Talk:Main Page/Archive 9

From ProofWiki
Jump to: navigation, search
Filing cabinet.png
This is an article of past discussions, from 11-Oct-2012 to 18-May-2013.
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.

Contents

We've gone social.

I've just setup a Twitter and FaceBook accounts for ProofWiki. If anyone is interesting in helping out with the facebook page let me know. --Joe (talk) 15:40, 4 October 2012 (UTC)

We've already had a comment on Twitter appreciating our Template:Handwaving ... Sheer class. --prime mover (talk) 16:53, 4 October 2012 (UTC)
Should anyone find or post locate something worthy to share with the world, don't hesitate to request a tweet. I for one have just posted about Group Example: inv x = 1 - x. --Lord_Farin (talk) 21:04, 10 October 2012 (UTC)

Spam

  • Changed to use reCAPTCHA. Hopefully this will solve some spam issues. --Joe (talk) 11:14, 10 October 2012 (UTC)
... and instantly a spam posting appears.
Interestingly the precise same message from the precise same username (Jeremykippler728) has just been deleted by me from another website I assist in moderating. That website has a horrifically bad problem with spammers, and (mow I think about it) may be a major contributor towards my chronically bad temper. --prime mover (talk) 20:13, 11 October 2012 (UTC)

I'm not sure what else to be doing to stop the spam. --Joe (talk) 20:15, 11 October 2012 (UTC)

Can you go back to what we had before yesterday? It was considerably more effective than reCAPTCHA which is showing the same weaknesses as I'm having serious problems battling elsewhere. --prime mover (talk) 22:30, 11 October 2012 (UTC)

I think I've mitigated the spam problem for the time being. Changing the question we were asking seems to have done the trick. It's not a question anymore, just a copy the following "blah...." into the provided box. --Joe (talk) 01:37, 22 October 2012 (UTC)

Xymatrix

Allow me to draw your attention to this fantastic package again. Through the hard work of sonoisa, the xymatrix commutative diagram drawing package is now implemented.

A few months ago, the more powerful (but also more difficult and cumbersome) back-end language for xymatrix, xypic, was introduced, which I have employed several times.

But now, with xymatrix, I feel that most PW users should be able to draw the CDs they need. Documentation for xymatrix is easily located through a web search, but I find http://www.jmilne.org/not/Mxymatrix.pdf to be a good source.

The package can be used inside a xy environment, like so:

[[Axiom:Modus Ponendo Ponens]]

I have adopted the convention to precede such diagrams with two colons :: so as to place it a bit more centered in the flow of text.

In case you find functionality of xymatrix that is not yet supported by the implementation, notify me; I'll pass it through to sonoisa. That's all for now - happy editing. --Lord_Farin (talk) 19:52, 20 October 2012 (UTC)

I'm currently not in the area of mathematics which requires me to use this - but I will give it a workout once I need to. What I see of its results looks good. --prime mover (talk) 05:23, 22 October 2012 (UTC)
Should the basic reference above be insufficient for your purposes, then check out this, or as a last resort (very technical but complete) this. --Lord_Farin (talk) 17:35, 22 October 2012 (UTC)


I am still sorting these issues out for myself, but does this mean, in particular, that you implemented commutative diagram capabilities by installing XyJax 'extension' of MathJax. Or, instead, are you running xymatrix as a math system separate and apart from MathJax, such that the latter is used to do most math typography, but for commutative diagrams, a separate system is used ? Thanks and regards --Pickle (talk) 04:09, 7 January 2013 (UTC)
We have the XyJax extension installed - it supports the \xymatrix command now. How to implement the extension is explained at its web site. --Lord_Farin (talk) 08:42, 7 January 2013 (UTC)
Noted and many thanks. --Pickle (talk) 13:02, 7 January 2013 (UTC)
Followed the instructions and I have mathjax running locally on a local mediawiki (hard drive). Also, the XyJax sample runs fine when I click on it. I have placed the fp.js and xypic.js files in their places as described. However, I can not render XyJax (using the code for the same commutaive diagram) on the MediaWiki, even though the rest of my MathJax code renders just fine in the MediaWiki. I suspect (given very limited technical skills) that this is because I am not making any use of the script code as given in step 6 of the XyJax install instructions (https://github.com/sonoisa/XyJax). The reason for that is that I have no idea where to put it. Obviously, I'd like to place somewhere so that I need use only once as opposed to placing into every page, but unfortunately, I have no clue as to how or where to place this in the case of a mediawiki install. Would be grateful for any advice on this issue since I believe I am very close. Thanks very much! --Pickle (talk) 03:58, 15 January 2013 (UTC)
I'm afraid I don't know how to do such a thing with MediaWiki. Sorry, I don't think I can help you with this - you will probably find adequate assistance on some MediaWiki help thread, though I can't point you to one of those. --Lord_Farin (talk) 07:44, 15 January 2013 (UTC)
May be I am way off base in how I'm approaching the whole thing (I wouldn't be surprised), but how are you running XyJax thru MathJax for ProofWiki ? I followed the directions for installing XyJax with the exception of step 6. (Is MathJax running locally or remotely via the CDN. And, for XyJax, I thought you needed to move those two .js files mentioned above (per the XyJax directions) into the two subdir's of the MathJax distribution. Is there a different way to activate XyJax in MathJax. As I say, the test file for XyJax works just fine when I click on it. And, I have MathJax installed per the directions given on the MediaWiki discussion on the topic.) Thanks for any insight or thoughts about your experiences in XyJax up ! --Pickle (talk) 12:22, 15 January 2013 (UTC)

Our setup currently comprises a local MathJax copy, amended with the XyJax files in the locations as suggested by step 6. I think MathJax supports configuration files and extensions being located elsewhere; this should be documented in its manual. It may even be possible to use the CDN version of MJ while using local copies of the XyJax files. As for the functionality of the extension, I think it's simply amazing, especially considering it's one man's work. Whenever I encounter problems with it, I send an e-mail to the developer about it; usually it's then resolved within the next few weeks. I'd recommend it to anyone desiring vector graphics CD support in browsers. --Lord_Farin (talk) 16:09, 15 January 2013 (UTC)

Thanks. I finally got it to work. The trick was adding to the .js code of the MathJaxConfig file the elements described in step 6. The particulars came directly from the developer who was very helpful. Thanks again for your help. Best regards --Pickle (talk) 05:03, 17 January 2013 (UTC)

Equation template amendment

I have amended the eqn template to put the "ro" column to the left of the "r" column so as to be able to build multi-line expressions on the right with greater effectiveness. See this in effect on the $> 1$ proof in Minkowski's Inequality for Sums (incidentally up for a major refactor). --prime mover (talk) 11:23, 4 November 2012 (UTC)

Axiom namespace

It occurs to me that Definition:Group Axioms, Definition:Open Set Axioms and also potentially the axioms for a Definition:Synthetic Basis should perhaps all be somewhere in the Category:Axioms namespace.

