Talk:Main Page/Archive 7

This is an article of past discussions, from 9-Jun-2011 to 28-Mar-2012. Do not edit the contents of this page. If you wish to start a new discussion or revive an old one, please do so on the current talk page.

Localisation

Thoughts on have localisation pages for separate languages(or possible even a whole other wiki(eg. fr.proofwiki.org, en.proofwiki.org, etc.)? --Joe (talk) 19:11, 9 June 2011 (CDT)

Are there techniques for generating these things automatically? If not, isn't it more sensible to let an in-line translation tool do it for you when you access the link?

We already have at least one page in this wiki in German, from someone who thought that would be a good idea. But I'm daunted by the maintenance aspects of having to do the work in multiple languages - unless, as I say, there's some automatic technique for handling the hard work. --prime mover 00:26, 10 June 2011 (CDT)

Definitely a `curator' would be needed for each language. I think it's omitting large community to exclude other languages -- though perhaps it ought to be restricted to the main languages in which maths is written, say Russian, French, and German, maybe Spanish (don't know what Asian languages are used). Probably most people could write in English, but I could write in French, I very much doubt that I would. --Linus44 08:47, 10 June 2011 (CDT)

S'il faut que je le faire ... ah, putain de merde, je ne le veux pas. --prime mover 12:35, 10 June 2011 (CDT)

I rather suspect a bot would be pretty terrible at translating stuff (just try asking google translate to switch a page to french/spanish/whatever now and see how readable it is. I tend to think that it would be better to stick to one language, both to avoid creating an insane amount of work to curate/translate/etc and because it would seem to lead to a higher quality/quantity of articles (compare other language wikipedias to the english one, for instance). Also, looking at the google analytics report for the past month: 12,000 visits from the us, 6000 from the uk, 3000 from india, 2000 each from canada and australia, 1500 from germany, 600 each from the philippines and pakistan, 500 each from south korea and france, etc. Also 80% of visits were from some flavor of english (no, I have no idea how they track that, but google analytics is awesome). Of course, feel free to disregard all of this, seeing as it comes from an American who speaks only english and rather limited spanish. --Alec (talk) 02:03, 12 June 2011 (CDT)

Agree that automatic translation is a non-starter; Google translate is next to useless, it just swaps each word in turn for something in the dictionary. Of the more intelligent options, comparing with a catalogue of properly translated phrases seems to be the best, but the biggest database (en français) I can find isn't really up to scratch (I tried "local field" but it just had some stuff about a meadow).

Also FWIW I changed my mind -- I think there's more to be gained by `sister sites' forming independently of proofwiki; any of the results here could be phrased, proved and categorised in ten different ways, and I see no reason to impose the existing structure on a site in another language.

It might be interesting to set up separate wikis as suggested, and demand only that the philosophy be followed. Perhaps a community might grow, and we can share stuff, manually translated as people are able/feel like it.

I would also add that the fact that the site is already written in English somewhat skews the data from Google analytics --Linus44 21:04, 12 June 2011 (CDT)

If anyone wants to set up a sister wiki in a foreign language, my voice is: go to it. If we find it's worth linking to, then we can do that. Until then I would suggest that Google Translate is perfectly adequate. If you can't understand the gist of what's being said then your grasp of your own language is probably a bit shaky. :-) As a basis for one to generate a different-language version of a given page it's a good start, at least for someone appropriately bilingual in those two languages.

But as I say: not me, not today. --prime mover 00:32, 13 June 2011 (CDT)

Just had another idea: who's up for working on a "mathematical translation dictionary" project which does translate "local fields" into "corps locaux"? Ça plane pour moi ... --prime mover 00:35, 13 June 2011 (CDT)

Interesting interactive mathematical research project

What's your favourite number?

I heard this guy talking on the radio this morning. Seems like a cool project - but more about psychology than mathematics.--prime mover 16:43, 22 June 2011 (CDT)

This is pretty cool. --Joe (talk) 16:47, 22 June 2011 (CDT)

New version of mediawiki

... is better tha 1.16.2. It doesn't have the tedious features of cut and paste where the focus would get lost.

Only problem is that it no longer allows global search and replace. However, I think most of the pages with "math" tags in have now been replaced with "dollar" tags, so that should not be a big problem. --prime mover 00:37, 23 June 2011 (CDT)

I can still get a list of the pages with math tags see them if I search for math> (yes, I intentionally left out the opening < to pick up any random loose closing tags) in Special:ReplaceText...

Incidentally, all of the remaining math tags are in the main/main talk, definition/definition talk, and user namespaces, with the exception of the MediaWiki:Common.js page and on User talk:Prime.mover and User talk:Joe (I suspect the ones on you guy's talk pages should probably stay, seeing as they are in reference to the switch and our no longer using the tag). --Alec (talk) 00:41, 25 June 2011 (CDT)

I spent ages looking for a way to search for pages with "math" in them! I just tried your technique. There are still a few such pages that need amending. I am fixing them by copying the text into an external editor, doing a search/replace there, and copying them back. It won't take too long. --prime mover 02:32, 25 June 2011 (CDT)

... all done, I believe. --prime mover 16:48, 25 June 2011 (CDT)

Feature: LaTeX in templates - can't use equals sign

If you are using a template such as Template:WIP or Template:Explain, and in the explanation text you include some $\LaTeX$, then beware you can't use the equals sign in it.

This is because the MediaWiki interpreter recognises it as a parameter indicator: LHS=RHS is interpreted to mean "replace instances of LHS in this template with the value held in RHS", examples of which can be seen in Template:Citation.

So if you find any pages with "explain" and "WIP" on them without any explanatory text actually included, this is probably because that text includes some $\LaTeX$ which includes an equals sign in it.

You can get round this by replacing = signs with \text{ equals }, as I don't believe there is an \equals tag (or whatever) available. --prime mover 16:43, 10 July 2011 (CDT)

Grading proof difficulty

One of the aims of this site is to allow accessibility to all. To that end, there are proofs which are simple and understandable by anyone, whatever their level of mathematical literacy, and there are proofs which require considerably more ability in order to get one's head round.

I wonder whether it's worth adding a "difficulty level" to the proofs we publish, so that casual browsers are able to choose to look at proofs which suit their level of expertise.

This is a quick thought which may or may not make any sense, but I'm throwing it out here so it doesn't get lost.

Anyone care to comment (positively or negatively)? --prime mover 18:22, 12 July 2011 (CDT)

I like the idea ... though it needs some thinking. --Joe (talk) 18:39, 12 July 2011 (CDT)

ITP 2011: Nijmegen, 27 Aug 2011

I have just returned from ITP 2011, the 2nd conference on Interactive Theorem Proving, held in Nijmegen, Netherlands between 22nd and 27th August 2011.

I had half an hour to present a talk discussing ProofWiki to a workshop in which several other similar-minded mathematicians had a similar opportunity. It was stimulating and offered up several new directions in which it may be advantageous to expand ProofWiki.

