User talk:David Eppstein

Hi, and welcome to my User Talk page! For new discussions, I prefer you add your comments at the very bottom and use a section heading (e.g., by using the "New section" tab at the top of this page). I will respond on this page unless specifically requested otherwise. For discussions concerning specific Wikipedia articles, please include a link to the article, and also a link to any specific edits you wish to discuss. (You can find links for edits by using the "compare selected revisions" button on the history tab for any article.)

Happy New Year, David Eppstein!

Happy New Year!

David Eppstein,
Have a prosperous, productive and enjoyable New Year, and thanks for your contributions to Wikipedia. Poepkop (talk) 16:08, 1 January 2016 (UTC)

Spread the WikiLove; use {{subst:Happy New Year 2016}} to send this message

.

Your GA nomination of Reuleaux triangle

The article Reuleaux triangle you nominated as a good article has passed ; see Talk:Reuleaux triangle for comments about the article. Well done! If the article has not already been on the main page as an "In the news" or "Did you know" item, you can nominate it to appear in Did you know. Message delivered by Legobot, on behalf of Spinningspark -- Spinningspark (talk) 22:42, 2 January 2016 (UTC)

Your GA nomination of Binary logarithm

The article Binary logarithm you nominated as a good article has passed ; see Talk:Binary logarithm for comments about the article. Well done! If the article has not already been on the main page as an "In the news" or "Did you know" item, you can nominate it to appear in Did you know. Message delivered by Legobot, on behalf of Jfhutson -- Jfhutson (talk) 22:43, 2 January 2016 (UTC)

Hi Ed, I got an email notification from you

but I didn't get an email. Do you think you could resend it? Thanks :-) Serendi^podous 08:16, 3 January 2016 (UTC)

@Serendipodous: I have no idea what you're talking about, and my name's not Ed. You sure you left this message in the right place? —David Eppstein (talk) 08:22, 3 January 2016 (UTC)

Oh, sorry. Clicked the wrong talk page I think. :-) Serendi^podous 08:33, 3 January 2016 (UTC)

Your GA nomination of Stars (M. C. Escher)

Hi there, I'm pleased to inform you that I've begun reviewing the article Stars (M. C. Escher) you nominated for GA-status according to the criteria. This process may take up to 7 days. Feel free to contact me with any questions or comments you might have during this period. Message delivered by Legobot, on behalf of Dr. Blofeld -- Dr. Blofeld (talk) 13:00, 4 January 2016 (UTC)

A very fine article, one of the most professional I've read in a long time. You being a computer science professor explains it I guess. Keep up the terrific work.♦ Dr. Blofeld 13:12, 4 January 2016 (UTC)

Thanks! —David Eppstein (talk) 18:30, 5 January 2016 (UTC)

consider

dear first consider if it is not reliable mention the reason without reason every edit you delete why...... lot people are here for editing wikipedia. — Preceding unsigned comment added by HannuMannu (talk • contribs)

Context: two reverts to Square (algebra) and Symmetry, and a comment on Wikipedia:Articles for deletion/Books written in unconventional ways. HannuMannu was later blocked for (among other things) edit-warring on the same edits to Square and Symmetry, which were reverted by two other editors. —David Eppstein (talk) 18:25, 5 January 2016 (UTC)

DYK for Binary logarithm

Cas Liber (talk · contribs) 00:02, 6 January 2016 (UTC)

Recently removed "linkspam"

Hello professor!

You recently undoed my edit with the change description "linkspam", where I linked to a website where I implemented the recursive maze-solving algorithm as written in the wikipedia article. Two years ago, I added the recursive algorithm to the "maze solving algorithm" page (https://en.wikipedia.org/wiki/Maze_solving_algorithm#Recursive_algorithm), with sample code in Java. I recently learned javascript, and was able to recreate the same algorithm visually, and I thought linking to it as a sample implementation was appropriate. Please let me know why what I did was not allowed, and if there is any other way to provide readers with a better understanding of the recursive algorithm.

Thanks in advance, Ofek Gila!

Ofek Gila (talk) 04:02, 7 January 2016 (UTC)

See WP:EL for what links are appropriate or not, and WP:COI for adding links to resources that you created. —David Eppstein (talk) 04:35, 7 January 2016 (UTC)

Cuckoo hashing

In reference to:

https://en.wikipedia.org/w/index.php?title=Cuckoo_hashing&oldid=prev&diff=698589296

"Cuckoo hashing is not as fast as quadratic probing in general-purpose hashing applications on most modern processors. However, cuckoo hashing may outperform the alternatives in various special cases, when there are other constraints on the design of the data structure."