Of course, that will mean the Sysyphean task of changing Definition:Group Axioms to Axiom:Group Axioms and all the appropriate pages (and also for Open Set Axioms but there are little links for that yet).

Anyone else think that's a good idea? Oh please say no, my fingers hurt. --prime mover (talk) 22:38, 5 November 2012 (UTC)

I'd say no. Almost every concept defined introduces its "axioms". You have probably noticed yourself referring to them as "conditions" as well. I can't craft a definitive argument, though. It seems so natural to have Peano's axioms in the Axiom namespace. It also feels natural (to me) to not include the "totally ordered field axioms" there. Maybe the benchmark is that a certain axiom system should be more-or-less universal, or at least universally accepted as a possible axiomatisation (like e.g. Hilbert style deduction axioms for PropLog). Abstract algebra and topology seem to be less suitable for such axiom systems. --Lord_Farin (talk) 22:55, 5 November 2012 (UTC)
As is apparent from my answer, there is a case for this change. However, I think that we're better off tidying proofs, refactoring pages and eviscerating stubs than letting our fingers bleed for such an ultimately aesthetic change. --Lord_Farin (talk) 22:58, 5 November 2012 (UTC)
I have an idea PM if you're quite serious about your fingers hurting. Perhaps make something that could take "[ [D:Set union] ]" and turn it into "[ [Definition:Set Union] ]" (that could of course be extremely difficult I don't know).
I am quite confused as to what an axiom is in any given context due to conflicting sources. Some sources call proof rules axioms, others make no distinction between an axiom and an axiom schema and others further still call the defining properties of a class the axioms of some theory about that class. I understand all these senses and that the essential quality they have in common is that they are starting points but I wish there was a bijection between words and concepts. I will have to leave it up to you. --Jshflynn (talk) 01:29, 6 November 2012 (UTC)
No I'm sorry, you've lost me again: why would I need to turn [ [D:Set union] ] into [ [Definition:Set Union] ]? It already does say [ [Definition:Set Union] ] thruout. --prime mover (talk) 06:16, 6 November 2012 (UTC)
As far as I'm aware when creating an article you have to type [ [Definition:Something] ] over and over when you want to refer to a definition page. It might be easier if upon saving a page some program could turn [ [D:Something] ] into [ [Definition:Something] ] (you know like \R instead of \mathbb{R}). The idea being that it would do the transformation once you press the save page button (so people in the future would still see it as [ [Definition:Something] ]). It is just an idea though and may save no time at all to someone such as yourself. --Jshflynn (talk) 09:46, 6 November 2012 (UTC)
Interesting that you say that. I suspect it not to be too hard to implement this; I'll discuss it with Joe. In fact, MediaWiki already supports a similar feature, which I happily use: if you type, say [[Axiom:Proof by Contradiction]], that becomes Set Union. Make use of it :). --Lord_Farin (talk) 13:12, 6 November 2012 (UTC)
On the note of making shortcuts, if it is simple to implement them, can we make one for \operatorname? --GFauxPas (talk) 14:26, 6 November 2012 (UTC)
I see that the shortcut LF mentioned works for other namespaces too (even nonexistent ones should you need to implement them in future). How delightfully efficient :)

Changing letters

I have recently seen that some editors prefer the use of the symbols $S$ or $A$ rather than $X$ to denote topological spaces and the like, because "$X$ tends to denote an unknown." I have no objection against this, but I cannot comprehend why this is an improvement. Would anyone mind to explain this (somewhat) more thoroughly? --abcxyz (talk) 00:06, 18 November 2012 (UTC)

Whim. --prime mover (talk) 08:45, 18 November 2012 (UTC)
It's good to have consistency of presentation in a wiki. When the topology work was started, a lot of the original work had $X$, a lot had $A$ and a lot had $S$, depending on who posted up what. Either we leave it like it is, with lots of different symbology, or we make an attempt to bring it into line. I've been working hard on generally trying to bring consistency to the site, and it cna be frustrating to find that what I'm doing has been undone. --prime mover (talk) 09:13, 18 November 2012 (UTC)
Fair enough. I haven't bothered to go through all the topology pages and see which symbols are more common than others.
Also, I have a question about linking to redirects (such as {{ProofWiki}} as opposed to [[File:PrimeMoversOffice.jpg]]). Would anyone mind to explain why this is done? I just don't understand it. --abcxyz (talk) 19:13, 20 November 2012 (UTC)
It is mainly for two reasons:
  1. Prettier text when hovering over a link
  2. In the case a page gets refactored in the future, only the redirect needs to be adapted; this saves mountains of work
I'm sure you understand now. As far as I'm concerned, the second argument is the more important of the two. --Lord_Farin (talk) 19:38, 20 November 2012 (UTC)
Thanks; I hadn't thought of that. It appears to be a good idea. --abcxyz (talk) 21:10, 20 November 2012 (UTC)

clock

Any chance of getting the system clock reset? It's 5 minutes or more fast. --prime mover (talk) 23:05, 19 November 2012 (UTC)

I shall see what I can do! --Joe (talk) 22:23, 20 November 2012 (UTC)
Fixed. --Joe (talk) 13:25, 21 November 2012 (UTC)
Much better. I no longer panic when I see the time on a page I've just edited. Thx. --prime mover (talk) 08:29, 22 November 2012 (UTC)

Of vs. for (moved from Talk:Equivalence of Definitions of Direct Limit of Sequence of Groups)

We now have Equivalent Definitions for T5 Space and this page Equivalent Definitions of Direct Limit of Sequence of Groups. There are various other names for this kind of page (e.g. "Equivalence of Definitions of ..."). We need a consistent convention. What will it be? I like "Equivalent Definitions" better than "Equivalence of Definitions" (it's shorter). --Lord_Farin (talk) 18:22, 26 November 2012 (UTC)

