Wikipedia talk:WikiProject Mathematics/PlanetMath Exchange

From Wikipedia, the free encyclopedia
Jump to: navigation, search

For older discussion, see Archive 1

Random variable/distribution[edit]

One of the major things I can see from this list is that on PM, whatsit random variable seems to be the convention, while on WP it is whatsit distribution. I like the latter, but I'm surprised to see that no whatsit random variable redirects exist. For example, normal distribution but not normal random variable? Surely the latter is a legitimate page title and thus a candidate for a redirect. It's a long time since I've thought about such things, but can anyone tell me if there is a good reason where there are so few redirects from ... random variable to ... distribution? This topic must have come up at some point, but I can't find any discussion of it. Ben Cairns 13:39, 1 Feb 2005 (UTC).

Note: I moved this here from Wikipedia talk:WikiProject Mathematics/PlanetMath Exchange/60-XX Probability theory and stochastic processes — see above thread on Talk pages. CryptoDerk 15:32, Feb 1, 2005 (UTC)
Crypto, man you have a lot of patience to move all those talk pages! Oleg Alexandrov | talk 16:09, 1 Feb 2005 (UTC)

Fair enough -- I figured it was a problem specific to probability, but I agree with your comment (now on the archived page). Still looking for an answer to the above, if there is one... PS: thanks for moving this here. Ben Cairns 23:01, 1 Feb 2005 (UTC).

New articles on PlanetMath?[edit]

Will we try to update our PlanetMath article lists as new articles are created on PlanetMath? Paul August 15:31, Feb 1, 2005 (UTC)

These days I will try to learn how to fetch wiki-markup code from Wikipedia, and how to upload it back. I have some hints that this should be possible. If so, it will be very easy to automate completely the process of maintaining the lists. Then, from time to time I would just need to run my script to update these pages with new PlanetMath articles.
And in connection with Ben's comments. If it becomes easier to maintain these pages, it will be possible to suggest more intelligent guesses for corresponding Wikipedia articles and then again rerun the script to update the pages (this will not overwrite any comments or anything already in there). Oleg Alexandrov | talk 16:09, 1 Feb 2005 (UTC)

Great ;-) Paul August 17:46, Feb 1, 2005 (UTC)

I imaging since we're signing our comments in the list after processing each article, it will be possible to compare timestamps. CryptoDerk 19:32, Feb 1, 2005 (UTC)

If the WP field is correct please remove the qualifier "guess"[edit]

I've updated template:Planetmath instructions to add: "if the WP field is correct please remove the qualifier "guess". I'm going through the lists and removing "guess" for the entries whose status has already been updated. Paul August 17:58, Feb 1, 2005 (UTC)

Mathematics bibliographies?[edit]

The "General Topology" list has the following entry: bibliography for topology. I don't think such biliographies currently exist on WP, although I'm surprised they don't. Do we want to include articles like this? I think it probably would be a good idea, but I'm not certain how they should be named and organized. I think all of Wikipedia should be included in such a discussion. Comments? Paul August 19:46, Feb 1, 2005 (UTC)

Such articles do exist on Wikipedia, List of publications in mathematics. Subheading for Topology is empty right now. Have fun :) CryptoDerk 19:53, Feb 1, 2005 (UTC)
There are actually several subheadings for topology there. That page needs some major work and reformatting to keep everything consistent :( CryptoDerk 20:31, Feb 1, 2005 (UTC)

Errors with PlanetMath's database & history[edit]

Last week I mentioned to Aaron that I couldn't view the history of PlanetMath articles. PM was only showing the current article and those versions modified in the past couple of weeks. He said he'd restore them from backups, but I figured I'd mention it here in case you all are looking, and because I've run into a new problem. Some articles I try to view on PM are just blank, for instance [1] and [2]. So, if an article is blank, don't change the status, just check back later. CryptoDerk 21:20, Feb 1, 2005 (UTC)

Update: Aaron says "some of those entries just need to be rerendered, from the rather turbulent period we just went through" and he is "still working on the version history". CryptoDerk 21:35, Feb 1, 2005 (UTC)
I've had a little trouble looking at some PM pages today. Sometimes they come up with something like, "Timed out waiting for render", which I assume means that their server is under significant load and can't always produce the page in time (?). Trying again later usually fixes the problem (and both the cited examples above worked for me). Ben Cairns 06:47, 2 Feb 2005 (UTC).
Further conversations revealed that the background process that renders has been broken for "a while". If it times out, reloading will usually cause it to render, but if you just get a blank page (like I was earlier), it may have to be manually rerendered by PM staff. CryptoDerk 06:56, Feb 2, 2005 (UTC)

Unclassified PlanetMath articles[edit]

There is a list of unclassified PlanetMath articles at http://planetmath.org/?op=unclassified that might need to be processed as well. Fortunately there is only one article there right now. CryptoDerk 23:59, Feb 1, 2005 (UTC)

Well, it is not unclassified because it cannot be. Hopefully articles stay there only temporary, until somebody puts them where they belong. I will keep on eye there too from time to time. Oleg Alexandrov | talk 02:03, 2 Feb 2005 (UTC)

Copying the page doesn't always work![edit]

The alt text for images won't always work. Specifically in cases of long LaTeX lines. As an example, see Conway's chained arrow notation. I'll update the relevant instructions to tell people to watch out for this, but I wanted to let everyone know here as well. CryptoDerk 06:35, Feb 2, 2005 (UTC)

I guess there is a limit to how long the alt could be. It is good you noticed that. Oleg Alexandrov | talk 16:00, 2 Feb 2005 (UTC)

Joke articles on PlanetMath[edit]

I've found two joke articles on PM so far, [3] and [4] so beware (I marked them as such in the lists). Despite being true, they are minor statements and are not called "Siegel's axiom" or the "fundamental theorem of transcendence". Hopefully other such jokes will be easily discovered. CryptoDerk 17:15, Feb 2, 2005 (UTC)

Unsuitable articles on PlanetMath[edit]

Of course, much of PM's content is appropriate for WP also, but I just discovered a set of articles on neutrosophic probability, statistics, sets and logic that appear to be quite unsuitable. It seems that articles on neutrosophic sets, logic and 'neutrosophy' in general have been deleted from WP (they are still in Google's index, however). My guess is that on WP it was deemed a neologism, original research or somesuch. I think I've noted them all as Not needed on WP. Again its something to keep an eye out for, but I'm sure we'll catch most of these cases. Ben Cairns 06:24, 3 Feb 2005 (UTC).

The vfd for this article is at Wikipedia:Votes_for_deletion/Paradoxism-Neutrosophy Rich Farmbrough 23:43, 28 August 2005 (UTC)

List of mathematical topics[edit]

Just a note to say that I've added a recommendation to enter articles in the List of mathematical topics to the instructions on the project page. Ben Cairns 02:28, 4 Feb 2005 (UTC).

I might one day write a script to keep that list up to date automatically, by crawling from [[Category:Mathematics]] to all math cagories, and from there to the articles. As long as a given article is categorized, and that category is in another category, all the way down to [[Category:Mathematics]], I should be able to find it.
I don't know if I will do this, need to talk to other people to see what they think. In the meantime, you of course have a point! Oleg Alexandrov | talk 02:39, 4 Feb 2005 (UTC)
I actually had the same idea, so I'd say: go for it (I won't have enough time this month). I was planning to use the database dump instead of crawling the web site. As you probably realize, you should generate the list on a separate page instead of overwriting the list; and I would put all the links on one page. -- Jitse Niesen 12:03, 4 Feb 2005 (UTC)

Making it easier to convert TeX to Wiki markup[edit]

These days I am working on copying things from complex analysis. I got tired replacing the dollars with Wiki math tags, and I wrote a very small script to do that for me, along with other minor housekeeping dueties. That might not be of any use of you, but made my life a bit easier, so I post it here.

#!/usr/bin/perl
use strict;
use diagnostics;

undef $/; # allow it to read whole file in one line

$_=<>;    # read all in $_

s/\$\$\s*(.*?)\s*\$\$/\<math\>$1\<\/math\>/sg; # get <math> and </math> instead of $$..$$
s/\$\s*(.*?)\s*\$/\<math\>$1\<\/math\>/sg; # get <math> and </math> instead of $..$
s/\\displaystyle//g;                       # get rid of displaystyle
s/\\rvert/\|/g;  s/\\lvert/\|/g;           # substitute \lvert, \rvert by plain old vertical bars
s/\<\/math\>\./.\<\/math\>/g;              # put the period inside the formula instead of outside
s/\<\/math\>\,/,\<\/math\>/g;              #  same for comma

print "$_";                                # spit out the output 

It reads text from standard input in Unix and prints text to standard output. I call it from XEmacs, so it takes whatever text I have in a buffer, processes it, and replaces it with the new text. Oleg Alexandrov 07:44, 13 Feb 2005 (UTC)

Just a note, some articles use `` and '' for quotes instead of ", which are standard here. Also, some articles there use \mbox, which needs to be taken out. CryptoDerk 02:56, Feb 15, 2005 (UTC)

Statistics[edit]

What do you guys think of Wikipedia:WikiProject Mathematics/PlanetMath Exchange/Statistics? It occurred to me that it might be nice to see a breakdown of how things are progressing, so I put together a few very basic statistics.

