse subject matter about a topic with how to do it.

Glrx (talk) 17:14, 14 January 2011 (UTC)

Cathode

Hello, I made the revision on cathode where the section was deleted. You stated the following, and if it's correct then I agree with you the deletion was misguided:

The above argument is confused about cell potential. To first order, the cell potential does not change whether charging or discharging; the designation of the positive and negative terminals does not change. In the Daniell cell, the copper electrode is positive and the zinc electrode is negative. The direction of the current does depend whether charging or discharging, so the designation of which terminal is the anode or the cathode changes with the mode.

However, it also means that the article could use the additional information of how a terminal is defined as positive or negative (or was originally). As the difference in definitions between (+) and (-) vs cathode and anode will make it more clear as to why there's a difference.

Therefore I made the following change at the end of the intro: The reason that the cathode can change designation (+) to (-), is that the terminal the cathode changes designation when the current changes direction while (+) and (-) do not. Take the Daniel electrochemical cell: as a galvanic cell the (+) copper is the cathode while zinc (-) is the anode, if it were elecrolytic (+) copper would be the anode and zinc (-) the cathode.

I think it would be preferable just to state the definition of (+) and (-) so it can be compared to the definition for cathode an anode, but this is the best alternative available to me (as I can't find how (+) and (-) are defined).

And thanks for correcting me!  :) Finniganawakens (talk) —Preceding undated comment added 02:42, 17 January 2011 (UTC).

Contested speedy deletion

re CTERA Networks

Respectfully, I contest your proposed deletion. This article topic is notable per Wikipedia:Notability (organizations and companies). "An organization is generally considered notable if it has been the subject of significant coverage in reliable, independent secondary sources." . This company is the topic of a chapter in "Cloud Computing Bible" , as well as articles in industry publications as ZDNet, VentureBeat, CRN, CTOEdge, Datamation, The Register, Network World, Channel Insider, LinuxPlanet, Small Cloud Builder, Network Computing, MSP mentor, etc. The article cites plenty of reliable sources independent of the subject. I invite you to look at my contributions list for yourself and see that I am a genuine editor, not a spammer. Frankly I have no idea what's wrong with the article, but I would be happy to collaborate with you on improving it, if you believe it needs improvements. Let me know if you accept my offer to collaborate. Marokwitz (talk) 18:35, 25 January 2011 (UTC)

Yes, there are many references in the article, but the references that I examined appear to be reworked press releases and not independent secondary sources. I also see passing mention rather than significant coverage. Furthermore, narrow publications are not given much weight.

Your other contributions are not at issue here. I did revert your Tape drive edit because it did not appear to have a WP:NPOV. The edit implied tape failures. I did not find the ref'd work, but I found other references to it that quoted the failure rate was only 60% and that the dominant reasons for restore failures were not tape failures but rather configuration errors or operator error. (FWIW, I am not pro tape.)

The article reads like a company brochure.

You should contest the speedy deletion on the article's talk page.

Glrx (talk) 19:24, 25 January 2011 (UTC)

The sources that I collected are certainly and definitely not reworded press releases. That's not true - they are reviews of products, a book on the topic of cloud computing, and interviews by respectable publications, reliable *secondary* coverage of the company. Among the 20 or so sources, there is a single press release cited. I would be happy to collaborate and resolve any issues that make it read like a company brochure, although personally it does sounds neutral to me, and uses similar wording as other technology company articles on Wikipedia. I think you are being unfairly harsh on me. Marokwitz (talk) 19:39, 25 January 2011 (UTC)

Quicksort

Hello Glrx. I was adding the information about the compiler to get accross the point that I added in just now. Let me know what you think. I just figured out what tail recursion was and I didn't understand it until I read parts of Sedgewick's paper. Specifically I didn't understand until I read the exact sentence I copied into wiki that you erased. I am hoping that my summary of the idea is closer to what you find appropriate for the article. I think it will help newbies understand. Nathan (talk) 04:41, 26 January 2011 (UTC)

Well, I'll applaud the effort, but I don't think the addition helps much. To me, the implications of tail recursion suggest as much, but those implications are not obvious to the casual reader. The subsequent (existing) comment about iteration confuses the issue (why speak iteration when QS is explained recursively; an iterative version still needs a partitioning stack, and that stack must have tail recursive management). Mentioning iteration is a monkey wrench. Your comment about the call stack addresses a piece of the issue, but also requires the reader to understand something that isn't clearly stated. The article doesn't say why adding to the call stack is an issue. Without RS's trick, the worst case call stack can be n deep: a bad partitioning at each level of m-1 and 1. With the trick (and tail recursion), the call stack is only log n (even though there are still O(n²) calls). Glrx (talk) 17:26, 26 January 2011 (UTC)

Asking for your review of an article I composed.

Hello Glrx,

First - thank you for your latest undoing of my edits. They were done in good faith, and your explanations in the edits helped me to understand better what to do in the future.

Secondly, recently I composed my first article, and I would be happy for your inputs on it:

http://en.wikipedia.org/wiki/Cramer_V

Best, Talgalili (talk) 17:48, 27 January 2011 (UTC)

Z3 argument

I will raise this point at the History of computer hardware article and perhaps some of them can talk to you instead. It is just wrong to go on edit warring against something which is written in black and white in a better venue without asking there. Dmcq (talk) 17:41, 8 February 2011 (UTC)

Andrea Doria

Hi, I've reverted your good faith reversion to the Andrea Doria article. Whilst the Italian Wikipedia is not a WP:RS, the links were not being used as references, but as interwiki links to the articles on the Italian Wikipedia where no articles exist on the English Wikipedia. This in in accordance with WP:LINK. See also H:ILL#Inline interlanguage links which explains this fully. Mjroots (talk) 06:30, 28 February 2011 (UTC)

My bad. I took out the links because they only offered text in Italian (eg, no images), the individuals' significance was covered in the English WP article, and a misunderstanding about foreign links. The links still seem a little odd to me. The right step seems to be link to stubbed en.WP articles that point to the it.WP articles. Anyway, thanks for reverting me. Glrx (talk) 01:35, 1 March 2011 (UTC)

LISP and Usenet

I removed that link too myself, but after some thought have reconsidered. Lisp well pre-dates the WWW, and the Usenet group is a long established "official" place for Lisp. If you read the top of ELNO, you will see that it makes exception for official sites. This is pretty much a long-established official site pre-dating the Web. Yworo (talk) 03:28, 2 March 2011 (UTC)

I also note that WP:OFFICIAL specifically states "website or other Internet service", which anticipates this situation. Yworo (talk) 03:31, 2 March 2011 (UTC)

I disagree. If you want to bring the link into the article, then open a discussion on the talk page to gather a consensus. How would such a link improve an encyclopedic article? What does it say that the article does not? Glrx (talk) 04:57, 2 March 2011 (UTC)

S/PDIF

You removed a quote in a reference. I add this quote for as a reference and not as a howto. This is a really a "hidden" statement. mabdul 01:49, 14 March 2011 (UTC)

Why should a "hidden" statement be added to an article? How does the quotation help the average reader understand the topic? The quotation is primarily of interest to those who want to connect different interfaces -- ie, those that want HOWTO information. I have no problem with the reference as a source. Glrx (talk) 01:57, 14 March 2011 (UTC)

arbitrary precision

Please state the reasons that you removed the references for arbitrary precision after 2007 and only kept the original 2 references (without publish year)? Wiki said Please help improve this article by adding reliable references after 2007. The new papers cited reflect recent development and applications of arbitrary precision in scinetific and engineering fields. —Preceding unsigned comment added by 220.128.221.62 (talk) 19:29, 9 May 2011 (UTC)

The refs in arbitrary-precision arithmetic were removed because they were not used / cited in the article. They are after-the-fact additions that were not used to compile the original article and do not have inline citations pointing to them.

The notice at the top of the article is not asking for references after 2007. It states that in 2007 the article was marked because its text did not have inline citations to references. Nothing requires those inline citations to be post 2007.

Wikipedia is not a how to do it manual. WP:NOTHOWTO One of the references (one that appears to have a conflict of interest WP:COI with the editor) is a how to exercise for a class. Some other references appear to have little relevance to the main article; the article is not, for example, interested in FPGA implementations.

Glrx (talk) 20:14, 9 May 2011 (UTC)

Dear Mr/Ms. Glrx, (1) Are you the editor of the page of arbitrary precision? (2) The papers written by academic researchers published in conference proceedings and journals as in the new references are very helpful to the wide reader in the world. (3) The texts can be edited, too, to reflect why those papers are necessary for readers with scientific and engineering backgrounds. Please read through all the papers before you brutally deleted them. (4) You seem to have a bad history of removing other people's editing, violating the value of Wiki. (5)I strongly object you being an editor if you are. Have not authored any journal or conference papers in the subjects of arbitrary precision? If not, please leave space for other experts. —Preceding unsigned comment added by Yuehwang (talk • contribs) 21:27, 9 May 2011 (UTC)

Monty Hall Problem.