(Undid revision 698587282 by Aaron Will Claims like this cannot be included unless they are backed up by reliable sources. Also the studies I've seen show linear probing as fastest, not quadratic probing.)

Quadratic probing is just a generalization of linear probing; linear probing *is* quadratic probing. If the hash function is perfectly uniform, quadratic probing has no benefit; in the academic references linked from the page, the authors are typically just using random data, and side-stepping the hash function itself. So they will just compare to linear probing as representative of both linear and quadratic, because there's no difference. But perfectly uniform hash functions are slow; faster compromise hash functions that are actually used in practice are prone to 'funneling,' and quadratic probing is the mitigation. So a real-world hash table will always use quadratic probing, not linear probing; but they're approximately the same thing. But I think it sounds goofy to say, without qualification, that "linear" is faster, because no high performance table is using linear probing, because it gives poor results with fast hash functions.

This is not a novel or controversial result, as far as I'm aware. It's been well-known for decades, and I'm sure it's discussed more carefully on the main hash table page.

I am happy to word this any way you want. But the article is misleading in its current state without containing some form of this remark, making irrelevant claims about how much faster it is than chaining, which is sort of like saying how much faster Shell sort is than bubble sort. I'd honestly be much happier if, rather than reverting my edit, you reworded it to address your concern. What do you think? Aaron Will (talk) 20:49, 7 January 2016 (UTC)

The studies I've seen have been based on reasonable quality hash functions, but with that caveat they show that linear probing is much better than other probing strategies, because of its improved locality of reference. Anyway, you're asking the wrong question. What I think about hashing is not important. What you need to do before inserting any such comparison is to find publications that actually address this issue, and report what they say about it, rather than putting your own editorializations into Wikipedia without sourcing. —David Eppstein (talk) 21:39, 7 January 2016 (UTC)

PS Another consideration is that both cuckoo hashing and linear probing allow for fast deletions; I don't know how to do this easily for quadratic probing. As for the claim that speed requires the use of a bad hash function, and an overly complicated probing sequence to make up for the bad hash function: I don't actually believe this. See e.g. http://arxiv.org/abs/1503.03465 for a pretty good hash function (I'm not sure whether good enough for linear probing, but better than commonly used bad ones) that is also faster than the commonly used bad hash functions. Again, though, what I believe isn't important, it's what the reliable sources say. But we definitely shouldn't say that it's not possible to have hash functions that are both high quality and fast, because the sources contradict this. —David Eppstein (talk) 23:12, 7 January 2016 (UTC)

Look I just want to fix the article. I don't have a cite. What about Primary clustering and Quadratic probing, which basically describe this problem and why quadratic probing helps, also mostly lacking cites? Can we just cite that? Linear probing and quadratic probing, for all purposes of this discussion, are the same thing. Linear probing counts using an, and quadratic probing counts using an + bn^2. You say 'good' and 'bad' hash functions, but the reality is that any performance hash function over general purpose data is a compromise.

'Good' hash functions typically don't have really bad funneling, but they still tend to have some degree of clustering for hashing inputs that are very related--this being the normal case in practice. And check out the paper you cited; they admit that the version they're actually doing speed tests on fails the avalanche test, a catastrophic failure for a practical hash function, and that a modification is necessary. But just having good 'avalanche' still doesn't guarantee you don't have clusters; just stealing and tacking on murmur3's mixing step isn't going to be enough to magically give them uniform output for related keys, even if it does make the test pass. The other thing to keep in mind is that, in practical implementations, quadratic isn't slower than linear, so there's never any reason to implement purely linear probing unless you don't really care about performance, or you are on a specialized platform that doesn't have multiplication, or something. Finally quadratic probing generally is only counts differently from linear after we've had a couple collisions in a row, which should already be an outlier case--otherwise something is wrong and our hash table is going to go way too slow. This isn't some novel result I just came up with; like I said, it's well-known, and any high-performance hash table implementation has at least this level of sophistication. (Ironically I checked my language's library just now, and it's using chained buckets, so it's not even participating in this conversation.)