I'd appreciate any feedback, or contributions (particularly counts of total and unique articles).

Also: if you read that page, you may notice a peculiar thing. More Wikipedia articles now sport the PlanetMath copyright notice (Template:PlanetMath attribution) than have apparently required merging or copying! This suggests either misuse of the copyright template, or a significant number of misclassifications. I am not a lawyer, but the copyright notice is ONLY required, I think, if copying actually takes place. Expression of an idea from PlanetMath in one's own words only requires a citation, using Template:PlanetMath reference. Conversely, if any text is actually copied, the article should be flagged as Merged or Copied, not More complete or Adequate (even if one of the latter is now true).

Cheers, Ben Cairns 02:26, 15 Feb 2005 (UTC).

Looks good. Regarding usage of templates, I think we all have been doing it correctly, but some people have been using them without consulting the lists. I know that prior to the project starting that someone (I believe Paul August) copied over an article without providing any link back to Planetmath. Perhaps he has since added the template, but hasn't updated the list yet. Also, User:Dbenbenn noted that amenable group had been copied over from PM a long time ago, so he added the template. Since then our article has been expanded a great deal, so when Mikkalai got to it, he flagged it as being more complete. These are just two instances, I'm sure there are more. CryptoDerk 02:52, Feb 15, 2005 (UTC)
CryptoDerk: thanks for clearing that up. I was getting a little worried, since I think that copyright is something to take seriously and we shouldn't saddle ourselves with superfluous licenses. Or maybe I'm just paranoid! Your explanation seems adequate to account for this. Ben Cairns 03:11, 15 Feb 2005 (UTC).
Interesting, how are you getting these stats? Are you doing it by hand? Are you using the stats from the project page, or are you recounting yourself? Paul August 02:58, Feb 15, 2005 (UTC)
Paul, I just copied Wikipedia:WikiProject Mathematics/PlanetMath Exchange into MS Excel (say what you will, it's handy for quick jobs), and edited the HTML into a table. Took about 5 minutes to do that, and another minute to generate totals and percentages. The drawback is that reviewed articles in categories that haven't been completely reviewed will not appear in the stats. The benefit is that any biases in article selection for review within categories will be mitigated.
To get the counts of pages using the templates, I went to 'What links here' for each template and counted by hand. Those numbers only go up to 500, so I'm sure this will become a rather imprecise method very soon. The total article count may also be rather poor... it should probably be just > 4000 since duplicates may mess things up. Probably it is the unique article count that is closer to the 4000 value. I'll make a couple of edits to reflect this. Ben Cairns 03:11, 15 Feb 2005 (UTC).
Addendum: I've now incorporated the above into the article, so it should be clear where the numbers came from. This was a bit of an oversight on my part, so thanks for the suggestion! Ben Cairns 03:50, 15 Feb 2005 (UTC).
Adjusted % not needed should be blank, and should not be added into the total. CryptoDerk 03:03, Feb 15, 2005 (UTC)
I don't think those values are meaningless, but I see your point. I wasn't too sure when I calculated them, but figured that they could be used to say meaningful things. Remove 'em if you don't like 'em—I don't feel strongly either way. Ben Cairns 03:19, 15 Feb 2005 (UTC)
Not counting the articles in the categories that haven't been completely reviewed is quite somewhat a drawback. I now work on Complex analysis, and besides me being lazy and having other things to do, and that's a huge category and merging is often not trivial. But it would be of course way too much pain to go count the articles in each category. Your stat is a good thing though, as it gives some measure of progress (even if it shows how much more work we have ): Oleg Alexandrov 04:14, 15 Feb 2005 (UTC)
Well... there's always the option of putting 'stats in progress' up for each category that we work on. The question is, do we want to go down this path? I don't think it is such a bad idea, if we also count the unclassified non-duplicates remaining so we can see where the work needs to be done. Ben Cairns 04:47, 15 Feb 2005 (UTC).
Never mind, I was too tired to think of what I was writing. But your idea with a stat in progress is nice, if one willing is put the extra effort. Oleg Alexandrov 05:13, 15 Feb 2005 (UTC)

Bah. I should bludgeon myself over the head. OF COURSE there are more articles using the templates than appear to have been processed: the apparently-processed articles are quite a bit fewer than the actually-processed articles! I was fooled by the whole problem of stats on the main project page vs. articles that have been processed but are not noted on the main page. Reasons like CryptoDerk offered (above) should account for any remaining discrepancy. Ben Cairns 05:38, 15 Feb 2005 (UTC).

As far as the number of articles we have, it is exactly 3921 (based on counting each id once). I updated this now in the statistics page.
At PlanetMath they have 4088 (which reminds me that I need to finish that update script, but I will do it after wix gets fixed so that uploads/downloads to Wikipedia are automatic). Oleg Alexandrov 06:14, 15 Feb 2005 (UTC)

Status updates[edit]

I've just added a 'status update' to 60-XX Probability theory and stochastic processes, despite there being 46 unreviewed articles (29 have been reviewed). I think that provided it's not overused in small or barely-examined categories, the benefits of this outweigh the disadvantages (such as the added clutter), but wanted to ask the rest of you what you thought. Ben Cairns 06:22, 15 Feb 2005 (UTC).

Encoding issues[edit]

Eventually all the ugly characters which did not get encoded properly in the autogenerated list will need to be replaced when an automatic update is performed (again, for now wix is broken).

In spite of looking very carefully through several pages I did not find any more than the two I listed at Wikipedia:WikiProject Mathematics/PlanetMath Exchange/encoding issues. If you run into more, just put them there. Oleg Alexandrov 04:31, 17 Feb 2005 (UTC)

Crypto, thanks and sorry. I got the wrong page for posting that. Oleg Alexandrov 05:13, 17 Feb 2005 (UTC)

Theorems, Proofs[edit]

Planetmath does one interesting thing that WP mostly doesn't: provide proofs for theorems. Of course, adding proofs would increase WP's math section ten-fold, but it seems to me that having such would be valuable in the long run (many years/decades from now). Any comments on current WP status/phiolosphy for proofs? WP:PMEX doesn't nmention this, I haven't seen this discussed elsewhere. (Unsigned from User:linas)

Actually, the main page lists as the first bullet PlanetMath frequently has an entry on "some theorem" and a separate entry on "proof of some theorem". In Wikipedia these articles should generally be combined into one entry under the title "some theorem". So, you should copy over the proof, but don't put it in a separate article. CryptoDerk 00:02, Mar 13, 2005 (UTC)
I think proofs certainly should be in the same article as the theorem. However, I would be very careful when to add proofs. This is an encyclopedia after all, not a collection of math papers. Proof should be added when they are not too hard, and when they actually add some value to the article, and not for their own sake.
And proofs must certainly be in their own section or subsection, to make it easy for people who want to skip over them. Oleg Alexandrov 00:16, 13 Mar 2005 (UTC)
I agree with Oleg that we want tnly beo be judicious in our inclusion of proofs. Paul August 01:20, Mar 13, 2005 (UTC)

OK. I suggest the creation of a new policy for Wikipedia math articles: the creation of secondary/supporting articles, which contain derivations, minor proofs, or detailed bibliographic references. Think something along the lines of "talk" pages, except that they're "proof" pages. They are out-of-line (since they're mostly boring). Thier use would not be encouraged, except in such cases where the validity of a formula may be questionable to some readers (e.g. may have been challenged in a talk page, or may be 'less than obvious' in some way).

The goal of such articles would be to improve the verifiability of the correctness of the content of articles given in wikipedia. I envision the day, many years from now, when WP is not 'just' an encyclopedia, but is a reference that is comprehensive, as comprehensive as e.g. Abramowitz & Stegun is for special functions. When it gets that big, there is the potential problem of asking whether a given formula is truly correct; the answer to which is invariably a proof. We may as well start talking about this now, and experiment with a few 'house styles' for handling this.