Pedantically, "Equivalence of Definitions ..." is more accurate, as "Equivalent Definitions ..." suggests that the definitions themselves are being introduced on that page - whereas the definitions themselves are in fact elsewhere. I confess I have not thought it through; but were we to use the latter it ought to be "Equivalent Definitions of ..." not "... for ..." (when I was refactoring over the weekend, "for" felt wrong) but I'm not in the mood to change them all at the moment.
If someone else were to volunteer to do the work, then making everything either "Equivalence of Definitions of..." or "Equivalent Definitions of..." would work for me. --prime mover (talk) 18:50, 26 November 2012 (UTC)
The new structure with definitions on subpages makes it much easier to have these pages adhere to a common standard, and to ensure that updates on the def. pages get reflected properly on other pages. --Lord_Farin (talk) 13:55, 27 November 2012 (UTC)

What to do about bad work

Some of the work on this site is shockingly bad. This has got to stop.

I recommend that people who post up unacceptable pages get punished for it. I recommend a fine of up to and including 1000USD and a ban from posting for up to and including life.

That's the only way we can get the quality of submissions up to a decent level. --prime mover (talk) 07:09, 27 November 2012 (UTC)

I don't deem it commendable to expel new enthusiasts adding a proof for a previously stubbed article because their writing style is not as we demand it. I do however agree that users notorious for posting substandard material (not that I know many) should be warned and if necessary restricted to their own user page until they demonstrate to know how to make stylistically bearable contributions. I fear that a too repressive regime (and I'd say we're already pretty relentless, only not finding the time to display it) would scare off (highly desired) new contributors, thus limiting the influx of information unnecessarily. --Lord_Farin (talk) 13:55, 27 November 2012 (UTC)
Bad formatting and lack of attention to appropriate writing style is not a problem. It's bad mathematics that I was suggesting needs to be eliminated. --prime mover (talk) 19:41, 27 November 2012 (UTC)
Would you care to give a rather more objective description of what you consider 'bad'? Plainly false, or incoherent, or phrased in specialists' language? --Lord_Farin (talk) 19:47, 27 November 2012 (UTC)
In particular, invalid arguments like affirming the consequent. Incoherent as well. Specialists' language is fine, it just needs someone to come in and say "This means ..." and so on. --prime mover (talk) 20:17, 27 November 2012 (UTC)

Repeating the hypotheses

When I was at school and my perspiring perpetually-defeated teachers tried vainly to bludgeon the basics of mathematics into the solid lump of bone that passes for my skull, one of the rules actually caught hold. That one is: when you're answering a question, it makes sense to repeat the basic facts of the question at the top of your answer so you have a precise grasp of your hypotheses.

In this context, I believe it does no harm to repeat the hypotheses of the theorem within the body of the proof (at the top) in order to clarify the statements which hold at any particular time.

It does no harm to repeat the hypotheses in the proof. OTOH, I think it would be bad to insist that "All hypotheses need to be repeated at the beginning of all proofs and all new work is to be abandoned until every proof adheres to this commandment". But there's a case for saying: if a hypothesis has been repeated at the top of a proof, it does not necessarily improve the proof by removing it just because you can. --prime mover (talk) 19:48, 1 December 2012 (UTC)

Fair enough. I'll take this bullet. --Lord_Farin (talk) 20:10, 1 December 2012 (UTC)
OK, fair enough. I removed the phrase to make the proof easier to read; it was at the end of a rather lengthy sentence. (Not that you said I "removed it just because I could", but this is why.) Moving it to the beginning of the proof also does the job. --abcxyz (talk) 21:57, 1 December 2012 (UTC)
Same thing happened somewhere else I think, I misremember exactly where. No worries. --prime mover (talk) 22:30, 1 December 2012 (UTC)

Differentiation and integration

Is there already a proof that the derivative of the integral of $f$ is $f$, i.e., that differentiation and integration are inverses? — Timwi (talk) 12:29, 3 December 2012 (UTC)

Cf. Fundamental Theorem of Calculus (first part, I think). --Lord_Farin (talk) 12:31, 3 December 2012 (UTC)

On results for concepts that have multiple definitions

We have recently decided to try and treat all (sourced) definitions for a concept on equal footing. An example of how the site is progressing can be found at Definition:Connected (Topology)/Topological Space. In time, Definition:Exponential will go through a similar remodeling, and I hope to add other descriptions of a Definition:Boolean Algebra. This brings me to the following point, concerning how results on such "multi-defined" concepts should be structured. As it stands, we have pages like Exponential as Limit of Sequence; this will of course become irrelevant given that we may define the exponential in terms of this limit. Such pages therefore naturally move to the associated Equivalence of Definitions result.

Up to now, such has not posed a problem. However, in the case of Boolean algebras, we have results of the kind Meet is Commutative, or, more relevantly, Join Distributes over Meet. In order to prove that a BA may equivalently be defined as a complemented, bounded, distributive lattice, it is then required to effectively prove Join Distributes over Meet to establish the equivalence. OTOH, given that a BA is a distributive lattice, the result is not even a result, but plainly imposed. Because there are rather a lot of such results involved in proving equivalence of the definitions, there are two, both unappealing paths that I see:

Firstly, that the pages remain in place, and have a "this is by definition" one-line proof (which seems rather irrelevant).
Secondly, that the pages become lemmata on the Equivalence of Definitions page, thereby unnecessarily lengthening that page (and I can foresee it will already be quite comprehensive from itself).

Of these options, neither of which is satisfying my elegance standard, I'd prefer the first, if one were to be chosen. We are also approaching a similar issue with the natural numbers. There are a lot of results which either prove basic facts about certain paradigms, or prove partial equivalence of definitions, or prove things by means of one paradigm that are trivial in another (without mentioning the equivalence).

It will benefit the consistency and completeness of ProofWiki if we pursue a meta-structural approach that treats, to the extent possible, definitions on an equal footing. In the long run, this will:

Make proving stuff easier, because the paradigm(s) most suitable to a typical can be invoked to achieve results that are nearly infeasible to prove directly from other definitions.
Make ProofWiki more appealing to all mathematicians, since there is no need for abandoning preferred frameworks and mindsets.
Make the site easier to navigate, and make it easier to locate what one is looking for; to this end, the "Also known as" sections should be as comprehensive as possible.