Hi! Just wanted to mention that your edits to Monty Hall problem have really helped the article. Good work. Guy Macon (talk) 16:19, 13 May 2011 (UTC)

Thanks. I just edited the intro again in an attempt to simplify Gerhard's addition. Glrx (talk) 16:56, 13 May 2011 (UTC)

Glrx: But please consider that the host, knowing about the actual location of the car resp. of the three objects, in intentionally showing a goat, offers an overall probability to win by switching of 2/3, whereas the "uninformed" host in opening just "a" door, and indeed coincidentally by chance discovering a goat, already has reduced the probability to win by switching to just 1/2 only. This is true even if he actually shows a goat indeed and not the car. I clearly named the peculiar reason for that, but you deleted that reason. Nevertheless this peculiar reason should be shown "somewhere" in the article, helping the reader to grasp also that misty fact. Regards, Gerhardvalentin (talk) 15:25, 16 May 2011 (UTC)

MHP

Hi - I've drafted "combined solution" sections previously, see e.g. [1] (in the show/hide section). Several times I've done this Martin hasn't seemed to truly object. Is this more along the lines of what you're thinking? -- Rick Block (talk) 04:07, 31 May 2011 (UTC)

That draft is not what I have in mind. It does not explain the solution methods, the language is awkward, and the presentation is confusing. It's trying to say too much too fast. Glrx (talk) 16:42, 31 May 2011 (UTC)

Not even close enough to be a starting point for collaborative editing? If not, then perhaps I'm misreading your response here. In any event, I encourage you to keep working on the article in a WP:BRD sort of way. IMO, one of the main problems has been folks have been too timid about making changes and only suggesting changes (not even specific changes) on the talk page (and then becoming discouraged when others have not unanimously supported their proposed changes). As long as most changes are not simply reverted, actually changing the content and allowing others to successively modify your changes (iteratively) is a much more productive approach. -- Rick Block (talk) 18:39, 31 May 2011 (UTC)

I'm skeptical of it as a starting point. I don't think you're misreading my main thrust. On the contrary, you appear to understand my position better than most.

I agree with your too timid comment, but there's good reason for people to be timid: there's a tremendous amount of material on the talk page. Each side of the dispute also extends a lot of deference to the other side; it's a detente.

I intend to do some slow edits. That means I will move stuff around without patching it. The patches would create a string of edits. I will offer reasons for the changes, but you've already called me on a move. I see how that turns out on the talk pages.

Glrx (talk) 19:17, 31 May 2011 (UTC)

Compressed Air Batteries

Hello Glrx,

Please read carefully our answers to each of your points and let us know as soon as possible how we can resolve the problems you have raised.

This article was proposed for deletion by Glrx (talk • contribs) on 2011-06-04 with the comment: Article is about a Pnu Power product and sounds in advertising. The sources quote employees and company press releases without their own evaluation, so the sources are neither independent nor secondary. The compressed air battery is not notable. The compressed air battery is really a component of the company's uninterruptible power supply. The product is about a small variation on the design of a large scale UPS: instead of storing energy in a flywheel or standby battery, the energy is stored as compressed air. It is not clear that the compressed air battery UPS merits coverage in the UPS article at this time: WP:UNDUE. The main editors, User:Johnhutch and User:Goodwinaj, have ties to Pnu Power and therefore have a WP:COI.

1. Article is about a Pnu Power product and sounds in advertising

The article is about Compressed Air Batteries - their principles and application. Pnu Power is the inventor and only manufacturer of Compressed Air Batteries, so mention of the company is unavoidable and valid (cf Wiki entry for the iPad or Dyson (company)). It is written in an informative and neutral style.

2. The sources quote employees and company press releases without their own evaluation, so the sources are neither independent nor secondary.

The information supplied is factual and open to independent questioning or challenge. We make no apology that the authors (John Hutchinson and Andrew Goodwin) are employed by the company - just as, no doubt, the Wiki entries for the iPad or [[Dyson (company)] were authored by company employees or associates.

3. The compressed air battery is not notable.

The evidence suggests otherwise. It is very much notable as a new technology for generating electricity, as a direct replacement for conventional batteries, with lower carbon emissions and maintenance requirements. The fact that it is a new technology is proven by the granting of international patents. Its effectiveness is proved with its adoption by companies including National Grid in the USA and UK. The use of a scroll compressor/generator to generate electricity is an entirely new application of scroll technology. This may be compared with Sir James Dyson's application of industrial vortex filtration technology to a new type of product i.e. the bagless domestic vacuum cleaner.

4. The compressed air battery is really a component of the company's uninterruptible power supply

This is rather like saying 'a car battery is really (only) a component that starts the engine and runs the electrics', as if that was unimportant. Whichever type of battery is used in a UPS system, it is an essential component. Compressed Air Batteries are indeed used as components in large UPS system. They are also used as standalone UPS/backup power sources in smaller applications, as a direct replacement for conventional batteries. It is worth noting that this year's Japanese nuclear disaster resulted from a failure of battery backup systems.

5. The product is about a small variation on the design of a large scale UPS: instead of storing energy in a flywheel or standby battery, the energy is stored as compressed air

We fear there is some misunderstanding here. Please read all of the references that are included with the article. A Compressed Air Battery is a major variation in the design of a large scale UPS. Not least, Compressed Air Batteries have no moving parts in standby mode, whereas flywheels must be kept spinning and consume significant amounts of power in standby. The potential for Compressed Air Batteries to reduce carbon emissions in this application alone makes them notable.

6. It is not clear that the compressed air battery UPS merits coverage in the UPS article at this time: WP:UNDUE

We disagree. Compressed Air Batteries as part of UPS systems are a new technology which are generating a great deal of interest from engineers and specifiers internationally. Pnu Power has just received an order for a large UPS system for a data center operated by one of the UK's largest financial services companies, seeking (among other factors) to reduce its carbon emissions. We should be in a position to name the company within a few weeks.

7. The main editors, User:Johnhutch and User:Goodwinaj, have ties to Pnu Power and therefore have a WP:COI

If this condition was applied universally, many Wiki entries would have to be deleted. At this stage of the development of Compressed Air Batteries, the main source of information about them unavoidably resides within the company that developed them. We have simply provided a factual account of how Compressed Air Batteries work, how they have been developed and their potential applications.

In conclusion, we strongly believe that Compressed Air Batteries represent an important new technology and are worthy of inclusion in Wikipedia. We would welcome you suggestions for further editing.

Yours sincerely,

John Hutchinson Johnhutch (talk) 12:34, 6 June 2011 (UTC)

— Preceding unsigned comment added by Johnhutch (talk • contribs) 09:59, 6 June 2011 (UTC)

Perfect Sort / Triangulation Sort

Is a kind of un-algorithm. It simply triangulates from what data is so far sorted (discouraged / takes extra time). But implemented (see code) it shows shows some surprising benefits. Is sometimes referred to being "a sort in which data becomes available for streaming rapidly" (and formidable multi-field). Some books refer to such a method but I haven't seen it one of them code it or compare it's benefits. Interesting for just that. Enjoy. User:Sven nestle2/Perfect Sort / Triangulating Sort

You need to cite WP:RS for this algorithm. Who has evaluated this algorithm? Also watch out for WP:NOR. Glrx (talk) 21:40, 6 June 2011 (UTC)

Excuse me Glrx did you have time to see my corrections yet?

*** I've completely rewritten please look ***

I am trying to provide a good topic. Please say if there is any reason to hold it back. (for example: did I not explain triangulation?)

I'm sure the claims I made are correct. I'm sure the theory is old (ie triangulation) yet simply collated and furnished. As to its "value compared to other sorts" maybe you'd like to research it :)

Note I really don't mind your input in the matter I'd rather enjoy it.

Wikipedia:Identifying reliable sources. "~ the reliability of a source depends on context. Each source must be carefully weighed to judge ~" I don't see many of the sorts you list req. to show source. Wikipedia:No original research. "~ that includes any analysis or synthesis of published material that serves to advance a position not advanced by the sources ~". I'm sure I didn't make any unfounded conclusions nor say much or any theory. Triangulation is old hat I wouldn't say "origional". Yet I believe the properties (rapid output easy debug high speed) are interesting to see (as well as a rapid recursion removal lesson). Yet I do see allot of sorts posted make claims about handling things that infact most sorts handle if tweaked to do so.

By the way the sort was incl. in a copyright early 90's but as I said it's GPL'ed now. (called "radix_nr" but I won't go into the theories why since maybe I shouldn't :)

But that's parsing hairs. The original post was bad and I'm not sure if you've looked at the new post.

Thank you, John H. — Preceding unsigned comment added by Sven nestle2 (talk • contribs) 16:35, 12 June 2011 (UTC)

The above is essentially what is on talk:sorting algorithm. The discussion should happen there. Glrx (talk) 17:05, 12 June 2011 (UTC)

ok thanks I'm new to wiki arbitration (talk • contribs) 17:23, 12 June 2011 (UTC)

Streisand effect

As a result, the Revere article has become much longer, and much better sourced -– a version of what Wikipedia users call the “Streisand Effect,” which is described as when “an attempt to hide or remove a piece of information has the unintended consequence of publicizing the information more widely.” [2]. CallawayRox (talk) 18:28, 13 June 2011 (UTC)

But Sarah Palin has not been trying to hide or remove her Paul Revere comments, so it cannot be an instance of suppression leading to wider publication. -- Glrx (talk) 18:44, 13 June 2011 (UTC)

Front-side bus

Hello. Not sure if you are familiar with the policies of Wikipedia:Silence and consensus and Wikipedia:Don't template the regulars. It is sad to see you suggest dispute resolution, instead of participating in the discussion. Hope you reconsider. Generally making civil arguments based on evidence is more helpful to collaboration than acusations and threats. W Nowicki (talk) 18:34, 15 June 2011 (UTC)

Where have I suggested dispute resolution on this matter? I explained my edits in the summary. You want to add material; the material has been challenged; it was not reinserted by another editor; if an editor wants to insert challenged material, then he opens a discussion on the talk page. You've now opened a discussion. Glrx (talk) 18:53, 15 June 2011 (UTC)

Your complaint on my talk page included:

You can post a request for help at an appropriate noticeboard or seek dispute resolution. [3]

The issue had already been discussed on the talk page and nobody objected to my proposed compromise. I mentioned the talk page discussion in my edit summaries. W Nowicki (talk) 16:58, 16 June 2011 (UTC)

"Van is capitalized when it starts a name"

Not Vincent van Gogh. See first line of Japonaiserie (Van Gogh). William Avery (talk) 22:47, 19 June 2011 (UTC)

If "van" starts the name, then it is capitalized: "Van Gogh". If the "van" is medial, then it is not capitalized: "Vincent van Gogh". If you look in the article, you will see plenty of "Vincent van Gogh" and "Van Gogh". Glrx (talk) 22:55, 19 June 2011 (UTC)

Jensen's Device

Hello, I saw that you have reverted one of my edits. Would you mind explaining a bit more why? Although having examples in several computer languages would probably not be appropriate for the Wikipedia article, I think a link to the rosettacode is precisely what the reader would need to grasp the concept thanks to a version in his/her language of choice. I found it very relevant, especially given that the Wikipedia article is not that long and detailed. Note that I do not have personal interests in rosettacode.org. Alkarex (talk) 16:14, 20 June 2011 (UTC)

That was a long time ago. Linking to other wiki's is not usually encouraged; wikis are not WP:RS. You agree that WP should not be a code repository. It also is WP:NOTHOWTO. Those are basic problems with rosettacode. Generally, WP wants external links to have encyclopedic content that the article does not yet have (yes, there are other reasons to include external links). In addition to all of that, most of the rosettacode code examples do not use Jensen's device; instead they simulate the device in languages by passing non-lexically scoped functions. In many ways, that butchers Jensen. Glrx (talk) 17:47, 20 June 2011 (UTC)

IGES

Hello, Glrx … I noticed your recent edits to IGES regarding ISO 10303 (STEP) … I was involved with the IGES Project for over a decade, and was responsible for STEP's use of Wirth syntax notation instead of Backus–Naur Form.