In the interim, how about adding a status category for PM articles: Derivation/Proof? The PM proof articles ... well, I don't really want to mark them as Needs to be copied since frankly the need is not so great, at least, not now; these can be defered. There's more important stuff now. But I do not want to mark them as Not needed in WP, since, frankly, they are needed. Here are two examples: [5] and [6]. linas 18:30, 26 Mar 2005 (UTC)

I still don't see why we should include proofs. It seems to me that the proper answer to the question "whether a given formula is truly correct" is to provide some references. I would place more trust in published, peer-reviewed proofs than on Wikipedia proofs. Furthermore, proofs are not encyclopaedic. I think that most wikipedians agree that generally, there shouldn't be proofs in Wikipedia.
That being said, I don't see any big problem with an article like Proof of Bohr-Mollerup theorem, so I won't oppose the creation of such an article. Similarly, I wouldn't mind adding the "Proof" status category. -- Jitse Niesen 21:48, 26 Mar 2005 (UTC)
I don't think Linas's vision is encyclopedical (no, I will not suggest again that Linas might be happier at PlanetMath). This is not the right place to discuss this anyway, as it is a very important topic for the whole math community here. I would suggest that Linas give a very careful thought to what he would actually like to do, and post it at the usual place, Wikipedia talk:WikiProject Mathematics. Oleg Alexandrov 22:17, 26 Mar 2005 (UTC)

Status of automated updates (it seems it works)[edit]

I finished and tested the code for automatically updating things from PlanetMath. I must say it took forever for me to finish that, both because of technical issues (like Wikipedia changing some of their protocols, etc), and due to my laziness.

But either way, I mirrored this project at User:Oleg Alexandrov/PlanetMath Exchange/, and then wrote over it the updates from PlanetMath. So, if you visit that page, go to each individual lists, and do a diff of the current version (with updates from PlanetMath) and the one before it (which is just a copy of what we have here), you will see what the changes are.

Two remarks are in order. First, on PlanetMath they sort the articles alphabetically, with the exception of the names of mathematicians, which are sorted by last name, and with the exception of article titles starting with formulas, which are sorted separately. I was not able to reproduce that, I sort everything alphabetically, so in a few places things are a bit displaced.

Second, now all the sections are sorted by section number. Before that, the order was a bit strange, and I suspect I might have inherited that from the PlanetMath structure. So in some (just a few) places, some sections are interchanged.

I looked at around 10 diffs and it seems my script does not screw anything up (except it removed some of Paul's comments which were not attached to a specific article, but rather, to certain sections; but that happened only in one section, and is I think, not crucial).

When you get to it, I would like to ask you to check some diffs too, to see if there are any problems. Then, at some point, we will need to stop the work on our project for around 15 minutes to do the updates. This does not have to happen anytime soon, as we do have still plenty of work.

Our exchange project is going well, and I (and hope you too) am happy. Oleg Alexandrov 01:14, 14 Mar 2005 (UTC)

Brilliant, Oleg! I checked 60-XX and 39-XX and they seemed to work fine. The only suggestion I would make is that if _NOTOC_ has been removed from the page, then it would save removing it again if it wasn't added in the update -- of course this is a very minor thing. Overall it seems to work very well; thank you! Ben Cairns 00:59, 15 Mar 2005 (UTC).
Ok, great, but how do I use this? I see for example, that 33-XX special functions on PM has changed. Should I copy those changes to here? If so, then how will I detect future changes? Note that I have been hacking 33-XX over here since you last updated over there. linas 18:41, 26 Mar 2005 (UTC)
I would need to copy over the new stuff from PlanetMath to my computer, then copy our project too, merge them, and upload back here. All of that is automated. One day when we feel we did a lot of work and want new stuff, I will do that. Oleg Alexandrov 18:57, 26 Mar 2005 (UTC)

PlanetMath license and reference categories?[edit]

I was wondering if it might be appropriate to add category listings to the Template:PlanetMath attribution and Template:PlanetMath reference templates, so as to identify those articles that use PM material under license or that cite a PM article. I'm not exactly sure on the suitability of such categories (although there are stub categories so why not these?) -- could someone else comment on the suitability of this? On particular concern is that it might clog up the category listing at the bottom of a page if it licenses PM text and cites an article. Perhaps there is a way to get around this? Cheers, Ben Cairns 01:07, 15 Mar 2005 (UTC).

Project covered in Quarto[edit]

I was very pleasantly surprised to learn that our project will be covered in the next edition of the Wikimedia Quarto (the official newsletter of the Wikimedia Foundatoin), which is currently being translated: see here. -- Jitse Niesen (talk) 3 July 2005 16:35 (UTC)

Good to know. Thanks. Oleg Alexandrov 3 July 2005 20:09 (UTC)
I'm the one who wrote it up, but that seems like ages ago. CryptoDerk July 3, 2005 20:25 (UTC)

Converting from PlanetMath format[edit]

I wrote a simple web-based script which takes as input the id of an article on PlanetMath and returns back the wiki-code for the article. It is not perfect, but it does save some work when copying stuff. If you try it out, let me know what you think or what can be made better. Here's the link. Oleg Alexandrov 19:22, 25 August 2005 (UTC)

OK I tried it out. I think this will be very useful, but I don't understand what is going on with the non-inline TeX, this is what I got for Id=2262 "Cycle" (this is the third version of the ouput after Oleg did his two fixes mentioned below Paul August 00:10, August 26, 2005 (UTC)):

(Start of output text: I modified the section heading level for the "See also" section)

Let S be a set. A cycle is a permutation (bijective function of a set onto itself) such that there exist distinct elements a_1, a_2,\ldots,a_k of S such that

f(a_i) = a_{i+1}\qquad \mbox{and}\qquad f(a_k)=a_1

that is

\begin{matrix} f(a_1)&=&a_{2}\\ f(a_{2})&=&a_{3}\\ &\vdots&\\ f(a_{k})&=&a_{1}\\ \end{matrix}

and f(x)=x for any other element of S.

This can also be pictured as

a_1\mapsto a_{2}\mapsto a_{3}\mapsto\cdots\mapsto a_{k}\mapsto a_{1}

and

x\mapsto x

for any other element x\in S, where \mapsto represents the action of f.

One of the basic results on symmetric groups says that any permutation can be expressed as product of disjoint cycles.

See also

This article incorporates material from cycle on PlanetMath, which is licensed under the Creative Commons Attribution/Share-Alike License. (End of output text)

Paul August 20:40, August 25, 2005 (UTC)

I've taken the above source and modified it slightly to create: Cycle (mathematics). It was very easy. Paul August 21:30, August 25, 2005 (UTC)

Hi Paul. I fixed that one now. The reason for the bug was that there are just a lot of ways to display math formulas as images, and I did not treat them all. Now that article looks better, but I don't know what to do about the eqnarray construct, it does not really have a WikiTeX equivalent I think. Let me know if you see more bugs, there should be many. Oleg Alexandrov 21:37, 25 August 2005 (UTC)
Actually, eqnarray can be split into several equations. I hope to look into that tonight. Oleg Alexandrov 21:43, 25 August 2005 (UTC)
I converted the the "eqnarray"s to "matrix"s and the element separating "\"s to "\\"s, see Cycle (mathematics). Paul August 22:08, August 25, 2005 (UTC)
Thanks, I now implemented what you suggested in the code. More bug reports welcome. Oleg Alexandrov 22:43, 25 August 2005 (UTC)
Looks great now! I will proceed to give it a workout. Paul August 00:10, August 26, 2005 (UTC)
I will now work on fixing other bugs. As soon as you find any, just list it here. Oleg Alexandrov 00:14, 26 August 2005 (UTC)

Would it be good to add the category mathematics as a default? Paul August 17:19, August 26, 2005 (UTC)

I will do that tonight. Actually, I can even add as categories what they list as AMS classification (see for example congruence). Some of those will not generate real categories, but some will. Oleg Alexandrov 17:40, 26 August 2005 (UTC)

Also maybe it would be a good idea to separate the "planetmath" template with a line by adding "----" on the line before it? See for example Cycle notation. Paul August 20:13, August 26, 2005 (UTC)

That was easy enough. I also put the Category:Mathematics for now, which will change to something more fancy later.
By the way, I think you noticed that my program creates wikilinks, which are just existing Planetmath links with the new wiki braket notation. Some of the links are wrong, obviously, as on Wikipedia we could have a certain article named differently as on PM. If at some point you fix a certain wikilink as you convert an article, I could make sure that this correction applies throughout any future conversions containing this link. For that one would need to create a conversion table, where one could add a link which one fixed anyway. So, this is just something to be kept in mind. Oleg Alexandrov 20:27, 26 August 2005 (UTC)
Cool work on that script. Regarding the line, I don't like the look of it. I never have. I also see it put on articles when there is a "This article is about so and so. For the farmer, see so and so" note at the top. Hate hate hate hate. CryptoDerk 20:23, August 26, 2005 (UTC)
I don't linke the line much either. Let us see if more opinions come in about this. I will leave it in for now. Oleg Alexandrov 20:27, 26 August 2005 (UTC)
Well I think it should be separated from the text in some way. Having it follow the last line of the text is confusing I think, especially if it is the last line of a section, because then it looks like it is part of that section. Paul August 20:29, August 26, 2005 (UTC)
By the way Oleg's script is way cool. I'm afraid he is gonna be in line for another barnstar for this — and I have none :( By the by the way, while we are on the subject of awards, and since Crypto has crept out of the woodwork so to speak, I should also say that he deserves an award (if he doesn't already have one, or even if he does) for his "Vandal Fighter" program. I've become addicted to it. It is my new favorite thing to do while watching TV. Paul August 20:46, August 26, 2005 (UTC)
Sorry for being away/inactive. I've spent the last few months winding down things and moving. I'll be back on CDVF again and I'm bringing in a few new coders to work on the interface, bug fixes, and minor improvements while I work on a plugin to CDVF that will make it twice as good as before. Maybe I'll get to a point where I don't feel guilty working on PMEX thinking I should be working on CDVF. CryptoDerk 20:49, August 26, 2005 (UTC)