I'm frustrated because I don't really like the way you're engaging me here. I don't think you should be reverting this statement merely because it's uncited, especially since you seem to concede that it's more-or-less true, modulo the foregoing distinction between linear and quadratic, which isn't really the key point I'm trying to make; my point is that Cuckoo hashing is not that fast compared to the pre-existing state-of-the-art. Please see WP:ROWN for why I feel this way. If I had some other general term for cache-friendly open addressing scheme, I'd use it, to avoid your concerns about 'quadratic,' but as far as I'm aware, there really aren't any other alternatives. Cache lines aren't big enough for there to be too many different strategies here; but quadratic probing is a generalization of all of them, as far as I'm aware. Aaron Will (talk) 03:59, 12 January 2016 (UTC)

A possibly interesting reference (though I'm not convinced it passes Wikipedia's tests for reliability): http://rcoh.me/pyhash.pdf. It claims that replacing Python's built-in dictionary with linear probing and tabulation hashing led to significant speedups on test data and smaller but still nonzero speedups on more realistic data. I'm also aware of recent work in computer security in which using a hash table that can be made to have collisions is viewed as a security violation (because it can lead to denial of service attacks and back door communication channels). But we really really can't take your personal experience as a source, and using another Wikipedia article for a source is no better. If those other articles are making similar assertions without proper sourcing then they need to be fixed. —David Eppstein (talk) 04:33, 12 January 2016 (UTC)

Your GA nomination of Stars (M. C. Escher)

The article Stars (M. C. Escher) you nominated as a good article has passed ; see Talk:Stars (M. C. Escher) for comments about the article. Well done! If the article has not already been on the main page as an "In the news" or "Did you know" item, you can nominate it to appear in Did you know. Message delivered by Legobot, on behalf of Dr. Blofeld -- Dr. Blofeld (talk) 23:21, 8 January 2016 (UTC)

Cliques = connected components in graph of equivalence relation.

Hi, I'm the foolish bunny who perpetrated this change you reverted about the connection between the maximal cliques and connected components in the graph representing an equivalence relation. I think of the clique as a set of vertices with a property (distinct vertices connected) and the connected component as a graph (set of vertices + set of edges). Even taking a clique as a graph, it still doesn't seem to be equal to a connected component: each connected component has a loop on every vertex due to the reflexivity of the equivalence relation; no such loops are implied by the clique. Could you explain in what way "the maximal cliques *are* the connected components"? Frentos (talk) 11:23, 10 January 2016 (UTC) PS As a matter of form, should I have created a discussion on the talk page for equivalence classes and just left a comment here directing you to the discussion? Frentos (talk) 11:30, 10 January 2016 (UTC)

If you think of a clique as a subgraph rather than a set of vertices, and ignore the loops, then the maximal cliques are exactly the connected components. Your wording made it sound like there could be multiple clques per connected component, which I think is misleading. Also, there is no difference between "maximal cliques of the connected components of G" and "maximal cliques of G", so your wording was unnecessarily indirect. —David Eppstein (talk) 16:31, 10 January 2016 (UTC)

I agree my replacing "forming" with "of" was misleading; thank you for reverting it. The existing wording, "maximal cliques forming the connected components of the graph" makes it sound like the maximal cliques (as subgraphs) are the connected components when they are proper subgraphs of the connected components, which I think is misleading. The reference (Devlin, p. 123) makes no mention of cliques, but explicitly notes reflexivity of the connected components, which cliques lack. It would be clumsy to add a qualification about adding or ignoring loops to make the statement correct; perhaps it would be clearer and more concise to remove the reference to cliques and say the equivalence classes are represented by [the vertex sets of] the connected components? — Frentos (talk) 20:38, 10 January 2016 (UTC)

Maybe, but I think it's important that these components are complete+self-loops, rather than just being any kind of connected graph. —David Eppstein (talk) 21:31, 10 January 2016 (UTC)

The correspondences I can think of between the equivalence relation ~ on the set X and the graph as described are:

1. Each element of X corresponds to a vertex (by construction)
2. Each s~t corresponds to an edge (by construction)
3. Each equivalence class corresponds to the vertex set of a connected component
4. The disjointedness of equivalence classes corresponds to the disjointedness of connected components
5. The reflexivity of ~ corresponds to the loop on each vertex
6. The symmetry of ~ corresponds to either no edge or 1 undirected edge between any distinct pair of vertices
7. The transitivity of ~ corresponds directly to any two adjacent edges being part of a triangle and indirectly to completeness of each connected component
8. The property that any two distinct elements of the same equivalence class are related by ~ corresponds to the completeness of a connected component
9. The property that any two elements of the same equivalence class are related by ~ corresponds to the completeness of a connected component + the loops on each vertex

1& 2 are redundant. The article is trying to say 3 (which implies 4) and mention the graph part of 7 (or 8 or 9) in passing in just one sentence. Maybe split it into two sentences: one for the equivalence classes and one for the reflexivity<=>loops and transitivity<=>component completeness? A simple diagram? I'd be happy to knock up a simple undirected diagram ala Devlin. Frentos (talk) 01:24, 11 January 2016 (UTC)

Tetrahedron

Stop changing my edits to tetrahedron Ghostasylum (talk) 16:34, 11 January 2016 (UTC)

Directed Acyclic Graph updates

Professor Eppstein, you asked whether a Bitcoin block processes data. A Bitcoin block is a group of transactions confirmed by the world's most powerful computer network (See Forbes article). The network processes financial transactions through the blockchain as a linked list. This leads to a competitive race using proof-of-work to verify transactions were not double-spent, something only previously possible with a clearing house or other financial trusted third party. There has been considerable talk around using DAG data structures in crypto currencies, particularly Bitcoin, rather than a linear linked list. This adds complexity to the system, but reduces confirmation time by orders of magnitude with network latency and CPU power become the only bottlenecks. At this time, DAGs are only in the proof of concept stage, but it is a potentially revolutionary use of the data structure.

Perhaps this use case belongs in another section, but I believe that prior to the discussion of DAG usage in Bitcoin, many people who are not data scientists, myself included, had never heard of this data structure. — Preceding unsigned comment added by Mfactor (talk • contribs) 16:26, 15 January 2016 (UTC)

A cookie for you!

Thanks for cleaning up Caleb Frank Gates. Howicus (Did I mess up?) 16:40, 17 January 2016 (UTC)

User:Arenash13

Hi, David. The above-mentioned user has recreated two articles, Local cosine tree and Cohen's class. Neither article was ineligible to me, but I found two other articles which seemed suitable redirects (Trigonometric functions and Bilinear time–frequency distribution) and have converted accordingly. Since the previous articles were deleted as copyright violations, I am not certain how these two new articles compared— in any case, as redirects, copyright issues have been obviated, but I wanted to give you a heads-up on the business so you could keep an eye on what's happening in the world of, er, of abstruse mathematics! Cheers! KDS4444^Talk 06:29, 18 January 2016 (UTC)

Ok, thanks for the heads-up. Anyway, the deletion was only over the copyright violation, so if that's resolved by redirection then I guess the recreation isn't a problem. I'll check the re-created versions and revdel them if necessary to clear out any new copyvio. —David Eppstein (talk) 08:14, 18 January 2016 (UTC)

Hurwitz theorem

It's not obvious that it is true for rationals too. At least, I can't see it. The only thing I did was to follow what's written in the article that was referenced to https://en.wikipedia.org/wiki/Hurwitz%27s_theorem_(number_theory). Take a look and tell me what you think.Wisapi (talk) 19:29, 18 January 2016 (UTC)

A barnstar for you!

	The Barnstar of Diligence
	For your job of doing what few have dared, which was to close this move review. Thanks for the tireless effort and extensive reading and reviewing you certainly had to endure. Tiggerjay (talk) 23:36, 18 January 2016 (UTC)

You're welcome! Thanks for the barnstar. —David Eppstein (talk) 01:11, 19 January 2016 (UTC)

[Redacted]

.... And thanks for the transclusion on the Garamond talk page. As I said, I'm probably going to start doing GA reviews once this one or maybe the next have gone through and I've had a chance to get a bit of feedback. Blythwood (talk) 17:38, 30 January 2016 (UTC)

Tic Tac Toe - "up to"