Anywho, I'm curious as to your experience in CAx interoperability such that you would know that IGES and STEP should be linked, and how you came to the article on IGES (this time) in the first place?

Happy Editing! — 70.21.17.51 (talk · contribs) _{18:54, 21 June 2011 (UTC)}

I've been aware of IGES for a long time. My experience with CAx interoperability has been limited but frustrating. There have been many articles about STEP. Glrx (talk) 20:28, 22 June 2011 (UTC)

Re: Flip-flop (electronics) et all

Hi Glrx, I flip-flopped the title of that page to perform a history merge, so all the edits to the article are in the right place. Graham87 04:13, 24 June 2011 (UTC)

Thanks for the note and your work. Glrx (talk) 04:21, 24 June 2011 (UTC)

FPGA

I am a FPGA design engineer & I added a reference to probably the only FPGA Conference out there (FPGA Camp - http://www.fpgacentral.com/fpgacamp ), would you please let me know why my post was removed. If a link to the most respectable conference on FPGA is not deserved on the wiki page, I wonder what is? — Preceding unsigned comment added by 75.36.138.157 (talk) 06:32, 24 June 2011 (UTC)

(re [4] by User talk:8.4.225.30) The link sounds more in WP:ADVERTising a technical conference rather than supplying information suitable for an encyclopedia. WP is WP:NOTHOWTO do something; it's not a technical journal; it's not a textbook; it's not a place to advertise confernces. An external link usually supplies information that hasn't made it into the main article yet. See WP:ELNO. Glrx (talk) 14:49, 24 June 2011 (UTC)

Josh's edits

«Moving various "signaling" to signalling pages is not welcome. Please revert your edits.»

I'd like... Have American users got abused? I'd like to simply close the machine and forget about it. I know that tomorrow, I'll wish to occur here again...

I'm going to create twins then. It'll take some time tonight or even tomorrow as well. I'd like to preserve the ability for thousands of English learners in Europe, in Russia, I don't know where else.

Yes, I've now read the current convention. It's hard to me to cut my own fingers. But I have no skills to deny. Only forget. But the latter is inappropriate.

Or, do you think I vandalised the American English. I only want to diverge them. I don't want the British English to die the way of utter merging. Please wait. I'm thinking.. Lincoln Josh (talk) 19:03, 25 June 2011 (UTC)

Execution of Osama Bin Laden

I speedy deleted it. Bearian (talk) 23:10, 29 June 2011 (UTC)

Thanks. Glrx (talk) 00:43, 30 June 2011 (UTC)

Deprodding of Osama bin Laden assassination

I have removed the {{proposed deletion/dated}} tag from Osama bin Laden assassination, which you proposed for deletion. I'm leaving this message here to notify you about it. Instead, I have started a deletion discussion at [[Redirects can't be Proded. If you think the redirect is inappropriate, than take it to RFD. (If you do so however, than please don't notify me.)]], which you may comment on. I have explained my reasons for doing so there. Thanks!Armbrust ^{Talk to me} _Contribs 15:44, 30 June 2011 (UTC)

By the way the same applies to the Assassination of Osama bin Laden redirect too. Armbrust ^{Talk to me} _Contribs 15:48, 30 June 2011 (UTC)

Edit war

You currently appear to be engaged in an edit war according to the reverts you have made on Pulse oximeter. Users are expected to collaborate with others and avoid editing disruptively.

In particular, the three-revert rule states that:

1. Making more than three reversions on a single page within a 24-hour period is almost always grounds for an immediate block.
2. Do not edit war even if you believe you are right.

If you find yourself in an editing dispute, use the article's talk page to discuss controversial changes; work towards a version that represents consensus among editors. You can post a request for help at an appropriate noticeboard or seek dispute resolution. In some cases it may be appropriate to request temporary page protection. If you continue to edit war, you may be blocked from editing without further notice. That Ole Cheesy Dude ^{(Talk to the hand!)} 22:32, 19 July 2011 (UTC)

Nomination of CTERA Networks for deletion

A discussion is taking place as to whether the article CTERA Networks is suitable for inclusion in Wikipedia according to Wikipedia's policies and guidelines or whether it should be deleted.

The article will be discussed at Wikipedia:Articles for deletion/CTERA Networks (2nd nomination) until a consensus is reached, and anyone is welcome to contribute to the discussion. The nomination will explain the policies and guidelines which are of concern. The discussion focuses on good quality evidence, and our policies and guidelines.

Users may edit the article during the discussion, including to improve the article to address concerns raised in the discussion. However, do not remove the article-for-deletion template from the top of the article.

B-tree reversion

Was there something incorrect in my change to the B-tree article? RMcPhillip (talk) 01:50, 3 August 2011 (UTC)

At least two things were wrong with your B-tree edit. First, your change was wrong; when the root node is full, it is split and a new root node is inserted above it -- which forces all leaves down one level (but makes room in the old root in case the level below splits). Same thing with other splits on the way down. Continued inserts on the right make a tree that has half full nodes except for the right edge. It is not deep on the right. Second, material in article space needs to be reliably sourced; your thinking/views/concerns on the matter is original research, does not have a reliable source, and should not be put in the article. You should raise doubts on the talk page. Glrx (talk) 03:27, 3 August 2011 (UTC)

You are correct on both accounts. I reviewed my source (Jim Gray's "Transaction Processing") to buttress my argument and to avoid "original research". And, darn it, Gray agreed with you! No insertion operation changes the height of the tree except root splitting, which increases it by one for all nodes.

And I overlooked using the talk page, which perhaps is a typical newby error. Thank you for your patience. RMcPhillip (talk) 10:49, 3 August 2011 (UTC)

Edit to link on Box Plot page

From your obvious commmitement and dedication to preserving the integrity of Wikipedia content, just wanted to let you know that: 1) I completely respect your decision to remove my suggested link and 2) Let you know that it was never my intention to pollute the page with commercial content. Instead, I was just trying to make an honest contribution to education as I'm passionate about finding ways to make math more interactive. Thanks for your hard work! Danharrisdanharris (talk) 16:02, 5 August 2011 (UTC)Danharrisdanharris

First, thanks for the comment. Second, I don't think your intentions were bad.

I'm not a final authority on anything around here, and if you still think the material is appropriate for WP, you can bring it up on the article's talk page. In the future, if you have a WP:COI, it may be better to propose an addition on the talk page to see if other people agree with it.

Generally, ELs should have information that belongs in the article but has not been included yet or is too detailed to include. Calculators and applets are difficult territory. WP articles cannot have them, so they must be linked. They often do not add significant material, and they may run counter to WP:NOTHOWTO policy.

Glrx (talk) 16:23, 5 August 2011 (UTC)

Thanks for the insight and suggestions to a newbie. I have tried to understand and follow all the published protocols and guidelines, but nothing beats good advice from a real person. Your recommendations make a lot of sense. I too struggled a bit deciding whether applets are consistent with the WP:NOTHOWTO policy. In the end I said to myself "if static pictures of singular examples of box plots are valuable and extensively used on the main page, then perhaps an interactive manipulative that displayed a box plot for any combination of values might be appropriate as an external link". Thanks again for your time and help. Danharrisdanharris (talk) 16:53, 5 August 2011 (UTC)Danharrisdanharris

MedCab

I have made comment at the above case and would once again appreciate your input. Steven Zhang ^{The clock is ticking....} 20:56, 8 August 2011 (UTC)

Question regarding your VoiceXML edits

Hi, thanks for correcting me regarding addition of the vxml entry item. What confused me is the presence of other similar items in this section. Is there a legitimate way of putting in a reference to a product offered by a private entity? Dpllpd (talk) 16:13, 23 August 2011 (UTC)

Wikipedia is not interested in advertising company products, so the simple answer is, "No, product offerings are not legitimate". WP is an encyclopedia, and external links should contribute information that advances the goals of the encyclopedia. See WP:ELNO. In most circumstances, linking to a private product is inappropriate because the link is about selling the product rather than providing reasonable information content. Some companies have detailed descriptions of what their product does, how it works, or historical information. If that information is not in the WP article, then linking to such a description can be appropriate.

Sometimes some inappropriate links are added to an article and nobody objects. It may not be a big deal. But then competitors start adding their links, and pretty soon a single link has grown into a long list of suppliers that does not serve the purposes of WP. My impression is the VoiceXML EL list has been trimmed back a few times. User 173.48.119.249 added the iec link, I deleted it, you added it back, then Mindmatrix deleted it and threw out some other inappropriate links.