While I am convinced (as is apparent) that this is the situation ProofWiki should pursue, I think it is of the utmost importance that all editors are aware of, and should be able to comment on, this new direction, before we start rolling back each other's edits and create a hostile environment where the covering of mathematical knowledge is no longer the paramount goal of every editor. Thus, please comment. --Lord_Farin (talk) 14:10, 4 December 2012 (UTC)

Completely agree.
There are two types of definiend here:
One, e.g. Connectedness, which is one particular concept which can be treated monolithically; equivalence is proved "in one step", so to speak.
The other type of definiend exemplified is that of natural numbers, which is defined by means of an axiom schema - and in this case there are several approaches. In this case, what is done is for separate threads to be established, each one aiming towards the "defining properties" of the natural numbers - and indeed, some of the proofs along the way may be trivial.
The transition from the definition-and-equivalences to the multiple-definitions paradigm can be fraught with trouble if other editors are simultaneously trying to "correct" stuff which is still in the process of being developed, but if we are rigorous about using (and respecting) templates indicating work in progress / refactoring / etc., and perhaps adding a field for a username to those templates which do not have them, then inter-editor friction should be kept to a minimum.
I would also exhort newer editors to learn the ropes before being too strident in their opinions of how they think things ought to be. Many of the structural conventions have evolved over time and (despite appearance to the contrary) do actually make sense. Some MediaWiki tools and MathJax constructs, for example, are specifically not utilised because they have been shown to cause problems (although with more recent releases of the s/w some of these may well have been resolved).
In short, if new editors find their work is being systematically altered, tagged, refactored etc. etc. then treat it as a learning experience as to how pages are to be styled rather than as a challenge to their creative abilities. They are invited to take note of the changes being made and adjust their style accordingly. There are reasons. --prime mover (talk) 15:03, 4 December 2012 (UTC)
As to that last point of PM, I'd like to stress that anyone who doesn't understand a certain edit is urged to ask the relevant editor for a justification. In this way, we are both ensured that people get the rationale behind paradigms such as house style, and that such rationale continues to exist in the first place. All editors are invited to put up examples of possible changes/advances in house style or otherwise on their user sandbox, and invite others to comment. This seems to me like the appropriate way to try and nudge the rules and regulations for editing in one's preferred direction (as opposed to simply posting pages in the newly desired format in the main wiki). --Lord_Farin (talk) 22:13, 4 December 2012 (UTC)
Sounds good. I have a question about proofs (e.g. Synthetic Sub-Basis and Analytic Sub-Basis are Compatible) that use multiple definitions of a concept (in that case, the Definition:Topology Generated by Synthetic Sub-Basis). Evidently the equivalence proof has to be mentioned somewhere. I can't think of a way to deal effectively with pages such as those. Any ideas or comments?
Less importantly, does anyone find the phrase "the definition" strange, when one of the multiple definitions is being invoked? --abcxyz (talk) 16:52, 16 December 2012 (UTC)
Obviously what's happened is that the proof was written before the refactoring was done on the definition page to split it into multiple definitions. Equally obviously, the proof needs to be rewritten accordingly. --prime mover (talk) 18:22, 16 December 2012 (UTC)
Both definitions were still invoked (before the refactoring), so I don't see how the refactoring matters. (This was always a problem; I just didn't bring it up until now.) --abcxyz (talk) 18:37, 16 December 2012 (UTC)
So I haven't got round to sorting it out yet. My fault. I will try to do better. I understand my employment is up for review. Please feel free to vote on whether I get to continue to be allowed to contribute. --prime mover (talk) 21:33, 16 December 2012 (UTC)
No, it's not your fault one bit. I posted the proof and I want to know if anybody has an idea about how to make it work better. --abcxyz (talk) 21:38, 16 December 2012 (UTC)
Link to the specific subpage holding the definition you need in order to provide the required property that will move your proof along? --prime mover (talk) 21:54, 16 December 2012 (UTC)
Sorry, what required property? --abcxyz (talk) 23:35, 16 December 2012 (UTC)
Whatever property of whatever object you are using whatever definition of to prove whatever theorem you're writing a proof of. --prime mover (talk) 06:51, 17 December 2012 (UTC)
If I'm understanding you correctly, I think I did that already. The problem is that both definitions are used in the proof. --abcxyz (talk) 04:23, 18 December 2012 (UTC)
So link to one definition at the point at which you use that definition, and link to the other definition at the point at which you use that one. --prime mover (talk) 01:04, 26 December 2012 (UTC)
That's exactly what I did. So that's enough because the equivalence proof is in the "Also see" section of the definition page? --abcxyz (talk) 16:37, 8 January 2013 (UTC)

Enough for me, at least if you refer to "Definition X of ...". --Lord_Farin (talk) 17:53, 8 January 2013 (UTC)

Like this? --abcxyz (talk) 18:00, 8 January 2013 (UTC)
Yes, that's what I meant. In this way, it is apparent to the reader that there are multiple definitions, which may be overlooked otherwise. --Lord_Farin (talk) 18:02, 8 January 2013 (UTC)

Sine and Cosine of Sum

I split this (and all its subpages) into Sine of Sum and Cosine of Sum. Please let me know how I did! — Timwi (talk) 15:00, 13 December 2012 (UTC)

What should we do with the page Sine and Cosine of Sum now? Delete? Redirect to one of the two (which is what it does now)? Turn into a disambiguation page? The latter two seem silly, but the first one breaks existing links. — Timwi (talk) 15:00, 13 December 2012 (UTC)