See ProofWiki:Current events/ITP 2011 (work in progress) for full details. --prime mover 13:03, 29 August 2011 (CDT)

What if we moved what is in current events now into Community portal somewhere, and put this in its place. Seems very current events like? --Joe (talk) 15:19, 31 August 2011 (CDT)

That could work - we could add info as it crops up of whatever other events are going to be happening in the MathWiki community. It's a lot busier out there than I thought it was. Seems it's an idea whose time has arrived. --prime mover 15:29, 31 August 2011 (CDT)

... Joe, would you be up for moving the "Is there anybody out there?" page into Community Portal? It looks like it may require layout skills over and above what I got. --prime mover 15:31, 31 August 2011 (CDT)

Sure, think it should go into Community portal or its own page? --Joe (talk) 15:38, 31 August 2011 (CDT)

It's big enough that it ought to be a link - unless we use a transclusion and stick it at the bottom. What do you think will look good? (Sorry for the delay, I got reading xkcd and just lost a quarter of an hour) --prime mover 15:50, 31 August 2011 (CDT)

I think we should put it on it's own page and put some links up somewhere, either in Community Portal, on the sidebar or both. I have to go out for supper, will get back to this later. --Joe (talk) 15:56, 31 August 2011 (CDT)

Okay, I have moved "Is there anybody out there?" into Community Portal as suggested, and disabled the contents list on that page so as to allow it to look okay, but it could do with tidying a bit. Anyone up for it? --prime mover 15:42, 1 September 2011 (CDT)

Change of Copyright Conditions

It has been suggested seriously that we need to change our copyright license. We are currently using the GNU Free Documentation License 1.2, but it may be better if we were to move to a Creative Commons license, as this allows more freedom of distribution.

In particular, if we use the same as Wikipedia and PlanetMath, i.e. that any material to be used by anyone, as long as its source is cited, we will be able to share freely any stuff back and forward with no worries. This is particularly important with regard to PlanetMath, as there is a lot of possible and desirable collaboration between the two sites.

But in order for this to happen, we need to make sure that all the contributors to this site are okay with this. I understand that this may be a difficult thing to achieve, as many (most?) of them are no longer active, and it may take some effort tracking them down and getting them to respond.

Advice needed from anyone who has knowledge of all this legal stuff, because I am not a lawyer. --prime mover 14:11, 30 August 2011 (CDT)

I know we've had this discussion before (Talk:Main_Page/Archive_3#Our_License). I don't have much knowledge on this topic either, so I'd be happy to differ to someone more knowledgeable. BTW, great job on the conference Matt. --Joe (talk) 15:17, 31 August 2011 (CDT)

Who was the "professional mathematician" who suggested we go to GFDL? Wikipedia uses CC SA not GNU. They obviously changed, so we can too.

My attitude is: if you want to use something from this site then you should be able to, although if you do, it would be nice if you say where you got it from. I don't understand how this will make it impossible to use it in a book that you have to pay for - although if the entirety of that book consisted of just stuff from ProofWiki I can see how that would be cheeky. --prime mover 15:57, 31 August 2011 (CDT)

So should I go ahead and switch over to a CC license? --Joe (talk) 15:05, 3 September 2011 (CDT)

My view is: if it offers specific advantages of shareability, then I see no problem doing so. But I don't know about the legal aspects of just changing over (and don't currently have the motivation to go and find out, as I don't know how much reading I'll need to do and there's other stuff I want to do instead), and I would welcome someone out there who knows about such stuff who can advise. --prime mover 16:07, 3 September 2011 (CDT)

Search box

Has something been done to the search box? Under Google Chrome it is behaving in a way it probably thinks is really clever, but in fact is irritating. This could be a function of development of Chrome's ongoing upgrade to HTML5 (the "Search Box" is now not just a bland text field). (Firefox doesn't have that problem, but it has other issues and I never use it.) Is there a way the HTML can be amended at infrastructure level so as to turn this "search" box back into an ordinary text input field? Try it out as it is, see what I mean. --prime mover 15:38, 1 September 2011 (CDT)

Is it better now? --Joe (talk) 18:35, 2 September 2011 (CDT)

You're a genius. What did you do? (I ask in a professional capacity, I'm trying to learn this stuff.) --prime mover 01:23, 3 September 2011 (CDT)

Nothing fancy, just a setting for the Vector skin.--Joe (talk) 09:54, 3 September 2011 (CDT)

Oh all right, I'll look it up. I'm supposed to be learning about MediaWiki as part of my day-job. --prime mover 14:34, 3 September 2011 (CDT)

This particular setting is from the Vector extension. If you have any questions or want to see some config files just let me know. --Joe (talk) 14:37, 3 September 2011 (CDT)

Cool - may do that. Chx. --prime mover 15:05, 3 September 2011 (CDT)

... unfortunately, the search box no longer behaves like a search box. It no longer dynamically displays the pages which so-far match what you have typed. --prime mover 14:33, 6 September 2011 (CDT)

I'll change the search bar for now so that it dynamically works for now. --Joe (talk) 14:41, 6 September 2011 (CDT)

H'mm ... works perfectly! I wonder whether it had got filled with spaces or something silly in my installation the other day. I'll do more experimenting next time something like this happens next time. --prime mover 15:11, 6 September 2011 (CDT)

Bot ideas and irc...

Bots

Does anyone have any ideas for some bots that would be useful?

A ProofWiki IRC channel

I was working on a project a few days ago and needed to question a developer about something. Turns out the project had an IRC channel, so just signed on and asked my question. Apparently I had found a bug, the developer fixed it in a few minutes, and I was on my way.

Does anyone think ProofWiki would benefit from an IRC channel of its own? Users could join up and chat, ask questions, get help with formatting, etc. --Joe (talk) 15:02, 3 September 2011 (CDT)

Do we have enough volume of traffic to make it worth while? On the other hand, will it be likely to increase the volume of traffic? More immediately important: will this cost? --prime mover 16:01, 3 September 2011 (CDT)

As for the suggestion: "Users could join up and chat, ask questions, get help with formatting, etc." users seem not to be very interested in asking about formatting, they just go ahead and use their own, preferring to do things the same way they prefer, rather than take the time to determine whether is such a thing as a house style, etc. And they never seem to stay around very long (with 3 exceptions) so with current levels of usage it's unlikely to fly ... I confess I probably wouldn't use it. --prime mover 16:04, 3 September 2011 (CDT)

I know I don't keep an IRC client open (and have no desire to), so I wouldn't be very helpful for this. Frankly, I tend to think IRC itself isn't widely used these days, and people can always ask questions by talk pages or email, which someone generally catches pretty quickly. --Alec (talk) 23:54, 3 September 2011 (CDT)

A useful bot

I recently changed Template:BookReference to allow a "prev" and "next" parameter for an experiment I was going to work on. To do that I had to change it to replace the 2nd and 3rd author parameters to "author2=" and "author3=" to keep it working.

What that means is that all multi-author works need to have that "author2=" and "author3=" added to them so the 2nd and 3rd authors carry on appearing.

Can we have a bot to replace all instances of, for example:

BookReference|Counterexamples in Topology|1970|Lynn Arthur Steen|J. Arthur Seebach, Jr.

to:

BookReference|Counterexamples in Topology|1970|Lynn Arthur Steen|author2=J. Arthur Seebach, Jr.

I've actually done a lot of the books by hand the last few days (stupid me).