Glrx (talk) 16:37, 23 August 2011 (UTC)

Articles you might like to edit, from SuggestBot

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

Stubs Pairwise sorting network Fritzing Peter Stockwell Odd–even sort Compensation (engineering) Self-resonant frequency American flag sort Norm (abelian group) Grid dip oscillator Anode ray Psi-Theory Thomas H. Cormen Frequency drift Grid bias Long-term stability Burstsort Mount Kaukau Ceramic resonator Batcher odd–even mergesort

Cleanup Osama bin Laden List of common misconceptions Relaxation oscillator Merge Best, worst and average case Semacode Reactions to the death of Osama bin Laden Add Sources United States Naval Special Warfare Development Group Audio filter Dominique Strauss-Kahn Wikify PEO Soldier Unibind Sip-and-puff Expand Array data structure RL circuit Algorithm

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 Nettrom (talk), 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 (talk) 20:06, 5 September 2011 (UTC)

Thousand pair cable

This is not dubious at all. Large twisted-pair cables are very common in telecommuncations, or at least they were before the advent of fibre-optic backbones. You won't see them in your home or small business, but in central offices (exchanges) and the frame rooms of large buildings/organisations they are ubiquitous. Also, I don't think that you should be asking for citations in sections that are merely a summary of a main article. The references should already be in the article. SpinningSpark 18:01, 7 October 2011 (UTC)

Awa Maru

My moving the article was in accordance with WP:NC-S. Steamships and motor vessels are prefixed, whereas merchant sailing vessels are not. If you still disagree, please start a discussion at WT:SHIPS. Mjroots (talk) 11:35, 11 October 2011 (UTC)

Deletion of my section on AC97 and Intel HD front panel auto sensing

"Header issues more involved" - was the note left by you. Sir you have removed an important piece of information that can be of great help to people trying to figure out how to make auto detection of front panel Headphone work across two different kinds of front panel connectors. I had posted in brief hoping that some one would take it as a seed and build up around it. Please elaborate your stand on this and also please advice me how should i go about it. My understanding is: add what you know and others will look at it and add what they know; and together we will have written a comprehensive text.

Your edits to AC'97 and Intel High Definition Audio made little sense:

The difference between AC'97 and Intel HD Front Panel connections is mainly in that: AC'97 uses 'normally closed' switch for return/sense whereas Intel HD uses 'normally open' switch for the Headphone port. So it is the jack and not the wiring that has to be altered to get auto sense working across the two specifications.

HD does not have a goal of compatibility.

The AC'97 mobo header uses analog loopback; the HD mobo header uses analog out with digital back sensing; in theory an HD implementation could work with an AC'97 case, but even Intel screwed that part up by not gating the analog return lines.

See also WP:NOTHOWTO. WP is not about how to fix stuff.

Glrx (talk) 22:07, 24 October 2011 (UTC)

I have verified both theoretically and practically. I have a Creative X-Fi which has Intel HD 'conformal' FrontPanel Header. I had to manually alter normally closed connection in the female audio front panel jack to normally open: and now auto sensing works as intended.

I suggest/request you to add a little bit to what i had written and|or tag it for further improvement.

User:San 013 (talk) 20:55, 29 October 2011sig added

Do you have any reliable sources that back up your claim? Wikipedia does not want to publish an editor's original research. - Glrx (talk) 01:51, 30 October 2011 (UTC)

.see Intel, Front Panel I/O Connectivity Design Guide, version 1.3, Feb 2005, A29286-005, pp 19–25. It describes both AC'97 and HDA front panel connector pin assignments. The pin assignments are such that an intelligent motherboard will sense pin 4 of the connector to figure out the front panel type. If pin 4 is audio ground, then the front panel is AC'97; if pin 4 can be pulled high, then the front panel is HDA. (See 2.3.5.1.1.)

In AC'97, pins 6 and 10 are front panel audio return lines (loopback); pin 7 is 5V power. In HDA, pins 7 is SENSE_SEND; pins 6 and 10 are digital jack SENSEi_RETURN lines that are shorted to pin 7 when plug inserted into jack.

The classic problem is an HDA mobo doesn't do the sense right or just assumes it is connected to an HDA dongle. Consequently, audio lines are connected to SENSEi_RETURN lines. When a loud audio signal comes along, the mobo freaks out thinking the jacks are being plugged/unplugged hundreds of times per second.

Glrx (talk) 19:07, 31 October 2011 (UTC)

thankYou for providing greater insight. I had looked up the Intel Standard pdf as suggested by a writer in this page: http://www.wirelessforums.org/alt-comp-hardware/soundblaster-x-fi-xtreme-audio-pci-express-card-57089.html - Find(Ctrl+F) "intel.com" in the page (the pdf link at the page is broken). In the Intel HD Audio standard document at: http://www.intel.com/content/dam/doc/product-specification/high-definition-audio-specification.pdf page no.: 220 the following is written: "Each jack must have an isolated switch (normally open), as shown in Figure 97, which closes when a plug is inserted into that jack"

please suggest the next course of action. — Preceding unsigned comment added by San 013 (talk • contribs) 21:36, 31 October 2011 (UTC)

I added stub sections to both AC'97 and iHDA articles that point to the Intel reference. Glrx (talk) 18:10, 4 November 2011 (UTC)

Binary search algorithm

Hello.

I've re-posted the url in the external links section. Consider this...

In my opinion: The primary purpose of any technical documentation is to disambiguate its subject matter. The url in question attempts to do just that. I would consider it an introduction using plain-spoken language that provides examples for those whom seek an entry level discussion on the topic. I submit that it be left in place for that reason. Thanks for reading! Msswp (talk) 23:44, 20 November 2011 (UTC)

The link is an apparent blog, so it is not a reliable source to WP. It also appears that you have a WP:COI with the material. Editors should be very circumspect about adding coi material. Other editors have also removed the link, so you need to get a consensus on the article talk page to re add the link. Glrx (talk) 06:28, 21 November 2011 (UTC)

Actually, its not a blog, it is my website however. How have you determined its not reliable? Quite the contrary. Also, in both removals, not once has any one stated that the subject matter is less than applicable. In all earnestness, has anyone even read the paper in question before removing its link? Msswp (talk) 09:49, 21 November 2011 (UTC)

In general, self-published material is not considered reliable. Papers published in journals usually have undergone a peer review. An editor wishing to add material to Wikipedia has the burden of showing that the material is appropriate; it is not the other way around: I don't have to show the material is inappropriate. I deleted the link because I believe it is inappropriate. If I didn't read your webpage, then why did I think that you wrote it? I trust the other editor has also read the page. Glrx (talk) 16:57, 21 November 2011 (UTC)

By all means, kindly direct me to a link that I may read by what protocol you deem the paper inappropriate. Also, do I have the right to appeal your decision to a higher authority? One thing I'd like to bring up is that the wp article is well... not so approachable with respect to a layman understanding its content; hence the reason I seek to add another vantage point. 2ndly, if you believe the material is not a good fit for the page, why have you not enumerated the reasons you feel this is so? See... here's the thing: if you disallow this addition and then simply invoke: because I say so its the same situation that compelled me to add my paper to begin with. Msswp (talk) 19:14, 21 November 2011 (UTC)

Please see WP:RS. I have no authority around here -- they don't even let me turn out the lights. To gain a WP:Consensus or to discuss an aspect of an article, you just start a new section on the article's talk page where you explain why the material is appropriate for the article. Other editors may then express their views. If a consensus to add is reached, then the material may go in. See also WP:BRD. You were Bold and added the link; you've been Reverted; now it is on you to Discuss. Glrx (talk) 19:43, 21 November 2011 (UTC)

Thanks Glrx, appreciate the heads up. I just wanted to add an additional resource, but its not meant to be. I'll let it drop. Msswp (talk) 19:57, 21 November 2011 (UTC)

Just browsing WP:BRD and I read; Expect strong resistance—even hostility... resistance I could imagine but hostility? Darn folks - maybe some fresh air is in order here... Back to reading for me. Msswp (talk) 13:41, 22 November 2011 (UTC)

Requesting clarification

I'd greatly appreciate understanding your thought process for this edit. Are the texts accessible to you? Why the COI allegation? AshLin (talk) 19:08, 13 December 2011 (UTC)

This morning I came across an edit to East China Sea by User:Timmonsgray. The edit added a further reading section, but the addition was about energy and politics. The ECS article has some discussion of energy resources and related border disputes, but they are on a tangent about EEZ rather than ECS geography. The editor did not add content to the article.

I looked at some other edits by that user, and they were just adding NBR links (often to newly created further reading sections) and not editing the article content. I formed the belief that the editor's purpose is to promote NBR's publications rather than Wikipedia's interests.

From some other research, I formed the opinion that the editor is the Assistant Director of Online Outreach for NBR. That's the suspected WP:COI.

I then reverted several edits -- including the one you cite above. The first title may have some information about global warming (an "ecological" challenge), but it also includes "security challenge". The other two titles suggest they are about security challenges specific to Pakistan (abstract covers population growth, food production, and failure to manage water resources) and India. Even a non-traditional security challenge doesn't lead me directly to global warming. The Pakistan link has a short discussion about global warming and its impact on Pakistan (page 5). The India link is about economic schism, personal security, and corruption; it has a short section on climate change at pages 14-15; melting glaciers cause flash floods; more storms; rising sea level.

Generally, I do not like the idea of a further reading section. If a work has something important for the article, then there should be a statement in the article that uses the work as a reference.

Note that the article already has abundant references to support its statements. I doubt the further reading links add significant details not covered by the references.

I have no problem with NBR publications being used as references. I do suspect, however, that much better sources are available.

I have a problem with an NBR officer salting articles with NBR links.

Glrx (talk) 20:39, 13 December 2011 (UTC)

Thank you for your response. I understand now. AshLin (talk) 03:11, 14 December 2011 (UTC)

Your poor behavior; NOTHOWTO doesn't apply to links

First, I don't really care enough about linking people to something I found helpful to keep pushing this. But I must say that first you just undo something without comment, which is not according to policy; then you state that Wikipedia is not a how-to. Now, I've been around here long enough to know that that is absolutely true, Wikipedia is not a how to. However it is not Wikipedia policy to not link to a how-to, so that policy does not apply in this case.