From what I see, it looks good. Tasks still to be done are to make all links to Sine and Cosine of Sum point to the appropriate page and to turn Sine and Cosine of Sum into a transclusion amalgamate of Sine of Sum and Cosine of Sum. --Lord_Farin (talk) 15:19, 13 December 2012 (UTC)
Is that really necessary?... — Timwi (talk) 15:31, 13 December 2012 (UTC)
Such pages form convenient entry points to PW for wandering people doing web searches. While they are not strictly speaking necessary, they serve largely to disclose PW to incoming traffic. --Lord_Farin (talk) 15:33, 13 December 2012 (UTC)
OK well, I did the “simple” thing of just transcluding the two (see Sine and Cosine of Sum now), but it means it doesn’t have the proofs, and getting to the proofs is unobvious (you have to know that clicking the link in the header takes you to a proof...) — Timwi (talk) 17:38, 13 December 2012 (UTC)
If there's an existing page split into several, it is never a good idea to delete such a page. As you say, it breaks links. And even once the internal links are all amended, there may be external links which access this site from all over the place which would suddenly arbitrarily fail, particularly for old pages. It is also usually a bad idea to leave it as a redirect, because e.g. "blah blah sine and cosine blah" suddenly going to page containing "blah blah sine blah" etc., the reader is going to justifiably feel annoyed at not finding the cosine there. This contributes towards the generally negative view in which this site is held in the great wide world outside. --prime mover (talk) 17:41, 13 December 2012 (UTC)
We'll have to find a way to make clear that clicking links will take people to proofs, either now or some time in the future. As it stands, several people (rightfully IMHO) have indicated that the system isn't self-explanatory for non-veterans. Ideas are most welcome. --Lord_Farin (talk) 17:45, 13 December 2012 (UTC)

Separation of Proof and Theorem pages

Take for example Arctangent Logarithmic Formulation.

How necessary is it to put the proof on a separate page from the statement?

My idea was that we put the theorem and proof on the same page, except where we have multiple proofs in which case we separate out the proofs into different pages.

Have I been doing it wrong all this time? --prime mover (talk) 07:00, 14 December 2012 (UTC)

To the best of my knowledge, no. No need for duplication unless it allows for proper referral to knowledge. Your current approach to refactoring is reasonable and should be retained. --Lord_Farin (talk) 11:44, 14 December 2012 (UTC)

Template:Refactor

Personally, I’m quite confused by the fact that you seem to use the same template for two completely different meanings:

  • “Someone is currently working on this page, please don’t edit it”, and
  • “A change to this page is desirable, please edit it”.

I think we should split this into two to distinguish these two meanings. — Timwi (talk) 11:51, 15 December 2012 (UTC)

And for the please edit it one, I think the template should mandate that whoever put the tag on the page should explain what it is that needs refactoring, otherwise the tag is not useful. — Timwi (talk) 11:51, 15 December 2012 (UTC)

My personal view is that until a person has a good overall feel for the structure behind the website, it would be preferable for that person to concentrate on adding new stuff rather than tidying up the existing structure. --prime mover (talk) 21:33, 15 December 2012 (UTC)
And how might one get such a “good overall feel”? By jumping in and doing it, not by staying away. — Timwi (talk) 22:37, 16 December 2012 (UTC)
The refactor template is to the best of my knowledge not used in the first sense. Template:Inuse is the thing supposed to do that, but it seems to have fallen into obscurity. As to the explanation: sometimes the work required is either blatantly obvious to its intended audience (site maintenance workers) or hasn't fully crystallised yet (in which case it functions like a stronger version of Template:Improve). Since until very recently, I and PM were the only ones in the "maintenance workers" group (he more so than I) there wasn't really a need. I'll try to explain things better in the future. --Lord_Farin (talk) 21:39, 15 December 2012 (UTC)
Perfect — in that case, I’ll be bold and remove the first sentence from the template. If you disagree, feel free to revert it and then perhaps we should move this discussion to Template talk:Refactor. — Timwi (talk) 22:37, 16 December 2012 (UTC)
Reverted it, and protected so that only editors with admin privileges can edit it. Sorry, but it was correct as it is: it has either been identified as a candidate for refactoring, or it is in the process of being refactored. That is how it is being used.
The bottom line is: if there is a maintenance template whose purpose you do not fully understand, then it would be better to leave it alone. Basically, such a tag was not meant as an invitation to all to contribute, but more as a flag to a particular few who have been involved in what is an ongoing major restructure of the fundamental architecture on which ProofWiki is based.
TL;DR: thanks for the offer but we have the refactoring under control because we know precisely where we are going with it. Please work on other stuff. --prime mover (talk) 22:55, 16 December 2012 (UTC)

FAQs

The same questions (perfectly legitimate and sensible) are being asked by different contributors concerning things like house style, philosophical approach, structural conventions and all sorts of non-obvious and seemingly arbitrary stuff which we impose on the entries on this site.

It might be a good idea to bag up some of these, e.g. "Do we really need to put \left and \right round every parenthesis and if so why?" and "Why did you delete my link to Wikipedia in the citations?" and "Why did you change my favourite notation?" and publish them on a FAQ page - so if they get asked again, we can point the asker to that page.

Thoughts? --prime mover (talk) 00:11, 16 December 2012 (UTC)

I think this is a good idea! --Joe (talk) 12:31, 16 December 2012 (UTC)
Okay - I don't feel like working on it immediately, but as and when questions crop up that have been asked before, I will probably start building such a page. --prime mover (talk) 13:19, 16 December 2012 (UTC)
I don't know why we didn't have this page earlier! --Joe (talk) 15:02, 20 December 2012 (UTC)
May I recommend implementing a system to uniquely identify the questions? This will lend itself to redirecting inquiries. --Jshflynn (talk) 18:36, 20 December 2012 (UTC)

Something wrong with the rendering of references

References rendering bug.png

Does anyone know why we’re getting the rendering bug shown in the screenshot on the right? I looked at the page source, it looks fine, so the bug must be with the parser or the template system... — Timwi (talk) 14:05, 20 December 2012 (UTC)

It's the parser, I think. Something is wrong with how it handles nowiki tags (e.g. around square brackets for references). Various talk pages are littered with my remarks on the matter. --Lord_Farin (talk) 14:27, 20 December 2012 (UTC)
Once the templates have been cleared it should render okay. Anyway, house style discourages footnotes. --prime mover (talk) 17:32, 20 December 2012 (UTC)
... oops: I notice that that was a page I wrote myself. Definitely needs a rethink ... I was infodumping a college course. --prime mover (talk) 18:19, 20 December 2012 (UTC)

Milestone approaching

We are 36 proofs away from the 7000th proof.

As this is a small number times a higher power of 10 (in this case a cube), this number would appear to have significance to a hypothetical race of intelligent beings which have a 10-based number system (probably because they have ten fingers or something silly and arbitrary like that).

On this website it is traditional to reserve such a number to be the ordinal of a particularly significant result. I therefore alert all active contributors to this site to keep an eye open for something significant which can be hooked onto this number, and make sure it is not something simple and/or trivial.