The problem is that "up to" isn't clear. It may have a precise mathematical meaning, but it's meaning in general usage isn't so clear. I think a casual user would translate this more or less as "depending on". But "depending on" in which sense? Inclusion or exclusion? A wording that is very explicit about whether the redundant games are being included or excluded would be better. Probably better to just say something like "xxxxx games, if including all mirrored and rotated versions of the same game". Battling McGook (talk) 14:53, 20 January 2016 (UTC)

revert

Why did you revert my edit here : https://en.wikipedia.org/w/index.php?title=Packing_problems&oldid=prev&diff=700952485 ? — Preceding unsigned comment added by Lovasoa (talk • contribs)

Two reasons. The main one was that it looked like a personally-controlled website of the type that WP:ELNO #11 says not to add. But also, as a genetic algorithm, it is a heuristic, not a fast exact solution, and therefore not a valid source for the sentence to which you added it as a footnote. —David Eppstein (talk) 21:26, 21 January 2016 (UTC)

Grodziskie

Hi David;

It's been a while and I haven't had much (any) time to spend on Wikipedia recently but I know you spent a lot of time on the GA review of the Grodziskie article. I've responded to your comments on the GA review page if you'd like to take a look. Neil916 (Talk) 06:37, 26 January 2016 (UTC)

YGM

Hello, David Eppstein. Please check your email; you've got mail!
It may take a few minutes from the time the email is sent for it to show up in your inbox. You can at any time by removing the {{You've got mail}} or {{ygm}} template.

EEng 19:16, 28 January 2016 (UTC)

Image use policy

A policy needs written coherently. Here's what the text modifies:

“ Except with very good reason, do not use px (e.g. thumb ”

If we're going to take such an extremely hard line, we also need to note at least the uncontroversial cases where using upight is simply impossible. I don't think it's out of the range of policy to say "however, uprgiht can't be used when you're looking at things that need more than a 10px level of specificity, such as very small images, and also, some of our features requeire px" when the policy would otherwise be read as "use it anyway, event hough it won't work"

I don't think the part you modified should be policy, either. It's a guideline and should be in the guideline, not repeated here. (The more frequent case when you can't use upright is in infoboxes and other templates that display images but aren't set up for upright image sizing.) —David Eppstein (talk) 04:54, 29 January 2016 (UTC)

I'd be fine with that, but I think we need both or neither. Adam Cuerden ^(talk) 05:04, 29 January 2016 (UTC)

Please from the bottom of my heart.Can you have a look?

User:Only seems to be the suckpuppet of User:GiantSnowman which seems to be a Wikipedia master eraser: Max Merkel, Alex Notman, Slaven Bilić and many others articles, he has a really long edits history, this user needs special watching ! Also User:Only acuse people for vandalism and deleting referenced text, User:GiantSnowman prefers to deletes around instead of doing the job properly, they have a mini band and act like superior users which is meatpupetry, is like mafie of wikipedia in my opinion !--176.221.34.226 (talk) 22:43, 29 January 2016 (UTC)

176.221.34.226, GiantSnowman is an admin, been around here a long time and is the person I goto for football advice. Only is also an admin and has been around here for a long time. Neither one are sockpuppets. They are reverting your edits because there are no references. You need to add refs. I would revert too. Please add references. Bgwhite (talk) 22:49, 29 January 2016 (UTC)

see [1] — Preceding unsigned comment added by 103.1.149.90 (talk) 22:59, 29 January 2016 (UTC) GiantSnowman deletes massive content of Wikipedia and expect other people to do the work, even if you go on other language wikipedia you can probably find reference! — Preceding unsigned comment added by 103.1.149.90 (talk) 23:01, 29 January 2016 (UTC)

I am worried for ENGLISH WIKIPEDIA PROJECT ! English wikipedia articles will be "white" blacked with no information because of this attitude !--103.1.149.90 (talk) 23:03, 29 January 2016 (UTC)

People are to busy to block sockpuppets instead of helping the quality and quantity of articles !! ANd because of this people will not edit anymore !! or will get block !! is a major problem in my opinion !! (I want the best in a bright future for Wikipedia !)--103.1.149.90 (talk) 23:06, 29 January 2016 (UTC)

Please worry less about sockpuppetry and more about formatting your references so that they look like references. I agree that some of the links you added at the end of your addition to Max Merkel look like properly published news stories about him that would be useful in the article, but the way you formatted them makes them look like spam and did not use them to source the claims in the text you added. A hint: if you're using the visual editor, you can insert a reference by clicking on the big "cite" tool in the toolbar, pasting your url into the box that comes up when you do that, and hitting the "generate" button. It will look a lot better than the way you did it. And if you're not using the visual editor, you can switch to it at any time in any article by editing the article the other way and then clicking on the big black pencil icon on the upper right. —David Eppstein (talk) 23:16, 29 January 2016 (UTC)