Anyway, you apparently care more than I do -- so whatever, but I just want you to be aware that your conduct here hasn't been exactly stellar (and looking at your talk page you have a bit of reputation as a wikipedia bully with too much time on their hands) -- you might want to follow the rules a bit closer yourself.

--Quasipalm (talk) 20:26, 3 January 2012 (UTC)

Quasipalm's comment is about edits to Merge sort and an external link to codecodex.com that has merge sort implementations in many different computer languages. Generally, WP only wants one implementation of an algorithm, and it prefers that implementation to be in psuedocode.

Normally, I would not respond to Quasipalm, but Quasipalm has also left a message for User:X7q. Here is the background.

Quasipalm's edit history shows few edits to sorting algorithm articles. Those edits are primarily about the external links to codecodex.com.

• On August 12, 2008, QP edited the codecodex.com EL at Bubble sort (changing 20 to 28 languages).[5]
• Apparently, somebody removed the link because on January 27, 2009, Quasipalm re-added the codecodex.com external link to Bubble sort.[6]
• On June 30, 2009, Quasipalm edited the Bubble sort codecodex link.
• Apparently, the Bubble sort codecodex.com EL was again removed.
• On November 5, 2010, Quasipalm added codecodex.com external links to Bubble sort[7] and Heapsort[8]; in addition, he edited the codecodex.com link at Merge sort.[9]

• I removed some external links to polyglot implementations back in June 2011 with the comment "del WP:HOWTO code respositories".[10]

Present conflict

• On Dec 30, 2011, Quasipalm added (without any edit comment at all) the external link codecodex.com to Merge sort.[11]
• I reverted the link with the comment "(Undid revision 468650854 by Quasipalm (talk) article would not include code repository)".[12]
• Quasipalm, without any edit comment, reinserted the link; in addition, Quasipalm sorted the interwikis (dropping de:interwiki in process).[13]
• X7q reverted with undo (i.e., no edit comment).[14]

Comment. X7q may not have understood the sorting of the interwikis, but he could have viewed it as some subterfuge to slip in the codecodex.com link. Given that the de: link was lost, the revert was appropriate. He further explained in his revert below.

• Quasipalm reverted X7q with undo (i.e., no edit comment).[15]
• X7q reverted with the comment "(Undid revision 469246803 by Quasipalm (talk) 1) don't mess with interwiki; 2) wikibooks has enough implementations too, stop spamming your site!)".[16]

Comment. Given Quasipalm's codecodex.com edit history, X7q apparently presumed that Quasipalm has some ties to codecodex.com.

• Quasipalm reverted X7q with undo (i.e., no edit comment).[17]

Despite Quasipalm's concerted effort to quietly introduce a link two editors had removed (and others had removed in other WP sorting articles), Quasipalm accused X7q of "Undo Abuse".[18]. Despite X7q's edit comment on the second revert stating why he reverted X7q, the accusation was about an undo without comment. Quasipalm's recent insertion and undo efforts had been without any comments.

• I removed Quasipalm's codecodex.com external link from merge sort with the comment "remove code repository; ]]WP:NOTHOWTO]] to program; encyc content same as current article",[19] left the sorted interwikis, but restored the de: interwiki with the comment "restore German interwiki deleted by Quasipalm".[20] The text content at codecodex.com is essentially the same as the WP article content. I left QP's efforts at sorting the interwiki links because those links are usually sorted. I did spend some effort to make sure that the edit was correct, and that is how I discovered that QP had deleted the German interwiki.
• I issued a WP:3RR warning to Quasipalm because QP was continually reintroducing the codecodex link despite removal by two other editors.[21] Quasipalm had made 3 reverts within 9 hours. Quasipalm has been around for some time, so I did not template a regular.

Quasipalm's response is the above missive.

Quasipalm's double standard is regrettable. It is OK for QP to hit undo without further comment (or even to add material without an edit comment), but it is not OK for X7q to just hit undo when an editor with a history makes an unannounced entry of a previously removed EL. Criticizing X7q for using undo without a comment when X7q had made his reasons clear is also inappropriate.

QP criticizing me for "undoing something without comment" is baffling. The only merge sort edit I made without a comment was this addition of two spaces] that was buried in a block of 4 edits (the other 3 carry comments) that immediately followed my commented undo of Quasipalm's commentless insertion.

WP is not a textbook or a tutorial about programming in many languages. If merge sort were a good article, it would not include multiple implementations of an algorithm.

My practice is to remove links to algorithm implementations when those links do not include additional encyclopedic material. The current article describes the merge sort algorithm adequately. It does not need versions in different computer languages.

Glrx (talk) 19:43, 5 January 2012 (UTC)

merge sort

This section is about the wiki article Merge_sort. I'm hoping you're the correct person to discuss this with. If not, could you suggest someone else?

A more generic description of a merge sort could be:

1. Divide the unsorted list into n sublists, each containting 1 object.
2. Repeatedly Merge sublists to produce new sublists until there is only 1 sublist remaining. (This will be the sorted list.)

Note that the top down version of merge sort recursively divides sub-lists until sub-list size = 1, before any merging takes place, so this decription would cover both top down and bottom up merge sort algorithms.

Currently there's a conflict between Sorting_algorithm#Merge_sort which describes a bottom up merge sort and Merge_sort which describes and animates a top down merge sort.

It's also not clear to me what is the advantage of using recursion to generate sets of nested indices or pointers to sub-lists to divide them down to size = 1 object versus simply starting off with n sub-lists of size = 1 object and using a small fixed set of indices or pointers and iteration to implement the merge. I assume that Knuth and Von Neumann described a bottom up merge sort, so again I don't understand why a top down merge sort is used as the primary algorithm at Merge_sort.

Sometimes a merge sort will use another sort algorithm to sort relatively small sub-lists:

1. Divide the unsorted list into into some number of relatively small sub-lists and sort them using some sorting algorithm.
2. Repeatedly Merge those sublists to produce new sublists until there is only 1 sublist remaining. (This will be the sorted list.)

I'm not sure if or where this should be mentioned in Merge_sort.

Rcgldr (talk) 12:05, 9 January 2012 (UTC)

I poked at it a little to give a bottom up version, but there are many problems.

Alg can be divide and conquer or inductive build. Describing both ways may be useful.

The algorithm info box doesn't hit the O(n) aux storage. The list version is deceptive about aux storage and how lists are managed. There was some confusing array stuff, too.

Anyway, the article discussion should be on the talk page -- not here.

Don't be shy about editing the article.

Glrx (talk) 20:51, 9 January 2012 (UTC)

OK, I'll go back to posting in the dicussion page. Mostly I was wondering why so many web sites describe merge sort as a top down process, when the original meaning (von Neumann, Knuth) was bottom up, and wondering when and why this change was made and who decided to make this change. Thanks for the response. Rcgldr (talk) 22:20, 9 January 2012 (UTC)

• made a minor fix to the bottom up pseudo code to handle the case where (i+width >= n), in which case A{i,n-1] is copied to B[i]. Rcgldr (talk) 23:12, 9 January 2012 (UTC)

I am officially an idiot. But hey, I have jury duty today. Glrx (talk) 23:18, 9 January 2012 (UTC)

Also changed the for loop so i is advanced by 2*width, i = i + 2 * width. I did like the way your bottomupmerge single if() statement handles bothy copying the remainder of sub-lists as well merging. Rcgldr (talk) 23:36, 9 January 2012 (UTC)

Originally, I started out with width = 2 and had a lot of width/2 ugliness, but then I realized width = 1 would clean stuff up (clear invariant comments) -- and I blew the increment. I also didn't have work array B, but it confused the notion of "in place". But I'm making excuses. Thanks for checking and fixing the code -- checking someone else's code is not fun. Glrx (talk) 00:05, 10 January 2012 (UTC)

Instead of copying B to A at the end of each inner loop, you could have BottomUpSort() initialize two pointers, PSrc = A, PDst = B, and then swap pointers (using a 3rd pointer PTmp for the swap) after each inner loop. After the outer loop, the function could return PSrc (which was just swapped with PDst), which would be the pointer to the sorted array (or it could check if PSrc == B and if so copy to A one time). It might ba a bit more complicated, but it would reduce the copy overhead in the example, and be more reflective of how a bottom up merge sort would be actually implemented. Or you could just add a boolean variable and an if statement to alternately call BottomUpMerge(A, ..., B) or BottomUpMerge(B, ..., A). Rcgldr (talk) 04:02, 10 January 2012 (UTC)

I was just going to put swap(A,B) at the end, but the copy is simpler to understand. A few days ago I replaced some clever code with code that was simpler to understand. Glrx (talk) 04:07, 10 January 2012 (UTC)

Perhaps just a comment before the copy stating that swapping pointers or a flag could be used to toggle bewteen BottomUpMerge(A, ..., B) or BottomUpMerge(B, ..., A) to avoid doing a copy on every outer loop. I was going to update my previous message to mention that a comment would probably be enough for your example pseudocode, but you already responded. Rcgldr (talk) 04:13, 10 January 2012 (UTC)

merge sort - natural merge sort - use with tape drives

I'm considering removing the pseudocode from the use with tape drives section, since the text added to the Merge_sort#Natural_merge_sort and Merge_sort#Use_with_tape_drives section explain the basic algorithm, and proper pseudocode for a natural merge sort is more complicated than what I would expect for a wiki article. Also I'm not sure how relevant it is to include a lot of text or psuedocode for a legacy tape based sort. Waiting for your opinion on this. Rcgldr (talk) 06:47, 13 January 2012 (UTC)

I need to think about this more.

Generally, if an article gets too long, then it may be time to split something off. A natural merge sort article may be appropriate. Expanding External sorting#external merge sort is another option. Pseudo code is appropriate because merge short has a run number issue (classic algorithm does not use an array of widths). WP:NOTPAPER

Proper focus is also an issue. For external sorting, comparisons go out the window; focus is I/O speed and passes. The article should discuss N-way.