Newcomers: take a look in the "Landmark Pages" category for an idea of what (at least in recent times) sort of thing I mean. Note that at this stage it is not necessary to have a proof immediately available if you need some space to work up to it. --prime mover (talk) 23:03, 24 December 2012 (UTC)

My suggestion is Law of Large Numbers. It's the first significant result of probability calculus, and we don't have it yet. If we restrict ourselves to the Weak Law of Large Numbers, it should also be possible to prove the Lemmas that are required to prove the theorem. If you find that LLN is too trivial, then I'd suggest the Central Limit Theorem. --Anghel (talk) 18:44, 25 December 2012 (UTC)
If you happen to be sitting in front of your screen when proof 6999 appears, then you are more than welcome to post up LLN as the 7000th, or, if you prefer, CLT. :-) --prime mover (talk) 01:02, 26 December 2012 (UTC)
CLT was already up. I put up something quirky, although nowhere near being able to put a proof up at the moment. --prime mover (talk) 21:26, 31 December 2012 (UTC)

Duality

I have been thinking about a way to connect dual pairs of theorems like Infimum of Infima and Supremum of Suprema. What I've come up with most recently is something like Template:About, i.e. a sentence below the theorem title. It could be along the following lines:

The dual of this [theorem/definition] is Infimum of Infima. See Duality Theorem (Ordered Sets) for more information on duality.

where of course there are three parts ("[theorem/definition]" and the two links) that need to be changeable by the template governing this. Possibly we want to say "information on duality for [concept duality applies to]". Maybe the last two links may be governed by a standard setting like cat=ordset or s.t. along those lines. Thoughts? --Lord_Farin (talk) 18:23, 3 January 2013 (UTC)

a) How well-defined a concept is "duality" apart from it being an intuitive thing?
It can be made quite formal, by e.g. interpreting a definition in the Definition:Opposite Ordered Set (defined along Dual Ordering is Ordering), and then translating that statement back to the original poset. This process is described in more generality on Duality Principle (Category Theory) (although stated there for theorems, it can easily be extended to cover definitions as well).
b) I'm not happy about it being right at the top of the page - too much of that and the statement becomes in danger of getting "lost" so to speak. In the "Also see" section sounds more appropriate, where most of these duals already exist - suppose it does no harm to emphasise their duality. --prime mover (talk) 19:40, 3 January 2013 (UTC)
Meh. The Also see section is an option, but I feel this detracts from the importance of duality. OTOH, a dedicated section titled "Dual something" could easily turn into overkill. If there were a nice option to stylistically emphasize the dual theorem, along with a few accompanying words, that'd be great. Maybe we have to settle with the dedicated section nonetheless. --Lord_Farin (talk) 20:46, 3 January 2013 (UTC)


So, I've decided to delve into the task of covering duality (for ordered sets only at this point). Would we like to anticipate on similarities coming up in the (hopefully not too distant future) concerning duality for ordered sets, groups, lattices, boolean algebras, categories (incidentally the last covers all the others) and call everything "Dual Blabla", i.e. "Dual Ordered Set", "Dual Ordering" (currently Definition:Opposite Ordering), "Dual Group" (Definition:Opposite Group), "Dual Boolean Algebra". Definition:Dual Category is already up. I have a feeling that this path will contribute to a more cohesive feeling for the whole site. --Lord_Farin (talk) 19:02, 4 January 2013 (UTC)
Can I just leave it up to you to see what it looks like? Maybe when I see it in action I'll like the look of it ... you know me, anti any idea that gets floated in a talk page, but once I see it working I usually quite like it. --prime mover (talk) 21:07, 4 January 2013 (UTC)

Question regarding use of copyrighted sources

I was wondering about sourcing proofs from copyrighted works, and what is permissible. Are there any guidelines on this from a legal perspective? To what extent can a proof be copyrighted? If a proof can not be copied in verbatim, to what extent must it differ from the original to avoid infringement? --Icarus (talk) 16:10, 5 January 2013 (UTC)

I'm afraid it's hard to put a definitive line in the inevitable greyscale area between "copied verbatim" and "rewritten in own words". However, if it's not self-evident that there are no legal issues, you're probably pushing it. That should be a safe guideline to work with. --Lord_Farin (talk) 20:12, 5 January 2013 (UTC)
I was curious, because I spent a significant amount of time working through Rudin's 'Principles of Mathematical Analysis', and writing out most of the proofs up through Chapter 9 with omitted steps and clarifying details filled in. A lot of this was done in LaTeX, and I would like to post it to this or a similar site, but I don't want to violate the copyright. When I asked the publisher about doing so, they denied me permission to use that much work. I guess one other way might be if they were willing to sell the copyright, and enough people were interested in contributing to purchase it in order to put it in the public domain. I don't know how likely either case would be though. --Icarus (talk) 21:57, 21 January 2013 (UTC)
The view I take is that the individual proofs, theorems and definitions which make up an elementary source work such as Rudin are not themselves subject to copyright. If one were to post them all up, one would note that practically all of it existed in other published works, in some form or other. You will probably find that the most of this work already exists on this site. All you therefore need to do (probably) is to add a citation in the "Sources" subsection to the specific paragraph in Rudin that contains the contents of the page. --prime mover (talk) 22:03, 21 January 2013 (UTC)
As a related side note, I wonder then if it would be useful, if it hasn't already been done, to have a page for each source that lists and links back to the proofs that reference it. --Icarus (talk) 22:15, 21 January 2013 (UTC)
The fundamental difference between what PM suggests and what you appear to have done is, that your work follows formulation and ordering of Rudin exactly (and subsequently expands upon it) and may thus be considered largely a derivative of orginal Rudin, while ProofWiki is an assembly of mathematical proofs formulated according to our own style premises - each on its own separate page. That such a work contains results coincident with Rudin is inevitable as they discuss the same matter (to some extent). This means it is permissible to add all the proofs you have explicated to ProofWiki and put them on the respective pages (which may or may not exist at this moment), while it won't be permissible to bundle them all and have the PDF distributed over the internet.
As for your note: Many of the sources that have already been covered to some extent at ProofWiki (and a lot of sources that haven't) are listed on Books. The "What links here" tool in the side bar provides a right inverse to the links in the Sources section. --Lord_Farin (talk) 22:19, 21 January 2013 (UTC)
If that is the case, I would be glad to publish it here as separate proofs. Do you know how I can legally verify that? Sorry, I'm just a little paranoid, especially after the publisher said that I didn't have permission to reproduce that much work.