Ok I've run into a coulple of more problems. PM's dependence relation uses "\prec" function which doesn't parse, perhaps it could be converted into "\le"? Converting "dynkin system" generates a bunch of cruft at the end of the article. I was able to just delete it. Paul August 02:22, August 27, 2005 (UTC)

Fixed the "dynkin system". That article, unlike others, is abandoned. :( That confused the script.
About "\prec", it can be easily changed to \le, however I will think more tomorrow if some workaround is available. (Also tomorrow hope to get to the cats thing). Oleg Alexandrov 05:33, 27 August 2005 (UTC)
PS Crypto, we definitely want you first of all catching vandals. :) I think there was more vandalism when you were off. (I feel a lot of peace in heart knowing that there is somebody out there taking care of vandalism. Today I saw somebody deleting a paragraph from real number, made me depressed, really). Oleg Alexandrov 05:33, 27 August 2005 (UTC)

I finished implementing the automatic categorization of articles, based on the AMS subject classification. See also Wikipedia_talk:WikiProject_Mathematics#New_math_categories for discussion about missing categories needed to make the correpondence between AMS subject classificaiton numbers and Wikipedia categories a well-defined function. Oleg Alexandrov 00:17, 28 August 2005 (UTC)

I replaced for now \prec with \triangleleft. This is a dirty hack, but can't think of anything else. Oleg Alexandrov 04:21, 28 August 2005 (UTC)

After the conversion to wikicode, some wikilinks are obviously wrong. For example, on Planetmath the link to "function" is for some reason called "image2.html" which is kind of silly. I am making a table listing PlanetMath links and their corresponding Wikipedia equivalents at User:Mathbot/Mapped links. If there is any link which you encounter and need to fix again and again, you could just write it there, and my script will take care of them in future occurences. Oleg Alexandrov 00:53, 29 August 2005 (UTC)

Oleg: PM articles seem to link "subset" to "inclusion", where we want "subset" (see for example 3692). I've added this to the above list. Does your program access this file dymanically? Or do you have to updated it manually? At any rate your program still produces a link to "inclusion". Paul August 04:30, September 1, 2005 (UTC)

Well, first of all my program does not access that file dynamically, only when I chose to download it (I can easily make it dynamic, see what I did at Wikipedia:Votes for deletion/Old, but then vandalism of that map of links could be of concern).

I figured out why that "inclusion" link did not work. On PlanetMath the original link is "inclusion2" (you can check that by visiting finite character, id=3692. So, that one had to be mapped (via User:Mathbot/Mapped links) to subset. I did it now, and after downloading the map it works. Oleg Alexandrov 05:03, 1 September 2005 (UTC)

Ah yes. Paul August 05:23, September 1, 2005 (UTC)

Statistics - perl[edit]

Anyone who wants to update or add statistics to a section, here's a short perl snippet I just used:

 while (<>){
        if (/Status:/) {
            if (/Needs to be copied/){ ++$a;}
            elsif (/Copied/){ ++$b;}
            elsif (/Needs to be merged/){ ++$b;}
            elsif (/Merged/){ ++$d;}
            elsif (/WP article adequate/){ ++$e;}
            elsif (/WP article more complete/){ ++$f;}
            elsif (/Not needed on WP/){ ++$g;}
            else {++$h;}
                ++$i;   
         }
 }
 $j=$i-$h;
 
 print "::Total articles: $i\n";
 print "::Reviewed articles: $j\n";
 print "::Copied: $b\n";
 print "::Merged: $d\n";
 print "::WP article adequate: $e\n";
 print "::WP article more complete: $f\n";
 print "::Not needed on WP: $g\n";
 /**/print "::'''Needs to be merged''': $c\n";/**/
 /**/print "::'''Needs to be copied''': $a\n";/**/
 /**/print "::'''Needs to be reviewed''': $h\n";/**/
 /**/print ":::~~~~\n";/**/

This follows Paul August's/Oleg's neat layout. Someone clever could build this into an automatic updater, if they like. Rich Farmbrough 15:41, 29 August 2005 (UTC)

The neat layout is definitely due to Paul. There will come a day, hopefully in a week or two, when I will set up mathbot to periodically add new info from PlanetMath as well as keep the stats up to date. Oleg Alexandrov 19:12, 29 August 2005 (UTC)
I think it takes up too much space, so I tried to replace the first few entries by a table. I think the table looks cleaner, but opinions may differ. I removed the signatures since they are not very useful IMO. -- Jitse Niesen (talk) 22:18, 30 August 2005 (UTC)
The table is cool. Packs the same info in less space. Oleg Alexandrov 23:23, 30 August 2005 (UTC)
I think the sigs are useful, it tells you who last updated the data and more importantly when. Paul August 00:11, August 31, 2005 (UTC)
Oh any I forgot to say that the table is a much better way of presenting the information (especially now that it has sigs, and there mostly mine, thanks Jitse ;-) — Paul August 17:53, August 31, 2005 (UTC)
Also I'm thinking about some tweaks to the table. Maybe make the columns narrower, using more abbreviated column headings? I wish now that we had chosen shorter status messages (e.g: OK, Better, Not needed, Copied, Merged, Needs copying, Needs merging) These would be easier to type, and less error prone, and typos make automated counting harder. Is it worth changing now? Also it might be nice to display percentage reviewed, and percentage processed? Paul August 18:17, August 31, 2005 (UTC)
Be bold, especially when hacking my edits to pieces :) Percentages are nice, but somebody has to compute them. I thought about some colour coding to indicate when all articles are reviewed, or all are processed. -- Jitse Niesen (talk) 11:46, 1 September 2005 (UTC)
Nice work on the table Jitse!. I think color coding would be great. I'm one of those silly people who are motivated by things like that. Thats why I think percentages might be nice, besides, I just figured one of you code wizards would simply wave your magic wand and conjure up some bot pixie to do all the work ;-) Paul August 13:30, September 1, 2005 (UTC)
Table is definitely better! (need to rerite the perl now....) Rich Farmbrough 20:28, 1 September 2005 (UTC)

Version 2[edit]


while (<>){
          if (/Wikipedia:WikiProject Mathematics\/PlanetMath Exchange\/(.*)/){
                print "|[[Wikipedia:WikiProject Mathematics/PlanetMath Exchange/$1|$1]]\n";
          }
        if (/Status:/) {
            if (/Needs to be copied/){ ++$a;}
            elsif (/Copied/){ ++$b;}
            elsif (/Needs to be merged/){ ++$b;}
            elsif (/Merged/){ ++$d;}
            elsif (/WP article adequate/){ ++$e;}
            elsif (/WP article more complete/){ ++$f;}
            elsif (/Not needed on WP/){ ++$g;}
                else {++$h;}
                ++$i;   
         }
}
$j=$i-$h;
$done=$b+$d+$e+$f+$g;

$red=sprintf ("%2.X", 255*($i-$done)/$i );
$green=sprintf ("%2.X", 255*$done/$i );
$red =~ tr/ /0/;
$green =~ tr/ /0/;

$istyle= "! style=\"background:#".$red.$green."00\" ";

print "$istyle | $i || $j || $b || $d || $e || $f || $g || $c || $a\n";
print "|~~~~\n";

OK this does the job but the colour scheme is not great. Feed it the page (cut and paste to a text file) and it will throw out the table entry, colouring it between green and red according to the level of completion. Completeion counts as merged, copied, not needed, more complete or adequate. It may crsh on Education which has no articles (div 0)! Rich Farmbrough 00:36, 2 September 2005 (UTC)

How about this color scheme?
Grey Not started
Yellow Started
Tan All reviewed
Green Complete

Paul August 03:26, September 2, 2005 (UTC)

Nice! Oleg Alexandrov 03:35, 2 September 2005 (UTC)
Pretty, I'll look at the code tonight. Rich Farmbrough 09:48, 2 September 2005 (UTC)