It's not a legacy algorithm. External sorting is still important. Tapes may not be used, but disks are. A friend of mine is responsible for sorting terabytes every night on a huge mainframe -- and the machine is severely taxed.

Glrx (talk) 18:15, 13 January 2012 (UTC)

I already updated the tape section. The previous pseudocode in the tape section needed some work, and since you added pseudocode for bottom up merge which is similar to a tape sort, and with the added description for natural merge sort, I deleted the old pseudocode and added references to your pseudocode and the natural merge sort description, along with an explanation that filemarks can be used instead of variables or an array to track sublist (run) sizes. It also links to the external merge sort page, but that currently is oriented to disk sorts, so I referenced it as a disk sort. I'm not sure if it's worth creating a separate page for tape sorting. I could add psuedocode specific to tape sorting, but other than us, who would read it? Rcgldr (talk) 08:12, 14 January 2012 (UTC)

merge sort - natural merge sort

• classic algorithm does not use an array of widths - That's only used for a natural merge sort where sublist (run) sizes are variable. I was thinking that the pseudocode to handle all the boundary conditions involved would be uneccesarily lengthy versus the general description of how the sublist size array is used. Rcgldr (talk) 08:12, 14 January 2012 (UTC)

Thanks for cleaing that section up. I got carried away with the sublist size array and completely forgot to mention using compares to determing the end of sublists (runs). It was getting into uneeded details. I'm not sure what else should be added to the article, especialy since we seem to be the only two updating it at this time. Rcgldr (talk) 09:37, 21 January 2012 (UTC)

polyphase merge sort

I added a response to Talk:Polyphase_merge_sort. The main issue is the article comment about idle drives, since polyphase doesn't change the fact that each pass runs at the transfer rate of a single drive (assuming that reading and writing are concurrent). The clever part of polyphase is it's ability to run well on 3 drives. I don't know if the inspiration for the polyphase merge sort was due to someone actually watching idle tape drives in a standard 4 drive sort. I'm also wondering if there's a point of diminishing returns in the case of 4 drives (I explain this in my response). It also seems that polyphase isn't a stable sort, becuase the runs are not merged in orginal order. Rcgldr (talk) 08:15, 14 January 2012 (UTC)

I split up part of the talk page into tape and disk sections. I created a table to compare polyphase with standard merge on 4 drives, and unless I'm missing something, I don't see the benefit of polyphase in a 4 or more drive situation. Rcgldr (talk) 09:39, 21 January 2012 (UTC)

Sigh. PPMS phases aren't quite the same as MS passes. PPMS runs aren't quite the same as MS runs. Glrx (talk) 19:54, 21 January 2012 (UTC)

You're correct. I was missing this very fact. I didn't realize the PPMS phase size would be so much smaller than a MS pass. I wasn't sure, which is why I never touched the article, except to add some more 3 term fibonacci numbers. I updated the talk page tape section with another table that shows PPMS is better. Rcgldr (talk) 23:28, 21 January 2012 (UTC)

Are there any reliable sources for using PPMS versus MS versus ?? in modern day disk sorts? External_sorting specifically mentions using a 16 way merge for a disk sort (treating the drive(s) as 32 sequential devices) with large I/O sizes, and these class notes that imply MS is as fast (or better?) than PPMS with > 8 devices, esort.html : In general, polyphase merging is better than balanced n-way merging when there's a small (> 8) number of devices. Rcgldr (talk) 23:28, 21 January 2012 (UTC)

I would expect Knuth to address PPMS v. MS in terms of theoretical performance, but he's not handy. His treatment may be narrow in only addressing limited sequential access. For large N, I can see PPMS source data rates being unbalanced due to run length disparity, but such a statement needs a reliable source; there may also be more cleverness involved. I believe actual implementations of tape-base PPMS and MS are much different than the basic algorithm.

I don't know about modern day sources.

Thirty years ago, the typical sort/merge utility that I saw employed a large N-way merge. But those systems were also toys in many respects because they did not support large file sizes and had poor I/O channels.

I would not expect PPMS to be used in modern day disk sorts because it does not exploit random access R/W. I believe one database provider switched from a straight MS to a disk-based PPMS, but that doesn't mean much because they were also fighting poor I/O rates.

I don't see External sorting or its talk page discussing a 16-way merge; maybe I'm blind. There is mention of a pedestrian 9-way merge. I think there are some multi-threaded balanced I/O merge sorts that might be interesting, but I don't know of any sources. I read the class notes (which lack a lot of detail) a little differently; tailored sorting algorithms will have better performance. I would not consider those class notes a reliable source for the quoted statement; I don't have an opinion on the statement either way.

Glrx (talk) 00:08, 23 January 2012 (UTC)

My mistake, I mixed up the external sort article with what I read a few years ago. As you noticed, External sorting mentions a 9 way merge with large I/O, but still that's equivalent to 18 pseudo drives. One of the articles I had read a few years ago mentioned a 16-way merge (32 pseudo drives). I emailed R. Clayton from esort.html and he replied: The 8 or less claim comes from page 165 in Sedgewick, where it's stated without citation. Flicking quickly through section 5.4.2 in Knuth doesn't come up with anything definitive, but the graphs and tables suggests the law of diminishing returns kick in at around seven tape drives. Wiki has a web page for Sedgewick. Rcgldr (talk) 09:50, 23 January 2012 (UTC)

• Getting back to the Merge_sort article, there are references to disk drives in Merge_sort#Use_with_tape_drives. It seems a separate sub-section should be created for use with disk drives, or the section title could be change to use with disk or tape drives. As for PPMS, there doesn't seem to be any citable source for or against using PPMS on a disk sort on a modern system, and now that I understand PPMS better, I'm not sure either. Other than disk references in use with tape drives, I don't see any issues with the Merge_sort page, and I don't plan to make any changes. My main concern was the inclusion of the algorithm for bottom up merge, and you added a very concise psuedocode section. Rcgldr (talk) 09:50, 23 January 2012 (UTC)

• In general, polyphase merging is better than balanced n-way merging when there's a small (8) number of devices. - I created a program to grind this out. The initial condition assumes data already distributed on the drives, same as the article. 10 devices seems to be the transition point where which is better (PPMS or MS) depended if run count was just above a MS boundary point and just below or at a PPMS boundary point. At 12 or more devices, MS is better at any run count > 1000. Rcgldr (talk) 04:37, 22 January 2012 (UTC)

source files for example sort programs

A pair of links to example sort programs I wrote, along with benchmark results. Note the sortp.zip file is large (8.6 MB) because it includes an example text file to sort. These programs can be compiled with Visual C++ express (free), in either 32 bit or 64 bit mode.

array sort programs, using STL vector<> template. msort.cpp is my own merge sort. sort.cpp uses STL stable_sort() where visual c++ uses insertion sort to create sublists of size 32, then uses merge sort, but you can change stable_sort() to sort() where visual studio uses quicksort. hsort.cpp is my hybrid radix / merge sort, it's the fastest for random data, but it's using 1GB of ram to sort a 32MB array.

sortv.zip

text file sort program, these work by sorting pointers to records. msortp.cpp does both normal and natural merge sort. You need to change the define for VARSZGRP to #define VARSZGRP 1 to enable the natural merge. An array of sublist sizes is used in either case to keep most of the code common.

sortp.zip

Rcgldr (talk) 14:58, 15 January 2012 (UTC)

Disambiguation link notification

Hi. When you recently edited Sons of Guns, you added a link pointing to the disambiguation page Mortar (check to confirm | fix with Dab solver). Such links are almost always unintended, since a disambiguation page is merely a list of "Did you mean..." article titles. Read the FAQ • Join us at the DPL WikiProject.

It's OK to remove this message. Also, to stop receiving these messages, follow these opt-out instructions. Thanks, DPL bot (talk) 10:43, 16 January 2012 (UTC)

Fisher–Yates shuffle, "Algorithm P" C example.

Hi,

On the Fisher–Yates shuffle history page I notice that you wrote the following comment when you reverted the page to the version before the addition of my C language example:

"(rvt C implementation; have pseudo code imp already; rand() % index is biased without accept/reject)"

I was slightly confused by the bias point, as I had looked up and followed "Algorithm P" from Knuth's The Art of Computer Programming Vol 2. I decided to have another look today. By chance I looked in the 3rd edition and saw the description was slightly different. On investigation I realised that the version I had looked up originally was from the 2nd edition. The 2nd edition makes no mention that "division by j should not be used to determine k" (j being the index and k being the random number), that requirement was added in the 3rd edition.

In the past I've always used the % operator to get random numbers within the required range and had no idea that this was not good practice. If instead I had used the following line, would that resolve your bias concerns? The line is suggested in the FAQ of the comp.lang.c newsgroup.

int randomNum = (int) ((double) rand() / ((double) RAND_MAX + 1) * index);

The full amended code can be seen here: http://pastebin.com/MW7TUTJ4

I would also like to understand why you feel there's no need for a C example of "Algorithm P"? I was just trying to improve the page by adding it. Is it really superfluous rather than helpful?

Best wishes, Matthew Mattstan (talk) 16:44, 4 February 2012 (UTC)

re Fisher–Yates shuffle

Your "improved" expression for randomNum is at least as bad as the original modulo. Think of the mapping the output of rand() onto the desired index range; if the index range does not evenly divide the number of possible outputs of rand(), then there will be bias -- some values of randomNum will be slightly more prevalent (by one count) than others. Floating point may introduce other ugliness. Accept/reject is needed. That's also why WP wants verifiable sources and not an editor's WP:OR or belief. The article itself explains both the modulo and floating point problem at Fisher–Yates shuffle#Modulo bias.