But there's still the one above and also:

BookReference|Probability: An Introduction|1986|Geoffrey Grimmett|Dominic Welsh

to be changed to:

BookReference|Probability: An Introduction|1986|Geoffrey Grimmett|author2=Dominic Welsh

I think that's all of them. --prime mover 00:34, 20 September 2011 (CDT)

Can the search and replace page do this? --Joe (talk) 10:07, 23 September 2011 (CDT)

Yes of course - I'd forgotten about that. Job done. --prime mover 12:46, 23 September 2011 (CDT)

Talk page hard to follow?

Does anyone else find following a conversation across several talk pages super confusing (Especially when referring to a past discussion )? Seems like the format:

Me
:You
::Me
...

would work a lot better. --Joe (talk) 16:35, 13 September 2011 (CDT)

Can't argue with that, but it has to presuppose that both correspondents have a watch on the talk pages they are posting to. More than once I've continued a conv on my own user talk page and my correspondent hasn't noticed I've replied. --prime mover 17:03, 13 September 2011 (CDT)

I see your point. Though, though I think that if someone posts something on your (in general) talk page, generally they are either asking a question or making a statement. It's up to them to check and see what you've had to say. If it's something important you want to tell them, then odds are you'd be posting on their talk page ... which they would get a notification about if they have it set up. ( If any of that made any sense?) --Joe (talk) 17:11, 13 September 2011 (CDT)

Fully agree, and go along with it from now. If people miss important stuff, then that's just a real shame. --prime mover 00:44, 14 September 2011 (CDT)

External Links on Home Page

Do we allow external link (to whatever websites) on home pages? Or does it count as "spamming links to external sites"?

I suspect that we ought to allow any user to put whatever they want on their home page ("within reason"). What say? --prime mover 18:39, 27 September 2011 (CDT)

Yeah, as long as it's not to a phishing site or something I guess. --Joe (talk) 19:29, 27 September 2011 (CDT)

GeoGebra extension

Thoughts? --Joe (talk) 17:59, 3 October 2011 (CDT)

Impressive. I will give it a good workout when I'm in the mood to get back to Euclid. --prime mover 00:35, 4 October 2011 (CDT)

Automatic verification of proofs

Hello ProofWiki users!

I have a thought (nothing concrete at this point) about developing a "wiki"/Q&A site for mathematical proofs, with automatic verification included. This would be similar to ProofWiki, with the difference that you would know the theorems and proofs in it are flawless, and aren't "hand waving".

Please see this (perhaps badly titled) blog post for more details.

I welcome feedback of any kind. I'm not a user of this wiki though, so I won't be coming here to check responses inline - please leave a comment on my blog or email me at [email protected] if you want me to see your feedback.

Ripper234 11:18, 6 October 2011 (CDT)

Linguistic ruling: "alternate" vs. "alternative"

I understand that in American English, the word "alternate" has evolved so as to be another word meaning "alternative". However, in British English, the word "alternate" is used to describe a sequence of things which are taken in turn, e.g. "a, b, a, b, ..." or as squares on a chessboard: "black, white, black, white ...", and can also be a verb with a slightly different pronunciation: "to alternate" meaning "to take one of two different values in turn" etc.

I would suggest that when there are two proofs or definitions posted up, the term "alternate proof" or "alternate definition" should not be used, but "alternative" used instead.

However, there is a linguistic problem even with this: because when there are three or more options, you technically should not use the term "alternative" because strictly speaking it means "the other option of two". If you have three or more choices, then "alternative" is not the term to use (putting contemporary media creations called "the third alternative" aside).

If "alternate" is just being used as a more complicated-looking (and therefore cleverer-looking) word for "other", then I suggest it should not be used at all, and instead just number the proofs "Proof 1", "Proof 2" ... etc.

If a proof has a description (e.g. "direct proof", "proof using induction", "proof using recursion") then use that, indeed, if you can encapsulate that description in one or two terms, but a simple numbering should be enough.

In short, shall we try to make an effort to lose "alternate" (and also "alternative", if you like) out of page subtitles? --prime mover 00:39, 7 November 2011 (CST)

Spamming accounts

We have a fair number of indefinitely-blocked users whose sole purpose was to act as advertising outlets for a variety of piffle. These, as I say, have been blocked and their contents removed, never to be seen again.

The question is: is it possible to delete them completely from the database, as though they had never been created in the first place? As it is, they cause the number of users to be artificially considerably higher than is accurate. --prime mover 14:05, 9 November 2011 (CST)

Humm... for some reason I don't get emails for this page. It would be possible to delete these accounts, but it would be a pain. As well, new accounts are still being created for spam purposes ... there were several today. The only difference is that they give fake email addresses. So I don't think deleting them would truly address the problem. --Joe (talk) 19:46, 9 November 2011 (CST)

Okay, so 10 new accounts today, all but two in the format (boring name)(dull name)(3-digit number) (and one which is 3 letters followed by 3 digits) which are obviously spamming accounts. Hence we're well on the way to 700 users, although I would guess that of the last 60, probably about 5 are actual serious PW accounts. Not suggesting any solution, just sayin. --prime mover 14:51, 15 November 2011 (CST)

It's pretty annoying, but for the moment I don't see anything we can do about it. The mail server is getting a ton of bounced authentication emails as well. I hoped the ReCaptcha and checking the IP's against SORBS would fix the issue. I'll have to do some research to see what options are available. --Joe (talk) 21:10, 15 November 2011 (CST)

I think I've defeated the spamming ... for now. --Joe (talk) 08:07, 24 November 2011 (CST)

I just got really confused for a second when suddenly all the edit links were gone and I had to reconfirm my email :p --Alec (talk) 10:43, 24 November 2011 (CST)

Congruence Symbol and POTW

Point 1: It's time for a new proof of the week, no? It's been a while since May 2.

Point 2: Anyone else get an extraneous space after the congruence symbol? Like in |$\cong$|, on my computer in both Google Chrome and Mozilla Firefox I get an extraneous space. A similar thing in $\sinh x$ between the "n" and "h", but it's not as noticeable. $\cosh z$ works fine. --GFauxPas 07:30, 17 November 2011 (CST)

It's a function of MathJax. I suggest we live with it. We do not want to spend a lot of time making fiddly reformatting of code when a new edition of MathJax may well have addressed the issue. We have already made a few adjustments where the rendering interferes with the interpretation, but I've got no interest in adjusting stuff where it reads perfectly adequately.

As for POTW, I expect someone will get round to it sooner or later, if they can be bothered. --prime mover 08:08, 17 November 2011 (CST)

Internal linking

I just had an idea for linking to pages explaining a notation. Something like this:

Then $f$ is described as multiplicative iff:

$m$ $\perp$ $n \implies f \left({m n}\right) = f \left({m}\right) f \left({n}\right)$

Click on the $\perp$. Thoughts? --GFauxPas 12:59, 17 November 2011 (CST)

I've thought about it, suppose it might work. But.

a) Before MathJax, links which consist of just LaTeX don't appear as links (i.e. in blue) so it's not easy to see. And in fact this one doesn't appear blue to me, but maybe that's because my eyesight is deteriorating (and it was never very good to start with).

b) You can't right-click on a LaTeX-only link and get the "Open link in new tab" (at least, not in Google Chrome and FireFox), you just get "View source". I argue that you really don't want to navigate away from the page you're reading when you are investigating the meaning of a symbol.