Version 3[edit]


while (<>){
          if (/Wikipedia:WikiProject Mathematics\/PlanetMath Exchange\/(.*)/){
                $title= "|[[Wikipedia:WikiProject Mathematics/PlanetMath Exchange/$1|$1]]\n";
          }
        if (/Status:/) {
            if (/Needs to be copied/){ ++$a;}
            elsif (/Copied/){ ++$b;}
            elsif (/Needs to be merged/){ ++$c;}
            elsif (/Merged/){ ++$d;}
            elsif (/WP article adequate/){ ++$e;}
            elsif (/WP article more complete/){ ++$f;}
            elsif (/Not needed on WP/){ ++$g;}
                else {++$h;} # not reviewed
                ++$i;   
         }
}
$j=$i-$h;
$done=$b+$d+$e+$f+$g;

if ($h==$i){ #not started
        $istyle= "|- \n";
}
elsif ($h+$b+$a==0){ #completed FFFFCC
        $istyle= "|- bgcolor=#FFFFCC \n";
}
elsif ($h==0){ #all reviewed E6E6AA
        $istyle= "|- bgcolor=#E6E6AA \n";
}
else { #Started FFFFCC
        $istyle= "|- bgcolor=#FFFFCC \n";
}

print $istyle;
print $title;
print "| $i || $j || $b || $d || $e || $f || $g || $a || $c\n";
print "|~~~~\n";

This seems to work.... Couple of bugs removed. Rich Farmbrough 11:55, 2 September 2005 (UTC)

Rich, I have a challenge to you. If you (1) like Perl programming (2) know how to install Perl modules, then you could write a script to periodically fetch data from our project, compute the statistics above, and spit out the data in the table. What do you think? Oleg Alexandrov 15:46, 2 September 2005 (UTC)
For further joy, you could try to make your script understand Paul. For instance, Wikipedia:WikiProject Mathematics/PlanetMath Exchange/08-XX General algebraic systems has 3 articles marked as more complete, but the table lists only one. -- Jitse Niesen (talk) 16:02, 2 September 2005 (UTC)
Lol. Yes I have a habit of doing that! That's what I was referring to when I wrote above that I wish we (meaning me) had chosen simpler status messages. I thought about changing those just now, but I decided to see if Rich is up to the challenge ;-) Paul August 16:50, September 2, 2005 (UTC)
Now I will start writing some of that Perl noisy syntax which Jitse hates. Rich's script needs to be a bit more robust. For example, if the status has the words "needs" and "copied", then that should be a good indication that the article needs to be copied. Deal with all the other cases simiarly. If the status does not match any of the options and is not empty either, record that as a problem to be looked into (most likely a misspelling somewhere). How about that:
 
       if (/Status:([^\n]*)/) { # after "Status:" and all the way up to a newline (after the newline the comments start)
            $status=$1; # the string following the words "status"    
            if ($status =~ /need/i &&  $status =~ /copied/i ){ ++$a;} # needs copying I guess
            elsif ($status =~ /copied/i ){ ++$b;} # copied already
            elsif ($status =~ /need/i && $status =~ /merged/i) { ++$c;} # needs merging
            elsif ($status =~ /merged/i ){ ++$d;} # merged already
            elsif ($status =~ /adequate/i ){ ++$e;} # wp is good enough
            elsif ($status =~ /complete/i ){ ++$f;} # wp more complete
            elsif ($status =~ /not/i && $status =~ /needed/i ){ ++$g;} # not needed
            elsif ($status =~ /^\s*$/) {++$h;} # empty status, that is not reviewd
            else {print "Huston, we have a problem. The status line is incomprehensible\n"};
  
                ++$i;   

Oleg Alexandrov 18:24, 2 September 2005 (UTC)

If and when I re-run, I will make the code more forgiving, unless the new status sybols below mean I don't have to. I will also look at the perl module to automayrte, but once I got up to speed I could stat a page in about 25 seconds, so unless it becomes a regular task it may not be worthwhile. Rich Farmbrough 14:47, 10 September 2005 (UTC)

Questions about adding new info from PlanetMath[edit]

Question one. Is it a good time to start adding new articles information from PlanetMath, or is it premature since there is still a lot of work? Just wondering.

If the answer to this question is yes (and even if it is no, still useful for future reference), another question I want to ask is whether it is worth preserving comments made by us at the beginning of sections, as opposed to comments to individual articles (see 03-XX Mathematical logic and foundations for an example). It looks to me that this kind of information would become partially obsolete when more articles are added.

Or maybe the right thing to do is to actually have this kind of information at each section, and which would be updated automatically. Would such per section statistics be desirable?

A third question is if links saying

Duplicate entry. See so and so...

which we have are useful or not. For me personally they looked more useful at the beginning, but don't look so anymore.

Answers appreciated. :) Oleg Alexandrov 01:24, 4 September 2005 (UTC)

Here are my thoughts. Yes it is a bit too soon to update. As far as the section stats, I started using them in 03-XX Mathematical logic and foundations, because it was so long (396 entries) with so many sections that I felt like I needed some way to keep track of where I was, and what was left to do. If they could be automatically updated that would be great. If not just leaving them there to be updated manually would be fine too. As Regards the Duplicate's I've never found them of any use. Paul August 03:52, September 4, 2005 (UTC)
1) Don't care; 2) Leave the comments; 3) Not useful. -- Jitse Niesen (talk) 11:18, 4 September 2005 (UTC)

OK. One day I will remove the Duplicate entries (not having to keep track of things across multile sections will simplify my code a lot). Also, unless Rich gets there before me, I plan to add statistics to each section and periodically update the main statistics table.

Paul, you mentioned about possibly shorter names for the status line. Do you have in mind anything in particular? Oleg Alexandrov 03:42, 5 September 2005 (UTC)

How about something like: OK, Better, Not needed, Copied, Merged, Needs copying, Needs merging? And perhaps we don't need to make a distinction between OK and Better? Just have OK? Paul August 04:25, September 5, 2005 (UTC)
Well, if we replace Wikipedia article adequate with simply OK, we could as well replace Needs copying with NC, and same for others, to get the same abbreviations as in the main page stats table. Opinions? Oleg Alexandrov 19:40, 5 September 2005 (UTC)
I had thought that there was some value in having the status fields be self-explanatory. Now I think that just using the abbreviations might be the best thing after all. And while I still think the distinction between merged and copied has value, I no longer think the distinction between adequate and more complete is all that useful. Paul August 00:24, September 7, 2005 (UTC)

Sounds good. Sometimes this week (or weekend) I will replace the old status lines with the new ones. Oleg Alexandrov 03:19, 7 September 2005 (UTC)

Ok then if we are going to go with abbreviations I would propose the following:
Status means PM article
N not needed
A adequately covered
C copied
M merged
NC needs copying
NM needs merging

I changed the status line and removed the duplicates. I checked all the diffs (I hope!) and seems there were no problems with my code. Can't say for sure since in some places the diff is a sea of red (removing the duplicates confused the diff). Some edits might have not gotten through as the server was behaving badly tonight, so I will do another pass tomorrow evening.

Oh, and I screed up the edit summary, it refers to Wikipedia talk:WikiProject Mathematics instead of Wikipedia talk:WikiProject Mathematics/PlanetMath Exchange. Oleg Alexandrov 03:49, 8 September 2005 (UTC)

Nice work Oleg. I will update the table on the project page to reflect the elimination of the "more complete" status, adding the numbers under "Bt" to "A", and eliminating that column, unless you prefer to perform some perl magic there too? I also think that the order of the status messages should follow the order in the "status table" just above here, i.e. N, A, C, M, NC, NM. Paul August 13:54, September 8, 2005 (UTC)
Here I think one could use a spreadsheet rather than Perl magic. :) I will do it tonight. Oleg Alexandrov 15:21, 8 September 2005 (UTC)
(That is, unless you get there before me. Oleg Alexandrov 15:24, 8 September 2005 (UTC) )
Ok I did it the old fashioned way by hand ;-) Paul August 16:21, September 8, 2005 (UTC)
Thanks. I should have done it yesterday, but it was getting rather late in here. Oleg Alexandrov 17:04, 8 September 2005 (UTC)

PMEX Proofs?[edit]

(Copied from my talk page by Oleg Alexandrov at 18:10, 5 September 2005 (UTC))

Could I talk you into adding a WP:PMEX status of Proof, with the understanding that its not needed on WP? I would be curious to know what fraction of NN are proofs. linas 17:45, 5 September 2005 (UTC)