WP articles about algorithms generally prefer a single implementation in pseudocode rather than an explicit language. I should have that link handy, but my mind fails me right now. Many editors want their favorite language represented, but that ends up being a nightmare. If everybody added their favorite language to an article, then the article would be clogged with multiple implementations that don't say anything new. WP is not a code repository. WP wants to convey the idea of the algorithm and need not get lost in details about declaring variables or the exact syntax required for a loop.

I do not follow random links, so I'm not looking at pastebin.com. The original C code[22] had a lot of superfluous details in it -- such as initializing the random number generator, initializing the array to be shuffled, an awful ARRAY_SIZE macro that cripples the code, and comments about including header libraries rather than actually including the libraries. Such detail is tangential to describing the algorithm.

That's my thinking.

Glrx (talk) 02:02, 5 February 2012 (UTC)

My RfA

Could you expand on your "There is a small concern that he wants perfect information" comment as I'm intrigued by it. Dpmuk (talk) 20:21, 13 February 2012 (UTC)

First, I wish you well on your RfA.

By reading your answers and going through some of your contributions, I get the sense that the level of proof you require for a statement might be very high. Consequently, you might stall on making a decision or reject marginal statements for lack of proof too easily. It's the downside of being too cautious. I'm not suggesting that you lower your standards; it's just something that crossed my mind. (I know you gave a pass to an admin for not having all the facts when he acted.)

Generally, I'm happy when someone claims that he's not qualified to answer a question. It shows that he's aware of what it takes to understand an issue. But I get leery if he starts invoking it too often. I'm not hot to trot on BLP issues either, but when I ran across one a couple weeks back, I looked at and tried to understand the the list, and then reverted the edit for no RS. If I didn't do it, then maybe nobody would. Begging off has a downside. If I don't start dealing with simple questions, then I'll never be qualified for more complicated ones.

Glrx (talk) 02:44, 14 February 2012 (UTC)

Fair enough. I take your point and thank you for expanding - it's useful to know. I'd like to think that when I beg off I do at least try to ensure it isn't forgotten. I suspect the high level of proof may well be an unintentional consequence of being an academic researcher where having to make a retraction can put a real blot on a career - and I'm not further enough along in my career to be likely to survive such a blot. I'll take on board what you say and try to bear it in mind when making marginal decisions. Dpmuk (talk) 02:50, 14 February 2012 (UTC)

Disambiguation link notification

Hi. When you recently edited Primer (firearm), you added links pointing to the disambiguation pages Muzzle and Notch (check to confirm | fix with Dab solver). Such links are almost always unintended, since a disambiguation page is merely a list of "Did you mean..." article titles. Read the FAQ • Join us at the DPL WikiProject.

It's OK to remove this message. Also, to stop receiving these messages, follow these opt-out instructions. Thanks, DPL bot (talk) 10:43, 15 February 2012 (UTC)

A kitten for you!

For the kittens -- could you consider smerging Joe Dan Mills Elementary School into the Austin schools article?

Bearian (talk) 22:42, 5 March 2012 (UTC)

Copied material from prod target to Austin Independent School District. Glrx (talk) 21:16, 11 March 2012 (UTC)

Feedback in oscillations

Hi Glrx. Have moved the Talk:Electronic oscillator discussion here as per Chetvorno's suggestion. Hope that is okay with you - please advise otherwise.

I agree that there is confusion around feedback - unfortunately the replies I'm getting aren't useful. I'm not trying to be disruptive; I have encountered a lot of confusion about what seems to be a simple enough concept, and most of that confusion I put down to the use of ambiguous terms. So I'm looking for some sort of guiding context to sort out the ambiguity.

Your comments about the restorative force may be a point in question. Agreed that the sign of the quantity isn't important. The fact that the force is "restorative" is enough to label it "negative feedback" (isn't it?) After all, the point of NF is to restore the status quo by opposing the change (positional difference) that gave rise to the feedback. Regards a positional difference vs a force... I understood that the direction of the force defines the type of feedback: positive = increasing the difference, etc.

And with that understanding, I don't see positive feedback as having any part in either oscillation or amplification. Where do we differ? What is your understanding? Trevithj (talk) 00:12, 6 March 2012 (UTC)

I'm not sure that I can help you. Oscillators are a difficult subject. I don't see ambiguous terms.

The sign of the loop gain is important. One cannot look at a particular gain (such as a spring rate k) within the entire loop, notice it is negative, and then assume that the entire loop gain is negative. My sense is that is what you are doing. Ditto for seeing anything that says "restorative" must mean negative feedback.

You cannot compare apples (position) and oranges (force). To see the loop, you must take the force and figure out how it subsequently influences the position. That negative restoring force is actually going to cause the pendulum to swing to the other side of the equilibrium point. A negative force on the positive side will create a negative position -- that's another negative factor. Very technically, the force gets integrated twice to develop a position; each integration introduces a sqrt(-1) factor. The loop gain at resonance is positive, so the feedback is positive.

To get an idea that something strange is going on, look at the velocity of the pendulum. As the pendulum mass is rising, that restoring force is decreasing the pendulum's velocity (a negative effect), but as the bob is falling, the restoring force is increasing the velocity (a positive effect). I have no trouble with that because I see a force being integrated and producing a 90 degree phase shift for velocity. If I integrate the velocity, I get another 90 degree shift for a total 180. With a negative spring constant, I get a positive feedback loop.

Oscillation and amplitude control are two different beasts.

Glrx (talk) 00:45, 6 March 2012 (UTC)

Thanks. You have been very helpful - I begin to see my source of confusion. It seems to come down to the point of reference, and so what therefore counts as a positive or negative quantity.

While I'm still not convinced that oscillation isn't a case of delayed amplitude control, hey - I've been wrong before!

Can we work an example to clear up one point? If I take an amplifier A with an input signal of 1mW, and an output signal of 10mW. I say A has a positive gain. I add a feedback loop B with a gain of (say) -3dB, so the feedback signal will be 5mW. I say B has a negative gain, but the overall loop gain is positive (5mW > 1mW). Is this terminology okay so far?

If so, does it necessarily follow that the 5mW feedback signal must reinforce the original 1mW signal, or can it negate the input somehow? Likewise, does a negative loop gain mandate a negation of the original signal, or can the smaller feedback still reinforce the input?

This is where I am hazy. Trevithj (talk) 23:31, 6 March 2012 (UTC)

No, you are confused. Do not use dB. B has attenuation (that what < 0dB means); the sign of the gain is in doubt. Glrx (talk) 01:16, 7 March 2012 (UTC)

Ah! Light shines! Of course, I'm assuming a log scale. So a negative gain does mandate negative feedback (and v.v). But wait a minute, dBs are a pretty common measure of gain. What measure are you using?

Also, going back to the above example, if the loop gain is positive, then the original input to A should keep increasing, as will the output. I would expect A to switch fully on, but instead it oscillates. What stops A from saturating? Trevithj (talk) 05:13, 7 March 2012 (UTC)

In your amplifier example, in addition to a loop gain of at least 1 (0 dB), the loop must have the correct time delay (phase shift) so the signal shifts in phase by an exact multiple of one cycle of the oscillation frequency (360N degrees) as it traverses the loop. Then the feedback sine wave will be in phase with the initial sine wave and they will reinforce. This is positive feedback. If the phase shift is an odd multiple of a half cycle (180 degrees) the feedback sine wave will be out of phase with the initial sine wave and they will cancel. This is negative feedback. --Chetvorno^TALK 19:11, 7 March 2012 (UTC)

Hi Chetvorno. Is that where the sign of the loop gain comes in... something like the cosine of the phase shift? Or does the loop gain have nothing to do with frequency? Trevithj (talk) 01:09, 8 March 2012 (UTC)

@Trevithj. Just look at numbers on a linear scale; the topic is (mostly) linear oscillators.

There are still many points of confusion, and I'm not sure how to point them out. When we're talking about a pendulum, we have an almost lossless system. The restoring force and loop gain arguments are in a resonator loop. A loop gain of (almost) one follows from lossless and conservation of energy. In the lossless ideal, the amplitude of the oscillation is set by the energy in the system. Start the system with a lot of energy, and the amplitude is large. The amplitude is not set by the loop gain.

In a resonator based electronic oscillator, there's an amplifier loop, but its purpose is not to create the oscillations but rather to compensate for the energy losses in the resonator. At small amplitudes, the amplifier adds energy to the LC resonator, so the amplitude increases (just like the pendulum swing would increase). At some point, the amplifier starts saturating, so it no longer adds energy to resonator but starts taking some out. That nonlinear mechanism sets the amplitude of the oscillator; there's an equilibrium point where energy into the resonator is balanced by the energy lost from the resonator (averaged over a cycle). But it is a subtle effect; the amplifier doesn't get stuck at one rail forever; there's a dynamic solution. Just because a gain is greater than one and there is positive feedback, that does not mean the output of an amplifier instantly goes to infinity.

Glrx (talk) 22:06, 7 March 2012 (UTC)

Okay, I'm wiling to accept that the amplifier is operating with positive feedback. Glrx, I don't think the amplifier "instantly goes to infinity", but it follows that after a time "the amplifier starts saturating". That is what I would expect positive feedback to do.

The above explainations seem to take the oscillation for granted. The impression I'm getting is that the resonator is being treated as a black box, and its details are ignored except as a source of the input signal. So the context of talking about the amplifier assumes an existing input signal (sine wave of desired frequency) that is presumably amplified to saturation via positive feedback.

My difficulty is that this comes across as a bootstrap argument: "an LC oscillator consists of an LC oscillator and an amplifier with positive feedback". It seems to avoid the issue of oscillation completely.