The intent of Copyright Acts around the world is not to copyright ideas. It is to copyright particular implementations and renderings. Therefore, since we will undoubtedly fundamentally alter wording and presentation to match house style and our definitions, and hence not much but the idea of proofs will be retained, it is reasonable to conclude that doing so will not cause Copyright violation. Admittedly that's not a legal verification (for which I would need to be an expert in copyright law and jurisprudence, which I am not) but it's as close as I can get. I hope such suffices for you. --Lord_Farin (talk) 22:47, 21 January 2013 (UTC)

PS: I hit cancel on an edit and it seemed to post it anyway? --Icarus (talk) 22:39, 21 January 2013 (UTC)
Joe, can you look into this? --Lord_Farin (talk) 22:47, 21 January 2013 (UTC)
I'll dissable the fancy editor for now until I have time to do some more testing. --Joe (talk) 02:53, 22 January 2013 (UTC)

Oddness in Chrome

On my machine, when you try to edit a page in Google Chrome, it changes from (Courier New) font to some other font (some naff san-serif font that's all tiny and narrow and not equal-width, making it difficult to edit). It does this in all applications, not just ProofWiki. Even does it on Facebook.

Anyone experienced this? Only just started doing it. Version 23.0.1271.97 m.

I've sent a feedback note to Google in case it's something their end. Any advice? I lack the patience to dig around in the source code to debug it myself. --prime mover (talk) 20:02, 5 January 2013 (UTC)

I use the exact same version of Chrome, and I still get Courier New. --Lord_Farin (talk) 20:08, 5 January 2013 (UTC)
I think a solution is to add the code:
textarea{
    font-family: monospace;
}
to your user.css. This will force the textarea HTML environments (like the edit window is) to use a monospace font. --Lord_Farin (talk) 20:08, 5 January 2013 (UTC)
Doesn't work. As I say, it's not a ProofWiki issue, it does it all over, on all websites.--prime mover (talk) 20:20, 5 January 2013 (UTC)
Sorted it. In case anyone else has the same trouble, it turns out that I had Diigo bookmarked (it's an insidious little beast) which hijacks the stylesheet when in a textarea. Burn it with fire. --prime mover (talk) 22:44, 9 January 2013 (UTC)

Expressions involving restrictions

Just a quick note:

I noticed that the spacing of expressions involving restrictions can look somewhat awkward. For example:

f \restriction_A: A \to Y

is rendered as:

$f \restriction_A: A \to Y$

which doesn't look good. I think that a solution is to use braces, i.e.:

{f \restriction_A}: A \to Y

which is rendered as:

${f \restriction_A}: A \to Y$

I have already done this in several places. --abcxyz (talk) 18:48, 8 January 2013 (UTC)

Well spotted. Maybe this technical note could be added to the Definition:Restriction page as well? --Lord_Farin (talk) 18:55, 8 January 2013 (UTC)
Sounds like a worthwhile structural paradigm: perhaps another standard section name below everything except "Also see" and "Sources" called "LaTeX technicalities" or some such. The exact wording of that subtitle may need to be discussed. --prime mover (talk) 19:07, 8 January 2013 (UTC)
Definition:Convergence in Measure uses "Technical Note" which we deemed to be pretty good back then. --Lord_Farin (talk) 19:10, 8 January 2013 (UTC)
I would say however that the section should be below Also see, because we don't want to burden the casual reader with LaTeX information they may well find obscure and irrelevant, aside of incomprehensible. --Lord_Farin (talk) 19:12, 8 January 2013 (UTC)
As we progress down the page we go to concepts less and less directly relevant to the item in question. As an "also see" is not directly about the subject, I would argue that the LaTeX for presenting the subject is more directly relevant, so therefore deserves to be above Also see.
Vote? --prime mover (talk) 19:43, 8 January 2013 (UTC)
The "Also See" section is about mathematics; the "Technical Note" is not. Thus, I would prefer having the "Also See" on top. --abcxyz (talk) 21:45, 8 January 2013 (UTC)
I completely agree with Abcxyz. The sources section is naturally at the bottom; this quite uniquely determines the position of the section. --Lord_Farin (talk) 21:51, 8 January 2013 (UTC)
Yep okay, no worries then. --prime mover (talk) 22:01, 8 January 2013 (UTC)
I'll just add this: the rendering of MathJax is different from how it renders in the MediaWiki LaTeX package, and so we had to go through and change a lot of the rendering. As MathJax evolves, presentational awkwardnesses are fixed and fiddly extras are no longer required. So it's possible that a recommended technical implementation on one version of MathJax may not be optimal on a later version (e.g. at the moment it is advisable to add a half-space before primes on certain letters, $T$ for example, otherwise the prime overwrites the letter making it obscured - it is possible that this may be fixed in future MathJax versions). So a technical recommendation for presentational optimality may change during the course of evolution of this site. --prime mover (talk) 22:07, 8 January 2013 (UTC)


I have added a "Technical Note" section to each of the subpages of Definition:Restriction.

I also noticed that, on this site, the restriction symbol (e.g. $\restriction_A$) is used as a unary operation on the relation/mapping. However, $\LaTeX$ apparently treats harpoons as binary relations, as it does for other arrows (e.g. $\rightarrow$). This causes an excessive (for my taste) space between the relation/mapping and restriction symbols. Thus, I suggest inserting braces around the restriction symbol instead, which results in:

$f {\restriction_A}: A \to Y$

In addition, this looks similar to $\preceq \restriction_A$.

Regarding the use of \mathop, its behavior seems to be the same as that of \sin (as in x \sin y, for example). That is, it treats its argument as a unary operation. When it is not intended as such, it can have an undesirable effect, e.g.:

$\operatorname{id}_X \mathop{\circ} \left({g \circ f}\right)$

which is produced by the $\LaTeX$ code:

\operatorname{id}_X \mathop{\circ} \left({g \circ f}\right)

One $\LaTeX$ command that treats its argument specifically as a binary operation is \mathbin. I have used it in Definition:Restriction of Operation. It also can be applied to $\restriction$ as in $f \mathbin{\restriction} A$.