Any opinons about that? I myself don't like proofs on Wikipedia, so I would be tempted to support such a Proof=nocopy status. On the other hand, puttting such a no-exceptions restriction could look a bit too much. Oleg Alexandrov 18:10, 5 September 2005 (UTC)
Or maybe I misunderstood what you want. You want the status line to have the word Proof. But the status line is not meant for such a thing. :) OK, if you want me to write a script to tell how many proofs are there, I am rather lazy to do that for me moment, as I don't see what that would gain. Oleg Alexandrov 18:13, 5 September 2005 (UTC)
I'm pretty sure that Linas wants the status line to have the word Proof.
I think we need only three statuses: Not reviewed, reviewed but needs more work (NM and NC), and done. If we decide that we need more information, like we have now, then I'd support splitting NN (not needed) in proofs, dictdefs and nn (non-notable), which is incidentally what I've done in "my" section (15-XX Linear algebra; too lazy to link). -- Jitse Niesen (talk) 18:39, 5 September 2005 (UTC)
I think some proofs are a valuable addition to wikipedia, specially those that shed light onto concepts instead of being a chain of calculations and logic deductions. For instance, take Pascal's rule. The combinatorial proof is intuititve and it could be nicely ilustrated with a few pictures (I'm guilty of not having done them yet ;) whereas the calculational approach has little added value. I saw he same situation on Rouche's theorem where the formal proof is uninteligible to someone who is or has not taken a complex analysis course, whereas I feel the geometric sketch tells much more about what's happening, even at the cost of leaving some gaps. So I don't think that proofs have no place in wikipedia just for being proofs. -- (☺drini♫|) 18:51, 5 September 2005 (UTC)
I'm glad you mentioned Pascal's rule, since I did not know about this article, and there were no links to it from either binomial coefficient or Pascal's triangle. Now fixed. -- Jitse Niesen (talk) 23:17, 6 September 2005 (UTC)

I think we all agree that some proofs should be be incorporated into Wikipedia, since they can have explanatory value. And I think there has so far been a consensus for not having very many proofs, since they will often get in the way of exposition. However I think Linus makes an important point (above: #Theorems, Proofs and at Wikipedia:WikiProject Mathematics/Proofs), that proofs (along with citations and references) can help with verifiability — and verifiability will be the single most important determinant of whether or not Wikipedia is ever considered to be an authoritative reference work. I don't see why though, we can't have our cake and eat it too ;-) Couldn't we, for example just include any proofs, which are not suitable for expository reasons, on the article's talk page? But, this is a discussion for a wider audience, for the purposes of this project, I think I see some value in marking in some way PM articles which contain proofs which have not been copied to WP. Paul August 22:23, September 6, 2005 (UTC)

I changed the status of all "my" proofs from "Not needed on WP" to "Not needed on WP: Proof". This should not break any script, and we can easily adapt any script to count the proofs separately (or not). -- Jitse Niesen (talk) 23:17, 6 September 2005 (UTC)

Hmm. Didn't mean to rekindle the debate of whether proofs are needed or not. Since reviewing is time-consuming, I thought it would be good to put a ticky-mark on an article to indicate its a proof, so it can be found later, if desired. However, I now realize that the majority of not-needed articles on WP:PMEX are probably going to be proofs, so its not really going to be hard to locate these. linas 23:48, 6 September 2005 (UTC)

Problem with PlanetMath site?[edit]

Is any one else experiencing problems viewing the images in PlanetMath articles? For example when I try to download this page: axiom of power set: My browser (Safari) reports errors of "can't connect to host", for each png on the page. I had to resort to using Oleg's Pmform tool just to be able to "see" the images. Paul August 16:55, 14 September 2005 (UTC)

I did some PlanetMath articles reviewing for our project these days and noticed the same thing. Wonder what is going on. Oleg Alexandrov 18:04, 14 September 2005 (UTC)
If you get on IRC, you can bug Aaron. He's "akrowne" on freenode. CryptoDerk 19:30, 14 September 2005 (UTC)
Nevermind. He's in Italy (even though he's still on IRC). CryptoDerk 19:39, 14 September 2005 (UTC)
For me, the whole site is non-responsive at certain times. I get no reply at all. It has been like that on and off all week. linas 00:55, 15 September 2005 (UTC)
I guess some asteroid hit this planet. Or maybe it is one of those bodies beyond Pluto which this week got too far from the sun. Oleg Alexandrov 01:25, 15 September 2005 (UTC)

Automatic stats[edit]

I am getting closer to updating automatically the stats for the project. Today I did two edits. The first, introduced a newline between rows and inserted the silly comment <!-- begin row -->. That is necessary, not because my bot is stupid, but rather because otherwise the wiki diffs become a sea of red. The comments above forces the wiki diff to compare things row by row.

The second edit actually updated the stats (semiautomatically for now). One side effect here is that the obnoxious mathbot replaced all the other contributors' signatures by its own. I could make the code not replace the signature in rows where no change occured to start with, but for now I am not sure it is worth bothering. The formatted table now looks a bit worse stylistically (you see some XX in places). I will trim them sometime. And eventually I plan to place a link under the table, which will refresh the stats on demand. Oleg Alexandrov 07:56, 29 September 2005 (UTC)

Excellent! The refresh button also sounds like a very good idea. My only quibble though is that the totals at the bottom were not changed. This seems like the sort of thing that the mathbot would be very good at. That seems much more important than the signatures issue (though maybe if the bot is replacing them all the time, we just shouldn't have a signature column). Jtwdog 16:45, 29 September 2005 (UTC)
The Automatic stats are great! As for Mathbot, he may not be stupid (that remains to be seen) but he is certainly an egotistical, glory-hogging, credit-stealing, upstaging little so-and-so! But I guess if he is going to do all the work, he should get all the credit. I hope he knows how to count! Perhaps if we are going to use mathbot exclusively to update the table, we could bring him down a notch or two, by dispensing with that column all together? Or perhaps we will need it for the occasional human intervention? Oh and having a link under the table to make Mathbot count on command sounds great! We should label it: "Go boy! Count!". I am having fun now imagining repeatedly commanding him to count over and over and over again until his little brain melts ;-) Oh yeah and tell Mathbot to lose the "XX"s. Paul August 16:51, 29 September 2005 (UTC)
Of course, the bot is just as stupid as its boss. To satisfy the egos of the editors of human flesh, perhaps the bot could count the signatures on all the subpages and summarize the results here, like "Alice reviewed 341 articles, Bob reviewed 20 articles and copied 3, ..." Oh, it's great to tell others what to do :) -- Jitse Niesen (talk) 17:11, 29 September 2005 (UTC)
Who the hell are Alice and Bob? I sure don't remember them reviewing any articles. Paul August 17:42, 29 September 2005 (UTC)
I had several good laughs over the above. :)
After some delay, the bot finally works, and the link for refreshing the table is in place. It now also computes the totals and the percentages. I would agree with removing the signatures altogether, they just change all the time and show up in red, distracting from the actual changes. Oleg Alexandrov 04:40, 3 October 2005 (UTC)
Great work Oleg! Some nits. Could Mathbot rather than entering a zero in the table leave the cell blank? This makes it easier to "read" the table, and for the NM, and NC columns, see more easily what has to be done. Also I think Mathbot has miscolored "97 Mathematics education". As regards the signatures, (show up in red?) , we could reinterpret the meaning of the last column to be, by whom and when, the last change was made to that particular category? I also think that JItse's Idea above would be nice if it were relatively easy. Recognizing people's work is always a good thing to do. Paul August 17:26, 3 October 2005 (UTC)

Fixing the zero and the miscoloring was easy enough. The other two things would require parsing signatures and comparing dates. Doable but not immediate. I will do it sometime before the end of the week. Other suggestions are very welcome; anything which can be automated must be automated. Oleg Alexandrov 18:22, 3 October 2005 (UTC)

Thanks Oleg. There are still some pesky zeros in the "Rv" column.
Done. There is still a zero at totals, but I think needs to be there, I mean if the space is empty for the total number of articles it might look strange, no? Also, for now I removed the timestamp from mathbot's signature, it just changes all the time. Oleg Alexandrov 19:29, 3 October 2005 (UTC)
I added a list of people who contributed. Feeback on formatting this needed. Mathbot does not yet add this automatically -- to be worked on. Oleg Alexandrov 03:02, 4 October 2005 (UTC)
I think the contributions list is good. I was surprised to see my name is at the top ;-) However, splitting up "contributions" by say "reviewed", "copied" and "merged", might reveal a truer picture of everyones contributions. (For example, I have confined my efforts mostly to reviewing.) At any rate, it is also quite gratifying to see my suspicions confirmed — "Alice" and "Bob" show up nowhere in the list! Paul August 17:31, 4 October 2005 (UTC)

I will refine my script then, sometime :) Oleg Alexandrov 18:39, 4 October 2005 (UTC)

Just a remark. If you start mathbot to update the stats, please check after it. Sometimes it truncates the page when submitting. I truly doubt the issue is with the bot, rather some server thing. Oleg Alexandrov 22:19, 4 October 2005 (UTC)

OK Mathbot does not seem to want to obey me. I order him to count, but he just seems to sit there pretending to be asleep (with my browser saying "contacting www.math.ucla.edu"). Do I need to give him a kick in the butt or something? Paul August 03:40, 5 October 2005 (UTC)
Can you try again? Maybe the UCLA server was down or something. I clicked on the link now (and I did it being outside the the UCLA network), and mathbot is ticking along nicely. Oleg Alexandrov 15:17, 5 October 2005 (UTC)
Actually, it is likely there was a bug in the script. After getting rid of zeros in the stats table (as per Paul) I was having poor mathbot try to add up empty strings with numbers. Surely confused my bot. Now it works well it seems. Oleg Alexandrov 17:53, 5 October 2005 (UTC)

PMEX and GFDL[edit]

There's a thread right now at Planetmath discussing the content exchange between the sites, and it has been raised the point that wikipedia may not be holding GFDL as well as PM does. So I would like some feedback. The thread is at PM's embedding entry. Basically, jac comments:

Copying or copy-and-modifying from Wikipedia is NOT fine in my view, since until the site admins (for both sites, actually) take action, such copying is not going to be GFDL compliant.