Seneff

I was wondering, whether you, since you are a computer scientist yourself, thought an h-index of 48 on Google Scholar [2] is enough to establish notability for Stephanie Seneff under WP:PROF#C1. I know the h-indices needed to establish notability vary from field to field, and sciences require higher ones than the humanities, so I wanted to hear what you thought. Note that the article is up for AFD. Everymorning (talk) 17:44, 31 January 2016 (UTC)

Thanks, I'll answer at the AfD. —David Eppstein (talk) 17:46, 31 January 2016 (UTC)

Insoluble

I came across this reference in "Albert Einstein Philosopher-Scientist". Can you give me an English wording so I can at least look it up.

Well I was going to paste a picture but I'm not able to. The best I can do us describe it: uE'vEgig. The u is probably mu. The first g may be a summa. It is in Ilse Rosenthal-Schneider's essay "Presuppositions and Anticipations". Gsteel57 (talk) 22:49, 31 January 2016 (UTC)

I don't have any idea what you're talking about or which Wikipedia article this pertains to. Could you give me some more context, please? —David Eppstein (talk) 23:04, 31 January 2016 (UTC)

Andreas Andersen

How do I get this article you recently deleted undeleted (as a draft is fine). I was in the middle of improving it with better references (which was the previous consensus for having deleted it). For example, I was going to add a link to one of his recent interviews: Interview with Dr. Andrew Andersen by Teimuraz Toumanishvilli. It probably should be named "Andrew Andersen" too since though he has used the name Andreas, that does not seem to be the name he is most notable for. Thanks. 50.126.125.240 (talk) 04:09, 2 February 2016 (UTC)

In this case, because it was deleted through a recent deletion discussion (Wikipedia:Articles for deletion/Andrew Andersen), the best way is probably to go to WP:DRV and argue that there is new evidence that the subject was notable that was not available when the discussion happened. —David Eppstein (talk) 04:19, 2 February 2016 (UTC)

That seems like much tedious lobbying to get back to where I was just a few days ago—*sigh* Is there no way to resolve this with you (the closing/deleting admin)? WP:DRV instructions suggest first appealing such here. I admit the article when I first saw it was a mess and mostly filled with primary sources that were not by themselves of much use to Wikipedia but it was tagged as having no sources so I added some and cleaned up the primaries (several are cited by other papers and I had not gotten to those yet). I expected that would buy me a little time (I do have other things to do than edit Wikipedia all day) and then suddenly the article disappeared having been slapped with a speedy G4. So now I have lost all my cleanup work hidden behind a delete wall (it took quite a while to find links like ISBNs and OCLCs, etc. to cited primaries that had just been dumped on the page). 50.126.125.240 (talk) 04:23, 2 February 2016 (UTC)

Well, if you want to work on it in Draft: space, I can restore it for you for that purpose. But you're going to have to go through DRV either now or later (to get it moved back into article space without getting G4 speedied again). —David Eppstein (talk) 05:07, 2 February 2016 (UTC)

I am surprised G4 speedy still applies for articles that successfully make it through WP:AFC (especially since this was based on notability). That said, after (painfully) reading the pissing match put up my the fanboys in the Wikipedia:Articles for deletion/Andrew Andersen entry from almost six months ago, I feel sufficiently disenchanted to not want to work on an argument against the deletionists. It is too bad too, I did enjoy researching him and his work and I do believe there is sufficient material to meeting notability requirements (his research topics are not the most notable to begin with and then they are scattered across multiple names and languages and conflated with other people with similar names). Perhaps I can revisit him later and until then work on other fronts. Thanks. 50.126.125.240 (talk) 07:07, 2 February 2016 (UTC)

Ross McKitrick

Hi David Eppstein (talk). I am wondering how we can somehow keep the information about the YouTube videos and knowledge claims in the videos by Ross McKitrick. Could we mention that they are self-published? They are very revealing in regards to scholarship or lack thereof. They were cited in our local newspaper recently and I listened to them all. I am surprised a university professor would produce them. I suggest we keep the material and ask other editors if the content needs to be deleted.Oceanflynn (talk) 04:39, 3 February 2016 (UTC)