What the answer is, IMO, is the compromise of writing underneath the line where it's used:

"where $m \perp n$ indicates that $m$ is coprime to $n$"

... and that, generally, is what is done. If not, then it ought to be. --prime mover 14:11, 17 November 2011 (CST)

Suggestions

1) I still think it would be a good idea to have a "suggestions box" type page, but whatever

2) How about a "random definition" option, like there is for random proof?

3) Would it be a bad idea to have a new category, "problems"? Like there are some classic problems that aren't definitions or theorems, like the falling ladder problem, birthday problem, largest open box, $\int \cos x e^x \mathrm dx$, finding a fake pearl out of 9 with only three weighings of a balance scale, pound of feathers pound of gold, etc etc. --GFauxPas 15:57, 12 December 2011 (CST)

1) Wotever. This talk page has proved adequate so far. Anyone else think it's a good idea?

2) The "random page" function comes with MediaWiki as part of its inbuilt functionality, but it excludes all pages not in the main namespace, i.e. everything with a colon in it. Therefore definitions and axioms etc. don't feature. I don't know how easy it is to get it to do this, but why not just a random page of any description?

3) We already have Category:Classic Problems - is this the sort of thing you were thinking about? To a certain extent I agree, but I think we'd have to ruthlessly subcategorise or else it would become too unwieldy a category. (The pearl problem BTW is out of 12.) --prime mover 16:11, 12 December 2011 (CST)

Oh okay, 12. But yeah, are any/all of the things I mentioned the kind of things that belong on PW? --GFauxPas 16:17, 12 December 2011 (CST)

If you have the patience to a) paste them up, and b) prove the answers, then go for it. --prime mover 16:39, 12 December 2011 (CST)

How do I make a category for "classic Calculus problems"? --GFauxPas 17:38, 12 December 2011 (CST)

Same way as you make a category for any sort of page. Add the category at the bottom of the page and it appears in red. Click on it and it takes you to the new category page. Then add the supercategory at the bottom. --prime mover 00:19, 13 December 2011 (CST)

Okay, sorry to come across all wikipedia and all that, but what is the decider as to what is considered a "classic problem"? Specifically, what makes Derivative of x to the x "classic"? Can't say I've even seen it in any of the calculus / analysis books I've got (not that I have that many) except perhaps as an exercise, or as a stepping-stone to get somewhere else. No matter, as I established a while back, I can't set myself up as an arbiter. --prime mover 16:25, 13 December 2011 (CST)

I don't mind you being an arbiter. It's just my impression from Khan Academy and my Calc I prof. that this is "the" example given to show people the "trick" of taking the logarithm of both sides, in the same vein that "the" example of Integration by Parts is $\int e^x \cos x \mathrm \ dx$. If you don't think it's classic enough, you can remove the tag without offending me. --GFauxPas 17:08, 13 December 2011 (CST)

Would it be a bad idea to have a new template like there is for {{AoC}}, something along the lines of:

Because of the multitude of trigonometric identities and the many surprising relationships between $e$ and trig. functions, this integral can be presented in many different ways. However, they all differ at most by a constant. ? --GFauxPas 10:27, 20 December 2011 (CST)

Don't like it much. But then I'm not a fan of this sort of imprecise language. If there are many different ways of presenting an integral, then (if they are genuinely different, and this different is enlightening) add them.

I think it's worth guarding against the tendency to just "talk about" results unless there's something genuinely informational to present. Saying "There's lots of interesting stuff that comes from here" without actually specifying what that stuff is doesn't seem to fit the philosophy of this site (which, yes I know, needs tightening up). Less is more. --prime mover 14:53, 20 December 2011 (CST)

New host

We have the same spam bots coming back from before you added the "add PW's name" option, Joe --GFauxPas 12:17, 5 January 2012 (EST)

On it. --Joe (talk) 12:34, 5 January 2012 (EST)

Does anyone else find this server a lot faster? --Joe (talk) 14:43, 5 January 2012 (EST)

It seems faster to me, yeah. --GFauxPas 15:09, 5 January 2012 (EST)

Not for me, not particularly. --prime mover 15:13, 5 January 2012 (EST)

I had noticed the increase in speed also. It is quite significant for me. How a convenient corollary. --Lord_Farin 15:28, 5 January 2012 (EST)

F'nality of MediaWiki 18(?)

... or whatever number it is.

Is here an option to get the app to alert an editor if they attempt to close a page with unsaved edits on it? I confess I have got used to that feature - and I will probably suffer from the loss of quite a few edits before I get used to the fact that it doesn't seem to be able to do that any more.

Apart from that - Good job. Let us know about hosting fees. --prime mover 14:44, 5 January 2012 (EST)

There are still a few extensions for the vector theme I haven't brought back online... that feature might be in there. I'll look into it. --Joe (talk) 14:46, 5 January 2012 (EST)

Problem with \vec

The $\LaTeX$ command \vec appears to be broken. Instead of putting a short arrow over the target character, it seems to superscript a square: $\vec a$ and $\vec b$ and so on. --prime mover 05:35, 8 January 2012 (EST)

Looks fine on my browser, what are you using? --Joe (talk) 16:06, 8 January 2012 (EST)

Hmm. On Google Chrome it's broken, Firefox is fine. I've been having trouble with GC (my browser of choice, don't judge me, it doesn't make me a bad person ;-) since the server move - although I wonder whether that was coincidental with a GC version upgrade (these have been increasingly flaky over the last few months).

I have no desire to check it in IE (any version) but I may give it a go in Opera and Safari tomorrow (if I remember) from my work PC (my job requires that all the commercial s/w that I'm involved in works on all the big 5 browsers). --prime mover 17:43, 8 January 2012 (EST)

Interesting, I'm using Chrome as well, only the dev version. Must just be a version problem. --Joe (talk) 17:48, 8 January 2012 (EST)

I'm on version 16.0.912.75 m. --prime mover 17:59, 8 January 2012 (EST)

17.0.963.26 dev --Joe (talk) 18:01, 8 January 2012 (EST)

Could explain it. No worries, I'm not doing vectory stuff atm anyway. --prime mover 18:08, 8 January 2012 (EST)

Looks fine to me in chrome 16.0.912.75 m... --Alec (talk) 14:46, 9 January 2012 (EST)

Utterly bizarre. And I've even cleared my cache and deleted all my history. --prime mover 14:58, 9 January 2012 (EST)

Change house style to $\mathbf v$ problem solved :P --GFauxPas 15:35, 9 January 2012 (EST)

I'm more than happy to agree. All those fiddly little arrows are so grade-school. --prime mover 02:37, 10 January 2012 (EST)

It works fine here, I'm using Firefox 3.6.24 for Ubuntu; actually, I'm using LXDE (Lubuntu), but should work on GNOME as well. --L.P. 23:03, 16 January 2012 (EST)

New User Messages

... don't seem to be happening today. Two new users a couple of hours ago still have not been introduced. --prime mover 13:22, 13 January 2012 (EST)

Fixed! --Joe (talk) 13:57, 13 January 2012 (EST)

Plural of premise

I've seen premises and premisses, both in authoritative sources. Which should I use?