I asked for comments got the reply

Simply put: I don't think Wikipedia is conforming to the GFDL's requirements for acknowledging GFDL'ed source materials, and I also think that they have set up their site in such a way that it is infeasible to conform to the GFDL when using their materials.

and a further analysis of most itens on the GFDL and his comments on each. At the botom he also states that

In my view, none of the articles can be copied (using them as references is of course OK, the same way you would use non-FDL'ed works) until further work has been done by the admins of both sites. When the appropriate steps WRT licensing have been taken, then all of the wikipedia articles should be usable.

so, I wanted more eyeballs on the issue to see what can be done from each side if indeed there's an issue (for instance I think WP articles taken from PM do not cite at least (or fewer if less) 5 authors from the original PM entries. Comments, anyone?

Well, it seems that nobody here knows (or cares about it), so I asked for guidance at Wikipedia:Village pump (miscellaneous). -- Jitse Niesen (talk) 21:29, 6 October 2005 (UTC)


Planet math links[edit]

The planet math reference template contains a trailing period that I would like to remove. Please add your opinion here. I am aware that this affects many pages. -- Tob 09:30, 12 October 2005 (UTC)

Why does Mathbot update so often?[edit]

I think the title says it all. Is that actually people hitting the update link, or are they scheduled? If they are scheduled, they happen way more often than they need to. 5 times a day is way excessive and I find it hard to believe that that is people hitting the link. do 5 different people actually look at the project page every day? I just find it odd that Mathbot updates so often when no one other than me is reviewing articles. not to mention it does make something of a mess of the history page. Of course, this isn't a big deal, but it seems like the sort of thing that should be easy to fix. Jtwdog 23:52, 21 October 2005 (UTC)

The mathbot updates as often as one clicks on the link. I can remove the link, but then you will be at the mercy of my whims, I will update the list when I feel like it (and ask for a wikibribe before satisfying any such request).
I guess you are suggesting that mathbot should refuse to update the table if nothing changed. That is doable, but I am not sure it is worth the trouble (I would need to write a function to do a comparision of "before" and "after".) An easier way is to just remove the datestamp of mathbot from the "last updated" lines. Then even if it attempts to update the tables, nothing will show up in the history if no actual changes happen. Oleg Alexandrov (talk) 03:26, 22 October 2005 (UTC)
I can implement a log which will post what IP address and web domain is responsible for each click, and make that log public. Is that going to make you feel a bit better? :) Oleg Alexandrov (talk) 03:28, 22 October 2005 (UTC)
Just charge everybody who clicks $1, to be paid to the Wikimedia Foundation. :)
I have also wondered about the number of updates, and I am very curious who is/are clicking all the time. Perhaps you can just have the bot run with a fixed frequency (once a day should be plenty). Another possibility is to put the stats on a separate page which is transcluded in this page; then the update wouldn't pollute the history. -- Jitse Niesen (talk) 13:25, 22 October 2005 (UTC)
While I would be highly curious about the contents of such a log, the problem is not of sufficient importance to justify it, methinks. I do like the suggestion of putting the stats on a separate page. This would make the history more readable, and discourage mischievious clickers. Jtwdog 18:42, 22 October 2005 (UTC)

Concerning the potential abuse of poor Mathbot, although I did write above that I was "having fun now imagining repeatedly commanding him to count over and over and over again until his little brain melts", I would like to assure everybody that it is not me! I wouldn't want to be reported to The American Society for the Prevention of Cruelty to Bots. Paul August 18:18, 22 October 2005 (UTC)

OK, I will put the stats on a separate page which is transcluded on this page. Oleg Alexandrov (talk) 04:41, 23 October 2005 (UTC)

Done now, we have Wikipedia:WikiProject Mathematics/PlanetMath Exchange/Table of topics and Wikipedia:WikiProject Mathematics/PlanetMath Exchange/Table of credits which show up on the main page via templates. One question. Both are big tables, and actually I wondered about it before, should they really show up on the main page, or is a link enough? Of course, then we would need one more click to see them, but people who are really interested in those tables would put them on the watchlist anyway to see when the bot updates them. Just a thought, wonder what people think. Oleg Alexandrov (talk) 12:14, 23 October 2005 (UTC)

I don't really think that's necessary. It's not as though there's much people need to scroll past the tables to find, and it's kind of nice to have them right there. My vote is for keeping them on the front page. Visible progress keeps people interested. Jtwdog 17:39, 24 October 2005 (UTC)

So, are proofs needed or not?[edit]

If not, I could have the bot put the status of all proof articles as N, meaning not needed. This would overwrite the status field when the status field is empty, and leave it the way it is otherwise. Oleg Alexandrov (talk) 08:40, 22 October 2005 (UTC)

I don't like that; I wish to keep a human element here. For instance, proof of Schur decomposition is a proof article, but I still decided that the proof should be included in Wikipedia because it is a constructive proof, and the construction is actually important in numerical algebra when one wants to compute a Schur decomposition. -- Jitse Niesen (talk) 13:25, 22 October 2005 (UTC)
I agree with Jitse. Paul August 17:56, 22 October 2005 (UTC)
Yes, good points. But putting the tag N on a proof could be a good initial guess, and may give us a better idea of how many articles are left to copy. Oleg Alexandrov (talk) 19:08, 20 November 2005 (UTC)

I found the culprit[edit]

We were wondering who is clicking so often on the mathbot link having it update the statistics. Well, I started keeping a log, at Wikipedia:WikiProject Mathematics/PlanetMath Exchange/Log. Except for a couple of clicks from Berkely last night, most are clicks from the google page crawler and an Inktomi search bot. Interesting. Oleg Alexandrov (talk) 23:02, 26 October 2005 (UTC)

This is as amusing as it must be frustrating. When I first checked the logs for my professional web page a few years back, I quickly noticed that the VAST majority of hits were from bots, particularly my own university's search bot! Eventually, I had to add a dummy link, invisible to graphical browsers, so that I could establish a rough baseline. No doubt you are able to put in place a more elegant solution than mine! Ben Cairns 10:55, 27 October 2005 (UTC)
A more elegant solution ... who cares? :) A bot calculates the statistics, other bots prod it to do work, and all left for humans is do their little part, reviewing PlanetMath articles. :) Oleg Alexandrov (talk) 12:39, 27 October 2005 (UTC)
Well, what could be more elegant than doing nothing? :) I was under the (false) impression that the 'culprits' had done a bad thing. These bots are pretty useful little triggers for things that need occasional updating. I wonder what else they could be used for... Ben Cairns 14:35, 27 October 2005 (UTC)
Yes, I would also like to know what else to automate. :) Oleg Alexandrov (talk) 14:54, 27 October 2005 (UTC)
Redirects? It is pretty silly, for example, that "Mathieu groups" doesn't redirect to "Mathieu group." This is a somewhat delicate process, but it could probably be automated to some degree. Jtwdog 18:56, 27 October 2005 (UTC)
That would be very complicated. We are talking about tens of thousands of redirects and without any background information about the articticles to redirect to, one can also get it wrong. About the "Mathieu groups" though, the Wikipedia convention is that in absolute majority of cases one should not link to plural anyway. Oleg Alexandrov (talk) 02:23, 28 October 2005 (UTC)

Missing table entries[edit]

Some of the categories (those between 31 and 45, in particular) seem to have dissappeared from the table of topics, and have been absent over several edits. What gives? Jtwdog 03:26, 8 November 2005 (UTC)

Fixed. Recently I also modified the way my bot fetches and updates the table of topics, so I think it will get likely to be corrupted.
However, if that happens in the future, you can just revert to an old working version and click on the link to rerun the bot. Oleg Alexandrov (talk) 02:37, 27 November 2005 (UTC)

Geometry proofs[edit]

I wonder if they are necessary or not. The current list of them are

Apollonius theorem (x2) Brahmagupta's formula* butterfly theorem double angle identity Erdős-Anning Theorem Heron's formula* Mollweide's equations parallelogram law Ptolemy's inequality Ptolemy's theorem Pythagorean theorem (x3, one is included) Simson's line Stewart's theorem tangents law triangle theorems bisectors theorem (x2) Ceva's theorem (and trigonomic version) Menelaus' theorem Pappus's hexagon theorem Pascal's mystic hexagram Desargues' theorem Morley's theorem pivot theorem Hadwiger-Finsler inequality Euler line

The number indicates that more than one article are present for that formula, and * means somebody has included them.