Nonsense

Please do not just remove nonsense posts from talk pages. It makes it harder to identify users who need to be removed from Wikipedia. I appreciate that you are trying to keep talk pages clean, but just deleting nonsense doesn;t help. Thank you. Robert McClenon (talk) 08:04, 4 February 2016 (UTC)

The nonsense is all still there in the talk page's history, if you need it to build a case against the nonsenser. But otherwise, it doesn't serve the purpose of talk pages, which is solely to discuss improvements to the article. —David Eppstein (talk) 08:47, 4 February 2016 (UTC)

My god, does everything have to be an argument? EEng 13:02, 4 February 2016 (UTC)

Thanks!

Thanks for the super quick COI edit! Ron Schnell 19:42, 4 February 2016 (UTC)

Please don't go!

Please reconsider, David. Your knowledgeable, well-written and lucid contributions are very necessary on that discussion page to counter the welter of ignorant, impenetrable and illogical contributions of the cantankerous. The MoS will end up being far worse without your continuing contributions... BushelCandle (talk) 12:06, 7 February 2016 (UTC)

Error in quickselect?

Hi I think the final line of the quickselect pseudo code has an error on the Quickselect page:

  return select(list, pivotIndex + 1, right, n)

should be

  return select(list, pivotIndex + 1, right, pivotIndex - n)

because n = (pivot index - n ) when n > pivot index

eg: n = 7 pivotIndex is 5; we want to search the right for 7-5 now not 7 — Preceding unsigned comment added by 2601:647:200:2FBA:2815:1A84:7D0:CC2E (talk) 23:39, 7 February 2016 (UTC)

Please read the comment at the top of that block of pseudocode. —David Eppstein (talk) 23:43, 7 February 2016 (UTC)

Practical Numbers

Do you think that greatest prime factor should get its own page, or that it should start out as a section of something else like prime factor?

This is of course a loaded question, because I would like to put on the practical number page:

From the above characterization by Stewart and Sierpinski it can be seen that if $n$ is a practical number, and $m$ is any number such that $gpf(m)\leq 1+\sigma (n)$, then $n*m$ is a practical number. It follows that if $m\leq 1+\sigma (n)$, then $n*m$ is a practical number. Also, because all practical numbers are powers of 2, perfect, or abundant, if $m\leq 2*n$, then $n*m$ is a practical number.

I haven't seen these statements anywhere else, so I'm worried this might border on original research. — Preceding unsigned comment added by Eassin (talk • contribs) 03:39, 8 February 2016 (UTC)

I would think it should go in Kempner function, since the greatest prime factor and the value of the Kempner function are the same for almost all n. —David Eppstein (talk) 03:49, 8 February 2016 (UTC)

Steven Gubser

Hi, David Eppstein I remove the link: http://www.gf.org/fellows/16630-steven-s-gubser, since I can't open it. Or say when I open it, the page show 404 - PAGE NOT FOUND for me, so I remove it. (Excuse me for I didn't explain at that time) If you find new available link, please update it.

When something like this happens, please try searching for an archived copy of the link on archive.org rather than just removing it. —David Eppstein (talk) 02:16, 10 February 2016 (UTC)

Editor of Wikipedia Lute88 several times posted photo sculptures of Shmatko

1. Editor of Wikipedia Lute88 several times posted photo sculptures https://en.wikipedia.org/wiki/Nikolay_Shmatko#/media/File:Shmatko-statue.jpg of Shmatko (original) http://www.kingofmarble-shmatko.com/en_vers/w264z.html , inventing a name. This can be seen in the pages of history.

2. Editor Lute88 at his discretion voiced sense of sculpture

1/ I ask to show permission from the author who can offer to place images of his sculptures for public review.

2/ I ask provide the description of the sculpture, voiced by the author.

3/ Please show an authoritative source, where the sculpture (original) http://www.kingofmarble-shmatko.com/en_vers/w264z.html is presented as kunniligus. In the photograph we can see, a kiss on the belly.

--Rerter 2 (talk) 17:04, 10 February 2016 (UTC)

Please let's keep the discussion on Talk:Nikolay Shmatko‎ where it belongs, and also focus more on the topic of the article and less on how much you dislike another editor of the article. —David Eppstein (talk) 17:18, 10 February 2016 (UTC)