Good question. I believe "premises" is the plural of "premise", and "premisses" is the plural of "premiss", which is also used. I have always preferred the first.

And a style ruling, please:

Also see

• Theorem name, or
• Theorem Name? --GFauxPas

Shrug. Don't really care, depends on how I'm feeling. The latter is better as it saves having to add a | field in the link. --prime mover 12:25, 22 January 2012 (EST)

Citing general documentation on the web

What does anyone else think about using sundry pdf's found on the web (usually as part of a set of specially prepared lecture notes) as source works for citation as references, for example:

Stanford handout

?

My view is rather not, as there is no indication of who wrote it. Also, web citations are ephemeral (we've found this earlier, citations entered into a proofwiki page one day had vanished by the next day), and unless backed up by hardcopy somewhere, IMO can't necessarily be trusted. Besides, except in cases of research topics, all the information can be found in a book somewhere, or a published paper that can be cited as an entry in a journal.

Perhaps I am oldfashioned by placing too much reliance on printed matter, but that's my position. --prime mover 01:44, 24 January 2012 (EST)

In general I would agree with you, as I'd indeed say that a hardcopy reference should be reported by a trustworthy user. If such is the case, generally that user has in fact taken the lectures corresponding to the lecture notes, making them a lot more reliable. In such a case, it is in my opinion acceptable to reference the notes in absence of any properly published book. --Lord_Farin 02:16, 24 January 2012 (EST)

Template:WIP

I hacked into Template:WIP with direct HTML code to allow for multiple correctly formatted lines of comment. --Lord_Farin 09:52, 26 January 2012 (EST)

Good idea. There may be other templates which may have similar work needed doing. --prime mover 12:22, 26 January 2012 (EST)

I have done all the commonly used (i.e., I recognized them) templates in this way. I encountered the gem Template:Inuse while processing (and while I'm at it, let me point out Special:Allpages as well. It is quite useful sometimes. --Lord_Farin 13:37, 26 January 2012 (EST)

Proof structures

Some while ago I started structuring iff proofs into two sections titled "Necessary Condition" and "Sufficient Condition", with both those titles being links to those definitions.

However, more recently we have adopted the more long-term strategy of splitting out multiple proofs into separate pages, each of which are likewise linked to from the section titles, e.g. "Proof 1", "Proof 2" etc.

It occurs to me that those approaches may become mutually incompatible, as we may want to separate out various proof sections into their own pages, so that clicking on e.g. "Necessary Condition" would take you, not to the page defining "N.C." but instead to that partcular (probably transcluded) section of the proof.

What we would want to do is then, having "unlinked" the title, add the words "Proof that the condition is necessary:" (or however the wording should be) underneath that title. Then we would be free to make that section title a direct link to the transcluded subpage containing just the proof of the theorem in that direction.

I have already been doing something similar to proofs of equivalence relations and orderings, and so on, and there may be other cases which use that same paradigm.

Comments? --prime mover 06:31, 18 February 2012 (EST)

This occurred to me too a while ago. I didn't feel like putting it in public as it appeared to be a minor issue. A really satisfactory solution hasn't occurred to me so far; I imagine that some time in the future we would like to use the MediaWiki extension idea that Joe threw up a few months ago. On yet another track one could argue that the splitting of these kind of proofs into multiple pages isn't really necessary; I haven't encountered an instance where I wanted to refer to (the proof of) only one side of the bi-implication. This is different with the multiple proofs case, as demonstrated by the changes processed to Derivative of Exponential at Zero last week.

For consistency and clarity, I think it would be best to have links in titles only when this means that the associated paragraph is transcluded. In cases like with the bi-implication, the creation of a link in a title should maybe best be avoided entirely.

Finally, an approach I have been coming to appreciate more and more when dealing with multi-staged proofs is a brief explanation at the start of the proof of what the main line is (basically, using a one-line introduction and summary encapsulating multiple detailed proof sections). This essentially removes the need for linking in titles in a quite appealing manner. --Lord_Farin 17:34, 18 February 2012 (EST)

One of the issues is that the texts often treat just one half of the bi-implication, and others treat the other half. I know it looks silly but I think it's worth separating out the pages in these cases, so as to be able to arrange the citations to match the pages. --prime mover 18:37, 18 February 2012 (EST)

In that case, feel free to use my sentence 'it would be best to have links in titles only when this means that the associated paragraph is transcluded' which then applies. This could be promoted to house style, but there is a humongous amount of changes to be processed. --Lord_Farin 08:36, 19 February 2012 (EST)

It does not all have to be done now. And I would rather "lead by example" rather than enshrine it in a house rule. But that's because I'm tired and I don't fancy doing either job. --prime mover 09:46, 19 February 2012 (EST)

Also see

I have removed a few links from "Also see" sections just now where the link goes to a concept which is already used in the main body of the page - in particular where the "also seen" concept is an integral part of the definition.

While there is no harm in including the concepts in the "Also see" section, it could add to the clutter of the page, and goes against the minimalist philosophy. On the other hand, gathering all such concepts together at the bottom could be seen as a useful approach (but would mean there's another 10000+ pages which need this done to them!).

What does anyone else think? It's a point at which I'm ambivalent about - I really can't decide which approach I prefer. --prime mover 06:35, 18 February 2012 (EST)

I feel quite strongly that closely connected concepts should be easy to locate; this is, AFAIC, not accomplished to a satisfactory standard when these are only accessible by some link in mid-page (maybe obscured by adapted text), which effectively forces one to read the entire entry when actually looking for something closely related. So I feel the Also see sections should rather be expanded than shrunk. --Lord_Farin 17:34, 18 February 2012 (EST)

Okay, that argument works. "Also see" will be encouraged to grow, in that case - but I still think it's worth being alert to putting everything in an "Also see" that's vaguely peripherally related - I see this as a section for strongly, directly related concepts. Whatever that means. --prime mover 18:39, 18 February 2012 (EST)

Also known as

For the situation where there are multiple names for a thing, it's been haphazard up till now. Sometimes the extra names will be in the same line as the definition: "A flibby widget (also known as a flabby tatter) is a ..." and sometimes in a section called "Alternative Name(s)" and so on, neither of which is ideal. Sometimes the alternative name and notation hang out on a nondescript street corner labelled "Comment(s)" or "Note(s)". Time we cleaned up the ghetto.

In keeping with the "Also see" convention, I propose an "Also known as" section as appropriate, where all the variant names and notations can be documented. And if we are going to keep with the "natural order" convention (which suggests that the further down the page you get, the more peripheral the subject matter), I suggest it goes just above the "Also see" section.