Should I just mark all of them as N, or link the proofs to their respective pages? Jeekc 10:41, 4 January 2006 (UTC)

Welcome to the project Jeeck, and thanks for contributing. Opinion is somewhat divided on whether or when to have proofs (see Wikipedia:WikiProject Mathematics/Proofs). Our current practice is to include some proofs where appropriate. So the answer really is just try to use your best judgment. By the way you may want to add yourself to these "participants lists": Wikipedia:WikiProject Mathematics/Participants and Wikipedia:WikiProject Mathematics/PlanetMath Exchange#Participants. the preceding unsigned comment is by Paul August (talk • contribs)
I am weary of proofs. If you include them, it is good if they are short, or elementary, or truly relevant, or go to the bottom, or just not adding the proof to start with. :) Oleg Alexandrov (talk) 21:33, 4 January 2006 (UTC)
My opinion is that you should at least put an external link to PM proofs in the WP articles if you choose not to include them in WP AdamSmithee 08:12, 5 January 2006 (UTC)
Yes this seems like a good idea. Paul August 15:47, 5 January 2006 (UTC)

Eep, an image![edit]

There's an image on http://planetmath.org/?method=l2h&from=objects&id=968&op=getobj which I don't know how to copy, since I don't think it falls under fair use. Since it IS PS of some sort (*.eps) isn't it possible to copy that source, then have the typesetter reconstruct it over here, or am I just going to have to omit it? - CorbinSimpson 06:32, 21 January 2006 (UTC)

The whole of PlanetMath, including the pictures (as far as I can see), falls under the GFDL, like Wikipedia. So I think you can just copy the PNG image to either the English Wikipedia or the Commons and tag it as GFDL, including the URL where you got it from. I assume you only have problems with the legal issues; if your problems concern the technical process, please let me know and I'll copy it for you. -- Jitse Niesen (talk) 13:20, 21 January 2006 (UTC)

Fetching new articles and proofs, revisited[edit]

Well, this project is one year old now, and a lot of work has been done. We have 3923 articles listed in here, while on PlanetMath now they have 4797 articles over there. How about copying over the links to the newer entries?

And on an unrelated topic, I would like to revisit the subject of proofs. The general sentiment, per a section somewhere above, seems to be that we should not disregard proofs by default. However, the truth is also that very few proofs have been copied so far, few proofs will be copied in total, and proofs, in general, are a lower priority. I would suggest marking proofs with a new flag, say P, and count those "reviewed", rather than "unreviewed", or at least, not counting them at all when doing the totals in the table. I would argue that this will reflect better our progress so far. Comments? Oleg Alexandrov (talk) 04:44, 15 February 2006 (UTC)

Well, IMHO to mark all proofs and count them as reviewed would practically be equivalent to marking them as N and forget about them. I would keep human judgement in the process.
On bringing links to new articles, this would obviously be a good idea. I don't know how you guys did it before, but if you did it manually let me know how I can help AdamSmithee 20:47, 18 February 2006 (UTC)
Heh, the human judgement on proofs. [7] Mathbot 19:35, 20 February 2006 (UTC)
I'm pretty sure it will work out fine. And if human judgement fails, I'm sure that democracy will do the job for proofs :-) AdamSmithee 20:54, 20 February 2006 (UTC)
I added new articles from PlanetMath. If you go through the lists, the new articles will be marked with new. Don't know if that's the best thing, and I can change it if there are other ideas.
So, we have now 4810 articles, up from 3923, and we are 35% done, down from 42%. Not too bad. :) Oleg Alexandrov (talk) 02:04, 6 March 2006 (UTC)
Great work Oleg. I like the new tag. And I agree with AdamSmithee, that we should continue to handle proofs as we have. Paul August 03:02, 6 March 2006 (UTC)

Might be worth another crawl; I noticed several articles at PM that weren't listed here. (unless I got confused by double listings??) linas 01:45, 8 June 2006 (UTC)

Not just yet — let's enjoy having reached 40% reviewed a little while ;-) (see this) Paul August 02:10, 8 June 2006 (UTC)
Seems that PM is alive and healthy, content-wise. They currently have 5664 articles, which is (5664-4810)/4810=18% growth in 9 months or 24% annual growth. Not too shabby. linas 04:35, 27 November 2006 (UTC)

Hello From AsteroidMeta[edit]

Dear Human Entities, I am jcorneli, lately of AsteroidMeta, the Fortress of Solitude of the PlanetMath multiverse. We would be interested in chatting with you about the relationship of PlanetMath and Wikipedia. If you wish, that conversation might unfold from http://planetx.cc.vt.edu/AsteroidMeta/Wikipedia. Otherwise I'll watch this page. --Joe Corneli

In our part of the multiverse, PlanetMath actually means we copy math articles from over there. Superman's services from the above fortress would be much appreciated, if that guy knows some math and how to type TeX. Oleg Alexandrov (talk) 14:54, 23 March 2006 (UTC)

Probably the most likely help you'll find from AsteroidMeta is that that's where we make sure that PlanetMath continues to exist, grow, and thrive. As to how we'll know what these things mean, it depends largely on what the PM user community and other correspondents have to say about it. Looking at PlanetMath as a source of articles is a good thing... but I'm also interested in exploring other possible collaborations between Wikipedia and PM. For example, code, finances, lobbying -- as the need may arise. In the long term, I'm anticipating that PlanetMath will just be one of several "Planets" devoted to technical content; getting some thoughts from Wikipedia people about how to approach this sort of growth would be a good thing. My initial thought is that PM and subsequent Planets could become the "micropedia" to Wikipedia's "macropedia", but that's a very abstract idea. --Joe Corneli

For things like "code, finances, lobbying", you'd need to contact the wikimedia foundation, the people who run wikipedia. We here are just powerless peons who read WP and PM articles. linas 04:59, 6 June 2006 (UTC)

Delete "Object not found!" entries?[edit]

I've been running across several PM entries for which clicking on the link gives the message "Object not found!" e.g. the following entry from the general topology list:

Should we just delete these entries when we find them? Paul August 20:58, 3 September 2006 (UTC)

Sorry for the late reply. I was away during the first half of september and could not check my watchlist.
Now, the above line is found in the section Wikipedia:WikiProject Mathematics/PlanetMath Exchange/54-XX General topology#54-00 General reference works (handbooks, dictionaries, bibliographies, etc.) of the bot-generated lists. The original planetmath it was generated from is http://planetmath.org/browse/objects/54-00/. If you compare the two, you will see of course mostly the same articles, but some articles we have like the above "every zero dimensional space is T_3" is missing on planetmath.
So I guess that PlanetMath deleted some of their articles or reorganized them. The solution is then I guess as Paul suggests above, to delete our corresponding links to dead PlanetMath pages. Oleg Alexandrov (talk) 15:32, 17 October 2006 (UTC)
You sure its not a programming glitch or corrupted database index? The message I get right after "object not found" is "Please report this to us!" which sounds like they weren't expecting this to happen. Due to my general non-desire to deal with email, I haven't actually reported to them. linas 04:19, 27 November 2006 (UTC)

"Category:PlanetMath sourced articles" proposed for deletion[edit]

A CfD has been called against Category:PlanetMath sourced articles and most other "categories by source". If people want to defend this category (and I imagine they will), comment at Wikipedia:Categories_for_discussion/Log/2007_March_24#"By-source_categories". Jheald 12:20, 25 March 2007 (UTC)

Time for an update?[edit]

It's been more than a year since the last update. How about running my bot to add new entries from PlanetMath? Oleg Alexandrov (talk) 02:49, 14 August 2007 (UTC)

Yes, it's time. Paul August 03:05, 14 August 2007 (UTC)
Done. The article count went from 4802 to 6940 (stats changes). Those PM people are not wasting time. :) I'll make an announcement at the math wikiproject. Oleg Alexandrov (talk) 02:28, 20 August 2007 (UTC)
The newly added articles have a new! tag, that is easy to search for and/or notice. Oleg Alexandrov (talk) 02:46, 20 August 2007 (UTC)

new grant proposal: towards a PlanetMath Books exchange![edit]

Please lend your support to this Individual Engagement Grant meta:Grants:IdeaLab/PlanetMath_Books_Project proposal I've put together with PlanetMath contributor Raymond Puzio. Inspired by the PlanetMath Exchange project, our aim with this proposal is to improve the PlanetMath platform and make it easy to produce mathematics textbooks -- for export to places like b:Wikibooks. Your endorsement of the grant proposal would mean a lot! And any comments prior to the Sept. 30, 2013 deadline will help us improve the proposal. --Arided (talk) 22:27, 27 September 2013 (UTC)