Trevithj (talk) 01:09, 8 March 2012 (UTC)

Good description. BTW, a mechanical analog of an electronic oscillator is a mechanical clock. The pendulum is the harmonic oscillator (analogous to the LC circuit or quartz crystal) and the escapement functions as a (very nonlinear) mechanical amplifier, forming a feedback loop to apply energy from an external source (the clock's mainspring or weights) to the pendulum with the correct phase for positive feedback, to keep it going. --Chetvorno^TALK 00:57, 8 March 2012 (UTC)

Hi Chetvorno. Sorry about the indent. We seem to have cross-posted. But the clock example is a case in point - the pendulum's oscillations are taken for granted, and we are left with a discussion about how losses are compensated for. — Preceding unsigned comment added by Trevithj (talk • contribs) 01:13, 8 March 2012 (UTC)

@Trevithj. Careful about the amount of positive feedback. If the feedback is positive but the loop gain is less than one, then the oscillation will die out. There's reinforcement and ringing, but there's not enough reinforcement to keep the system going.

@Glrx. Semantic differences here - I'd say that if the oscillations die out then by definition there wasn't reinforcing. Again, an issue of frame of reference? Trevithj (talk) 21:22, 8 March 2012 (UTC)

Well, what if there were no feedback, the resonator is injected with energy, and the resonator took 1 second to decay to the noise floor. Then some feedback is added, the resonator is energized, but not it takes 100 seconds to decay to the noise floor. I'd say there was reinforcement -- just not enough. I don't see it as a frame of reference issue. Glrx (talk) 22:19, 9 March 2012 (UTC)

Oscillator start-up is another difficult topic. Some clocks must be started externally. A grandfather clock, for example, needs somebody to start the pendulum swinging. I think that is also required for the typical chronometer escapement. Some clocks are designed to be self-starting, but its been a long time and I've forgotten the details.

Electronic oscillators are usually started with noise. The gain of the amplifier at low amplitudes is greater than one, so any noise near resonance will be amplified and grow. Oscillator start time is a function of the thermal/shot noise level (thermal noise power is kTΔB), the number of times that level must be amplified by the loop gain to get to the desired output level (the number of times around the loop), and the time it takes to make one loop. I said don't use dB, but it is often used to look at oscillator startup. The noise power might be -174dBm in 1Hz; assume the BW is 1Hz. Say the oscillator output is 10dBm. Therefore the noise must be amplified by 184dB. If the loop gain is 1dB, then we need on the order of 184 loops to start.

That makes sense. Good example. I suggest that an oscillator that has looped 183 times sounds like it has already started though. Reference point for started? Trevithj (talk) 21:22, 8 March 2012 (UTC)

But if the gain is only 0.1dB, then 200 times around the amplifier only gets the output 20dB above the noise floor. It's not the number of times, but rather the output power level. Glrx (talk) 22:19, 9 March 2012 (UTC)

So there's a design trade-off. To guarantee quick starting, the gain must be significantly larger than one. A large gain, however, implies there must be significant limiting, so there will be a lot of distortion and poor stability. See Meacham's article at Wien bridge oscillator.

@Chetvorno. Yes, and some escapements are fabulous machines.

Glrx (talk) 01:53, 8 March 2012 (UTC)

Maybe the stumbling block is the harmonic oscillator (LC circuit). An electronic oscillator doesn't necessarily have to have a harmonic oscillator (for example, RC oscillators don't) (although I believe it does have to have energy storage devices such as capacitors or inductors to create a phase shift) All it needs is a circuit that forms a feedback loop, with an amplifier in it to replace the energy dissipated in resistances to give a loop gain of one, and the correct phase shift around the loop, 2πN. --Chetvorno^TALK 02:16, 8 March 2012 (UTC)

I suspect you're right re the stumbling block. Perhaps separating the oscillating from the amplifying would be useful. It is the former that I'm trying to fathom.

If the LC circuit was replaced by a resistor of equal reactance - what would the circuit do?

If the LC circuit wasn't amplified in any way - would it oscillate? Ever?

130.216.36.84 Trevithj (talk) 21:22, 8 March 2012 (UTC)

There are many stumbling blocks. Using a resonator-based helps keep some notions separate. In an RC oscillator, the amplifier must must create complex poles and control the amplitude.

Resistors don't store energy, so the modified oscillator might just hiss slightly above the noise floor.

Without an amplifier (and not energy source), the LC tank would just sit there. If it had some initial energy, it would have a damped/decaying oscillation.

Glrx (talk) 22:19, 9 March 2012 (UTC)

Well, maybe an initial explanation in the Electronic oscillator article of how the LC tank oscillates would be helpful. Then PF can be introduced as a way of preventing the oscillations from decaying. Trevithj (talk) 08:11, 10 March 2012 (UTC)

There's already a good explanation in LC circuit. --Chetvorno^TALK 09:50, 10 March 2012 (UTC)

Its a very good explanation. Clearly LC oscillation is due to a causal loop, but positive or negative? This is very frustrating! Conclusive literature refs are hard to find.

BTW, I humbly accept that PF is a significant contributory cause to oscillation. My original concern stemmed from it being neither necessary nor sufficient cause. I'm now less sure how much emphasis to place on that. Trevithj (talk) 22:54, 13 March 2012 (UTC)

I think from the beginning your understandable questions stem from an unfortunate difference in meaning of the term "harmonic oscillator" between physics and electronics.

• In physics, a "harmonic oscillator" is a passive physical system with a restoring force which causes it, when displaced from its equilibrium position, to oscillate. Examples are masses on springs, pendulums, and, in electrical circuits, LC circuits and quartz crystals. But it doesn't have a source of energy, so it cannot produce continuous (constant amplitude) oscillations. Because it has friction, or (in an LC circuit) resistance, when given an initial energy in the form of a push, it will produce damped oscillations that die away to zero. The restoring force that returns a mechanical oscillator to its equilibrium position is not the same as feedback, positive or negative. It is a conservative force, which means that it doesn't change the net energy of the oscillator over a cycle. The gravitational restoring force that accelerates a pendulum when it is swinging down toward its bottom equilibrium position, converting its potential energy to kinetic energy, decelerates the pendulum when it swings back up, converting the kinetic energy back to potential energy. It can't add energy, it just switches the energy back and forth between two different forms (analogous to the LC circuit where the energy flows back and forth between an electric and a magnetic field). This leaves the friction force, which always opposes the direction of motion, to dissipate the energy, slowly reducing the amplitude of the oscillations to zero. Without an external driving force giving it "pushes" each cycle, this kind of harmonic oscillator cannot function like an electronic oscillator and produce continuous oscillations.

• In our Electronic oscillator article, the term "harmonic oscillator" is used for a circuit consisting of an amplifier with a feedback loop around it. The amplifier is a source of energy; it converts the DC electric current supplied to it to sine wave energy; it produces a signal at its output which has a greater amplitude (greater energy) than its input. Therefore it acts like an external driving force, giving the oscillations electronic "pushes" to replace the energy dissipated by resistance in the circuit. The sine wave of each oscillation passes around the feedback loop. As it passes through the feedback network the resistance in the network dissipates its energy into heat, reducing the amplitude (voltage) of the oscillation when it reaches the amplifier input. The amplifier adds energy to the oscillation, restoring it to its original amplitude by the time it gets back to the beginning. If the amplifier was taken out of the loop, or if its gain was reduced so the loop gain was less than one, the feedback loop would still oscillate if given an electronic "push" in the form of a pulse of current. But the oscillations would die out exponentially to zero , just like in the physics harmonic oscillator above. Only the energy added by the amplifier each oscillation to replace the energy dissipated in the resistance, can produce continuous oscillations.

You can see that the two meanings of "harmonic oscillator" are different. Roughly, if you add a positive feedback loop with a source of energy (amplifier) to the first ("physics") kind of harmonic oscillator, you get the second ("electronic") kind. We really should take the term "harmonic oscillator" out of this article, replace it with something else. --Chetvorno^TALK 00:51, 15 March 2012 (UTC)

If "harmonic oscillator" in electronics naturally implies amplification, and amplification naturally implies PF, then yes, I can see the difficulty. Regards a new term - how about "driven harmonic oscillator"? But then we're back to the division between the amplifier and the LC circuit again. (sigh) It is still difficult to distinguish between negative feedback and restoring force - the latter sounds like an example of the former. -- Trevithj (talk) 01:37, 16 March 2012 (UTC)

Disambiguation link notification for March 17

Hi. When you recently edited Binary search algorithm, you added a link pointing to the disambiguation page Recursive (check to confirm | fix with Dab solver). Such links are almost always unintended, since a disambiguation page is merely a list of "Did you mean..." article titles. Read the FAQ • Join us at the DPL WikiProject.

It's OK to remove this message. Also, to stop receiving these messages, follow these opt-out instructions. Thanks, DPL bot (talk) 10:39, 17 March 2012 (UTC)

AfD nomination of Plummer v. State

An editor has nominated one or more articles which you have on for deletion. The nominated article is Plummer v. State.

Your opinions on whether the article meets inclusion criteria and what should be done with the article are welcome; please participate in the discussion(s) by adding your comments to Wikipedia:Articles for deletion/Plummer v. State. Please be sure to sign your comments with four tildes (~~~~).

You are welcome to edit the article during the discussion to improve it but should not remove the articles for deletion template from the top of the article; such removal will not end the deletion debate. --Guy Macon (talk) 19:43, 17 March 2012 (UTC)

Objection to Proposed Deletion Notice of 27 March 2012

==Please Remove Template:Proposed deletion/dated 27 March 2012== and Refer to "Talk:SpiderGraph chart" for Author's "Objection to Deletion" and Response to the concerns mentioned in the 3/27 "Notice of Proposed Deletion" of the WP Article "SpiderGraph chart." Gregory L. Chester 00:00, 1 April 2012 (UTC) Gregory L. Chester 00:13, 1 April 2012 (UTC)

email address added: [email address deleted by Glrx] Gregory L. Chester 00:15, 1 April 2012 (UTC) — Preceding unsigned comment added by GregLChest (talk • contribs)

I removed my proposed deletion of SpiderGraph chart, but I still believe the article should go. I will nominate the article for deletion later. Glrx (talk) 16:02, 1 April 2012 (UTC)

A barnstar for you

		The Modest Barnstar
		Thanks for your recent contributions! 66.87.0.210 (talk) 20:19, 2 April 2012 (UTC)