I will think about getting it together to specify the order of sections on a page sometime, but life has taken a turn for the busier recently and I no longer have the empty acres of spare time that I used to have ... --prime mover 16:47, 19 February 2012 (EST)

Springer On-Line Encyclopedia of Mathematics

This has recenlty been wikified: http://www.encyclopediaofmath.org/index.php/Main_Page ... Note that they are using the same technology as ProofWiki. Where we lead, lesser entities follow. --prime mover 15:55, 22 February 2012 (EST)

It's like the Wikipedia version of ProofWiki --GFauxPas 21:18, 22 February 2012 (EST)

I still think our approach is better. ;-) --prime mover 17:31, 23 February 2012 (EST)

Oh of course, I meant it's the wikipedia version of proofwiki in the sense that the sentences are hard to follow, it's messy, there aren't enough links to other concepts, and the LaTeX is poorly rendered. --GFauxPas 17:36, 23 February 2012 (EST)

Question

Other than properly citing the source, under what circumstances am I allowed to incorporate material from an outside source into ProofWiki articles? Abcxyz 23:35, 2 March 2012 (EST)

Depends what you mean by "incorporate" and "outside source". Copy/pasting from Wikipedia etc. is okay as long as it is converted into our house style. Otherwise adhere to the terms of the copyright from where it comes. --prime mover 01:04, 3 March 2012 (EST)

For example, is posting a variation of [1] allowed? I'm not exactly sure whether that is allowed under the copyright of that article.

It says:

"Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at

http://www.jstor.org/action/showPublisher?publisherCode=maa.

You might want to do that. --prime mover 14:01, 3 March 2012 (EST)

I'm guessing that it is very likely allowed, because that has been done on Wikipedia (in widely read pages, and using articles from the same time period and publisher as the one I mentioned) and probably a lot on ProofWiki as well. Comments? Abcxyz 00:41, 16 March 2012 (EDT)

Another question

Is there anywhere on the site where its policies are defined? I have several questions and would not want to bother people if they are already answered. For example,

1. I and another person are the authors of proofs now hosted on Citizendium. Is it reasonable to also host these proofs on ProofWiki? The proofs are of the orthogonality of the Associated Legendre functions and solutions to the Sturm-Lioville equation.
2. SInce understanding these proofs depends on understanding the supporting theory, is it reasonable to copy the theory pages for the underlying concepts from Citizendium to ProofWiki? For example, the Associated Legendre function theory page and Sturm-Liouville theory page.
3. The person with which I collaborated on these proofs is now deceased. I would like to acknowledge his work on them. Is this allowed or does ProofWiki have a non-attribution policy for all contributed content?
4. Are there people who will help put the proofs and theory page mathematics in house rules form?

Note all of the material on Citizendium is licensed under the Creative Commons-Attribution-ShareAlike 3.0 Unported license. Dan Nessett 13:52, 6 March 2012 (EST)

Let me respond as far as I feel comfortable that I can answer correctly and unambiguously. Well, actually, as I am a bit of pressed for time, let me just assure you that I and prime.mover are more than willing to help you out with house style. The other points will have to wait a bit. Sorry for that. --Lord_Farin 14:09, 6 March 2012 (EST)

House style won't matter because it can always be adjusted.

Everything here is GNU Free Documentation Licence 1.2 at the moment, but there is a question as to how we should set it up to be as free as possible. Copyright (in the eyes of at least one contributor) is an even bigger and more unnecessary millstone round the neck of the universe than organised religion.

If existing work has an attribution notice which needs to accompany it, then make sure that's present.

It's a fair bet that, although the works they come from may themselves be under copyright, the actual proofs that they document are probably far older than that and therefore not subject to copyright.

Copying pages byte for byte is frowned upon as we are then open to the criticism "Don't bother about going to ProofWiki, all they are is a poor copy of Citizendium / Wikipedia / whatever." Copy the contents by all means, copypasta the equations, etc. but unless (a) it's your own work and (b) it already fits house style perfectly, then take a looksee at what we already have and try to fit it to that. --prime mover 15:34, 6 March 2012 (EST)

The chances that any actual proof is subjected to copyright are slim, especially when it does not concern any very recent insights into previously unsolved problems. I would say that one shouldn't bother too much. --Lord_Farin 17:59, 6 March 2012 (EST)

The concepts behind a proof cannot be copyrighted, and I agree that no new fundamental ideas appear in the proofs. However, copyright is a complex issue. Any modification of previously published results constitutes a derivative work that is subject to copyright. I tend to agree that copyright is an impediment to information sharing, but licenses such as the GNU Free Documentation License and CC-by-sa 3.0 help get around most of the serious problems.

When we (my collaborator and I) contributed the proofs to CZ (which we created from information in text books, but which are not direct copies of anything we found), we licensed it to CZ. That license allows its reuse as long as attribution is given for authorship. In the case of the proofs I cite, there are only 3 or perhaps 4 contributors, so attribution would be easy as a footnote.

In regards to copying verbatim, I understand your point about possible criticism. However, the proofs are pretty concise and changing them cosmetically just to ensure they are different than what appears on CZ seems a little silly. The advantage of having them appear in both places is if either site goes dark, they would remain available to the internet community. An additional advantage is it is far more likely that someone looking for these proofs would automatically visit ProofWiki and it is less likely that they would look for them on CZ.

In regards to the theory pages. The only reason I mentioned them is that without the theoretical background, it isn't clear why anyone would be interested in proving the orthogonality properties. However, another viewpoint is anyone looking for such proofs would already be familiar with the underlying theory. So, the real question is whether ProofWiki has an objective for self-contained content. Dan Nessett 15:59, 7 March 2012 (EST)

We have an aim for self-contained content. Moreover, we even strive for full rigour from axioms. Copying verbatim will most likely be impossible because of necessary adjustment to house style. In any case, this discussion justifies probably the creation of a 'link to CZ' template and/or a notification on ProofWiki:Copyrights that material may have originated on CZ (which I believe doesn't contradict CZs licence as it permits the creation of 'collections'). Oh, and thanks for calling PW a better place to look for a proof than CZ ;). --Lord_Farin 16:06, 7 March 2012 (EST)

What he said.

Your last paragraph expresses a viewpoint which conflicts with the philosophy of ProofWiki. Every proof is dependent upon previously-established results and definitions, which in turn are dependent upon more fundamental results, and so on back until the axioms are reached (ultimately ZF(C) and whichever axiom schema of propositonal and predicate logic). The orthogonality properties are to be proved (whatever they are, my knowledge of such stuff being limited to my memory of my undergraduate days). The idea is that anyone should be able to come upon any page, and have all the information required (via links to simpler, more basic stuff) to understand that page. The concept that the only people expected to understand a page are those who already understand it is at odds with our stated aim. --prime mover 16:14, 7 March 2012 (EST)

WRT to copying verbatim, perhaps my writing was unclear. I didn't mean the proofs wouldn't require modification to conform to house style rules. I have no problem with that. What I meant was cosmetic changes in content would seem a bit silly. For example, changing the proofs so that one proof line is divided into two or two lines are combined into one just to ensure the proofs on ProofWiki are different than those presented elsewhere doesn't make a lot of sense to me. There might be reasons to make such changes in order to improve clarity, but that is a different issue.

I take the last remarks by prime mover to mean ProofWiki would benefit from importing not only the proofs, but also the theory pages. Have I interpreted those remarks correctly? Dan Nessett 16:37, 7 March 2012 (EST)

yeah --prime mover 16:41, 7 March 2012 (EST)

I have imported the proofs and theory pages into Sandboxes in my user space. I have not yet started working on bringing the text into house style conformance. Before getting to that, I have a couple of more general questions:

1. In the proofs and theory text, there are references to links that exist on CZ, but not on ProofWiki. Mostly, these are to concepts such as the Kronecker delta, physics, mathematics, etc. How should I handle these? Should I remove the link reference and replace the word with plaintext? Should I turn these link references into external link references to WP, CZ or some other content? Should I do something else?
2. One proof is so small that a TOC is not automatically generated. In the CZ version there is an explicit callout to __TOC__ to force a TOC. Should I keep that reference or remove it?
3. How do you want to handle licensing attribution. You mention using a template for CZ. Do you want me to import the template used for that purpose on WP? Do you want to create your own template? Do you want to do something else? Dan Nessett 13:20, 8 March 2012 (EST)

I will try and answer 1 and 2. P.m will have to do 3 as copyright and citing are his natural domain (at least, more than they are mine).

Concerning 1, eventually the pages will need to be linked to what already exists on PW, so the links should be replaced with the appropriate ones to PW pages (note that the particular pages might not exist at this point). It is also most likely that we decide that it's best to split these pages into several, so as to ensure that pages handle on concentrated subjects (note that this does not apply to the proofs; look around a bit for how we usually handle them).

Concerning 2, we don't enforce the appearance of a TOC on every page. Therefore, the reference may be removed.

By the time it's been fleshed out, you usually find that (what with an "Also see" and "Sources" section) there are enough sections to trigger the (default) TOC appearing anyway. --prime mover 16:54, 8 March 2012 (EST)

Let me conclude by saying that I appreciate your contribution of approximately 40k characters worth of information to PW ;). --Lord_Farin 13:34, 8 March 2012 (EST)

What's TOC? --GFauxPas 16:06, 8 March 2012 (EST)

Table of contents. --prime mover 16:22, 8 March 2012 (EST)

A further point to add to the above: "It is better to beg forgiveness than to ask permission." In this context, banging up loads of pages of good proofs in a not-quite-right format is IMO a better way to go than waiting till you've learned all the rules before posting anything. Someone will (huff and puff and) go in and fix all the styling issues and make sure it adheres to house style and philosophy. As long as you're happy with that (many contributors throw their teddybears out of the pram when their favourite notation is changed to the house convention, which causes friction all round) there should be no problem. :-) --prime mover 17:06, 8 March 2012 (EST)

Dissatisfaction with Analysis

I'm not fully happy with the way analysis is being treated at the moment, but I haven't got a solution and I need input.

Lots of results are applicable to both real and complex domains, and some are applicable to rational numbers as well. (Some may even be applicable to hamiltonians and even octonions but that's another thing altogether). There are two approaches:

1. Merely apply the result to the complex field, and take it for granted that as reals and rationals are also complex the result holds for them too. But that doesn't work for me because:

a) Some results may specifically not hold e.g. for rationals because e.g. they're not complete. So we need to make sure that results that do apply to both are explicitly explained as doing so.

b) While the complex result often does encompass the real case, the latter is important to be treated separately because then they can be accessible to those who have not studied complex numbers, and it has been believed necessary in the past to establish the results solely in the context of real numbers (whether this still applies in the modern world I wouldn't know, I left academia decades ago).

2. Indicate at the top which domains are relevant, e.g. Let ... be a sequence in $\R$ or $\C$ ... and then later on use something like "This follows because of (result for reals) or (result for complex)" which is messy.

3. At this stage, not to try and piggy-back the complex results into the same page as the existing real result, but start anew with the result for complex numbers, then link between the two with an "also see". Once that has been established, the commonality between the two can be explored.

Any thoughts on this? (The above is not written as carefully as I wanted to because I'm currently being distracted by something else, sorry.) --prime mover 03:33, 12 March 2012 (EDT)

You have quite successfully made your point for option 3. (which, I presume, was the goal of this wall of text ;) ). It seems that we could take that path, then give some of the proofs in the real section by merely stating that they are implied by the complex case. But that's not even strictly necessary. I suggest to proceed with option 3. --Lord_Farin 05:17, 12 March 2012 (EDT)

Rendering of small spaces

Seems that MathJax may have fixed an issue with "\," as a LaTeX command, as it now seems to put a small space in place. We can, if we want, replace all instances of (for example) $\left[{a .. b}\right]$ with $\left[{a \, . \, . \,b}\right]$ as it was in the beginning. Thx fo Abcxyz for noticing this. --prime mover 03:49, 16 March 2012 (EDT)

I recall it's very fiddly to do that, and the visual presentation gain is, at least to me, somewhere around zero (maybe even negative). So I suggest to not bother. --Lord_Farin 03:53, 16 March 2012 (EDT)

There are some situations where I think it would be very beneficial. Specifically, the "prime" mark never stood out enough to make me happy, e.g. $f\left({x}\right)$ and $f'\left({x}\right)$, vs. $f \ '\left({x}\right)$ And in general, I don't like when characters touch each other when they're not supposed to. (insert lewd joke here). --GFauxPas 08:44, 16 March 2012 (EDT)

So, see how this looks:$f \,' \left({x}\right)$ ... aha, perfect.

As I say, it appears that MathJax may well have fixed something. I have fixed a lot of instances of touching characters (in the UK we say: "Ooh Matron ...") by adding backslash space to the LaTeX as appropriate, so we may do well to replace these with backslash comma instead. The way I see it is: if there are instances where we can improve the look by amending this, then it's a nice thing to do when we're in there, but we don't want to start a project of going through the entire database and changing them all systematically. --prime mover 08:52, 16 March 2012 (EDT)

... and looking at some of the work done on primes on various pages, this is a seriously good result. This site is in serious danger of looking almost professional in places. --prime mover 16:09, 16 March 2012 (EDT)

I'd say it's bound to as we are effectively taking books and applying full rigour to their results; good thing that the presentation will catch up to this standard. Keep up the good work us ;) --Lord_Farin 16:11, 16 March 2012 (EDT)

ProofWiki Extension

We've ( Lord_Farin mostly) been developing an extension at test.proofwiki.org. Check it out (read "try to break it") and post your feedback.

Site: test.proofwiki.org
User: proofwiki_test
Password: proofwiki.orgtestwiki

--Joe (talk) 16:30, 18 March 2012 (EDT)

Is this the subsection extension that was discussed a while back? --Alec (talk) 23:26, 18 March 2012 (EDT)

Probably. More information can be found on User:Lord_Farin/Long-Term Projects/Extension. I will try to put up some elementary guidelines to usage soon. --Lord_Farin 03:19, 19 March 2012 (EDT)

The Big 5000

We are on the brink of registering 5000 proofs on ProofWiki. Admittedly this number has gone up more quickly than usual recently, because of a lot of recent refactoring.

It makes sense that the 5000th proof be an important one, and it would look a bit flat if it were the result of a simple refactoring of something trivial.

So here's the challenge to the active users: to post up the 5000th proof - and make it a good one!

For these purposes we need not be too concerned that all the supporting links are in place. If it turns out to be a sea of redlinks, then so be it. It would give us an impetus to fill in a lot of gaps. --prime mover 03:44, 21 March 2012 (EDT)

I am rapidly converging towards Carathéodory's Extension Theorem for measures (saying that, given some side parameters, it is enough to specify the values of a measure on a generator). This theorem is very important as it grants existence of Lebesgue measure, making the field of measure theory actually relevant and incorporating the rationale for its definitions. Sounds like a good candidate to me. --Lord_Farin 05:14, 21 March 2012 (EDT)

Perhaps I should extract out the corollary to Homogeneous System has Zero Vector as Solution and give it it's own page as the 5000th proof --GFauxPas 09:06, 21 March 2012 (EDT)

Capitalisation convention

It is house style to put the titles of pages in 'Title Capitalisation'; i.e., important words should be capitalised. However, I have determined inconsistency as to whether the word 'under' should be capitalised or not... This may be an issue because page titles are case sensitive, and thus it should be preferable to have a consistent solution for this, but neither seem satisfactory. Opinions? --Lord_Farin 08:56, 22 March 2012 (EDT)

The glib answer is: if "under" is important (e.g. "Primes Under One Hundred have Less than Three Digits", then it's capitalised, but if it's juat a filler word, (e.g. "Integer Multiples under Addition" then it should not be. (I suspect your usage is as the first type.

The more this site progresses, the more I regret starting that capitalisation style. It looks smart and professional, but it's a bit imprecise in execution. If only we'd used first-letter-only capitalisation, or every-word-capitalised ...

We can of course always use redirects. It bulks up the proof numbers, which is a bit of a downer, but it allows flexibility. When to impose a redirect? When the title you guess gives a redlink. --prime mover 13:31, 22 March 2012 (EDT)

The annoying and ill-featured search engine of MediaWiki has pushed me to use Google with the prefix 'site:proofwiki.org'; this works brilliantly to find results you know must exist. The problem with title capitalisation is as you mention; the benefit of it looking better is a great good IMO. (Un)fortunately, you express mostly what I had already came up with myself. There may be cases where I didn't adhere to this in the past, please change these upon encounter. --Lord_Farin 13:38, 22 March 2012 (EDT)

Re: the search system, can't we incorporate Google's search into PW? Cf. what they have at Paul's math notes --GFauxPas 15:59, 22 March 2012 (EDT)

Enhancement for equation template?

Recent considerations of complicated equational derivations makes me wonder whether it might be a good idea to add two more columns to the equation template: lll and rrr, for columns further to the left and f.t.r respectively to ll and rr.

The idea is that we can then do things like:

\(\ds a\)

\(=\)

\(\ds b\)

\(\ds \land \ \ \)

\(\ds c\)

\(=\)

\(\ds d\)

\(\ds \implies \ \ \)

\(\ds e\)

\(=\)

\(\ds f\)

and put the $\land$ symbol into a separate column between the equations and the implication sign.

As it is, it is not obvious where the $\land$ symbol sits in the hierarchy.

The code would be:

 {{begin-eqn}}
 {{eqn | l=a
      | r=b
      | c=
 }}
 {{eqn | ll=\land
      | l=c
      | r=d
      | c=
 }}
 {{eqn | lll=\implies
      | l=e
      | r=f
      | c=
 }}
 {{end-eqn}}

Should be straightforward to do and will have no effect on existing equations.

Comments? I'm on the case if it gets approval. --prime mover 18:51, 22 March 2012 (EDT)

I've had situations where that would have been very useful. So I'm for it. --GFauxPas 18:57, 22 March 2012 (EDT)

It appears that two columns 'lo' and 'ro' putting operators between 'll' and 'l', and 'r' and 'rr', resp. would do the job in a, IMO, neater way, isn't it?

Also, when adding more and more columns, it's becoming apparent that some lines are going to be quite wide; probably wider than our 'short sentence' house style would deem appropriate. So in time, a more durable solutions needs to be crafted. --Lord_Farin 19:16, 22 March 2012 (EDT)

Keeping in mind that I am new to ProofWiki, either of the two approaches (i.e., "lll" and "rrr" or "lo" and "ro") would be useful in the proof I am working on. Dan Nessett 22:05, 22 March 2012 (EDT)

lo and ro or lll and rrr, either would work.

I take the point about wideness of equation blocks, but I disagree. The whole point of this is to make it possible to split up excessively long equations into shorter bits and put them onto more than one line. What we're providing here is an alternative to the temptation to do something idiotic like:

\(\ds a\)

\(=\)

\(\ds b \land c = d\)

\(\ds \implies \ \ \)

\(\ds e\)

\(=\)

\(\ds f\)

Yes, it's still possible to build equation blocks which look stupidly ugly, but that's the fault of the writer not the tool. --prime mover 02:13, 23 March 2012 (EDT)

In this particular example, it appears that the $\land$ sits (upon implementation) on the 'o' spot; not really a problem, but this makes posting the stuff a tad harder because this probably means that all of the eqn's have to be manually adjusted at more points. I strongly vote for the 'lo', 'ro' solution because it allows to align the stuff better (i.e., put the sign in 'lo'/'ro' a lot closer to the equation than it would sit when instead 'lll' and 'll' would be used). Oh, and I agree that possible misuse is not a reason for not implementing it. --Lord_Farin 03:13, 23 March 2012 (EDT)

I don't understand why assigning it the name lo or ro is different from assigning it the name lll or rrr. It's just defining columns in a table.

WTF I'll just do it. --prime mover 13:50, 23 March 2012 (EDT)

The point I was trying to make was, that in case fancy CSS stylesheets were used, it could possibly matter what the column is that contains the operator. But apparently this is not the case. --Lord_Farin 13:55, 23 March 2012 (EDT)

The underlying mediawiki is just a table, and similarly the html generated is also just a table. I've added both lll and lo and similar ro and rrr so we have freedom to choose. This is what it looks like: Congruence Relation induces Normal Subgroup. Look for the \land symbol. --prime mover 13:58, 23 March 2012 (EDT)

Symbol for $\sigma$-algebra

Here on ProofWiki, I've seen the symbols $\Sigma$, $\mathcal A$, and $\mathfrak A$ for a $\sigma$-algebra. Is one of these to be preferred over the others? –Abcxyz (talk | contribs) 11:43, 23 March 2012 (EDT)

I prefer $\mathcal A$, $\mathfrak A$ is hard to read; $\Sigma$ can lead to confusion with sum notation paramount in measure theory. In case an event space is denoted, I'd still prefer $\mathcal E$ over $\Sigma$. Not sure if anything is enforced as house standard (except that $\mathfrak A$, being in the 'fraktur' or gothic alphabet, is generally discouraged). --Lord_Farin 11:52, 23 March 2012 (EDT)

Fast glib answer: it's a sigma-algebra, it should be denoted by a $\Sigma$. --Alec (talk) 23:31, 23 March 2012 (EDT)

Seconded. $\Sigma$ is also used in probability theory (at least from the sources I managed to find to cite). --prime mover 04:28, 24 March 2012 (EDT)

Okay, I yield. --Lord_Farin 05:11, 24 March 2012 (EDT)

Are we starting a project to replace all of the $\mathcal A$ with $\Sigma$? --Lord_Farin 14:03, 26 March 2012 (EDT)

Not consciously. It's not a high priority in my opinion. --prime mover 16:37, 26 March 2012 (EDT)