Correct me if I'm wrong on any of the above. Comments? --abcxyz (talk) 02:59, 10 January 2013 (UTC)

I'm (guardedly) in favour of replacing "mathop" with "mathbin" when the operator in question is binary, and using "mathop" instead for unary ones. As per current style, suggest the curly braces are used only when the operator itself is multipart, so e.g. you'd do f \mathbin \restriction A rather than f \mathbin{\restriction} A. This matches current usage of mathop and saves time when inserting new instances. --prime mover (talk) 06:13, 10 January 2013 (UTC)
Oh yeah, of course. Sorry, I just forgot. --abcxyz (talk) 16:51, 10 January 2013 (UTC)
Apparently the original suggestion to use "mathop" came from some $\TeX$ noob who encountered it on a quick web search. Braces around the harpoon seems a good idea because we want to treat it as a postfix operator. There is no dedicated $\TeX$ command for that, sadly. It may be that in certain rare cases the braces do not provide an adequate solution but the situations I can think of where this happens are very unlikely to be on PW, so we may decide to work with this (for the sake of brevity of source code). Thank you for your input on the matter, abcxyz. --Lord_Farin (talk) 09:03, 10 January 2013 (UTC)


Compare and contrast:

  1. \sum_{i \mathop = 1}^n: $\displaystyle \sum_{i \mathop = 1}^n$
  2. \sum_{i \mathbin = 1}^n: $\displaystyle \sum_{i \mathbin = 1}^n$

On my setup \mathop wins -- no contest. --prime mover (talk) 19:21, 10 January 2013 (UTC)

Yes, I know. But, as I said earlier, \mathop can sometimes create problems when the right-hand side starts with parentheses. For example:
\sum_{i \mathop = \left({m + n}\right)/2}^n i
renders as:
$\displaystyle \sum_{i \mathop = \left({m + n}\right)/2}^n i$
We would need to add \, after \mathop =, which leads me to think that we might as well use \, = \, instead of \mathop =.
By the way, I think that a $\LaTeX$ command for a relation is \mathrel, which inserts more space than \mathbin. As one might expect, that doesn't do anything because = already acts as a relation. --abcxyz (talk) 22:36, 10 January 2013 (UTC)
No no and thrice no. We do not want to go adding arbitrary amounts of whitespace into the formulae just to make it look good on the browser you happen to be using.
\sum_{i \mathrel = 1}^n: $\displaystyle \sum_{i \mathrel = 1}^n$
is indeed rubbish, but:
\sum_{i \mathop = \left({m + n}\right)/2}^n i: $\displaystyle \sum_{i \mathop = \left({m + n}\right)/2}^n i$
works fine.
Bottom line: mathop as currently being used is fine. IMO all this farting about is a complete waste of time and effort. --prime mover (talk) 23:02, 10 January 2013 (UTC)
The point about restrictions was valid enough, but I am inclined to agree with PM that with \mathop we have something that works, and therefore, since it's only a source code issue, stick with it; after all, it's not like we're adding lines of cumbersome code that could be replaced by a single character. --Lord_Farin (talk) 23:09, 10 January 2013 (UTC)
Okay, if you don't like inserting white space, fine. I'll just add that the unequal spacing before and after the = when preceded by \mathop also occurs on my personal copy of $\LaTeX$, not just on my browser. I'm curious why you're so adamant about not inserting white space (in this case). I can't immediately see why.
By the way, I replaced \mathop{\mathcal R} with \mathrel{\mathcal R} on Definition:Relation Compatible with Operation to make the spacing less lopsided (and a bit larger). Anything you'd like me to know about it? --abcxyz (talk) 04:31, 11 January 2013 (UTC)

How about we open a Technote namespace instead of putting them in definition space? --Dfeuer (talk) 04:44, 11 January 2013 (UTC)

How about we just do something useful instead? --prime mover (talk) 06:06, 11 January 2013 (UTC)

Parentheticals

The house style insists on $\left({blah}\right)$ in all cases rather than $(blah)$. There are three problems with this, ranked from most important to least important in my opinion:

  • It's much harder to read in source form, particularly when nested.
Try expanding the font size - that might help.
  • It takes longer to type.
Cut'n'paste works here.
  • It requires different structure for $\left\{{a} \,\middle|\, {b}\right\}$ than for $\left\{{a:b}\right\}$. --Dfeuer (talk) 22:15, 8 January 2013 (UTC)
The house style remains unchanged. --prime mover (talk) 22:22, 8 January 2013 (UTC)
Even if it takes longer to type, it's only by a couple of seconds (for me). As for your last point, I think \left\{{a \mid b}\right\} does the job. --abcxyz (talk) 22:24, 8 January 2013 (UTC)
\left\{{a \mid b}\right\} does not automatically size:
$\left\{{\dfrac a b \mid \dfrac c d}\right\}$ (yes I know it's meaningless mathematics)
as against:
$\left\{{\dfrac a b} \,\middle|\, {\dfrac c d}\right\}$
One of the main reasons behind the rigorous enforcement of \left{ ... }\right is purely because of the resizing. Cut'n'paste on a given construct is more straightforward, you don't then have to put the left-right in. The irritation about adding the spaces either side of the \middle is a nuisance, I agree - but it's minor because that construct is not often used. --prime mover (talk) 22:34, 8 January 2013 (UTC)
Y'all do know of the mightily useful option of right-clicking a piece of rendered TeX, going to "Show Math As > TeX Commands" to obtain a small window with the TeX generating that construct, no? (Admittedly I only found out one minute ago, but that's not the point.) --Lord_Farin (talk) 22:37, 8 January 2013 (UTC)
Occasionally use it, but (a) muscle memory has developed facility typing \left etc., and (b) when I'm editing a page with a lot of f \left({x}\right) etc. cut'n'paste is easy. --prime mover (talk) 22:42, 8 January 2013 (UTC)
Muscle memory FTW :). --Lord_Farin (talk) 22:43, 8 January 2013 (UTC)
Let's not get hung up on my third point, which is relatively trivial. $\left({ \left({ f \left({x}\right) + g \left({y}\right) }\right) h \left({z}\right) }\right) ^ 2$ is just inherently harder to read in source form (and thus to keep track of while writing) than is $((f(x)+g(y))h(z))^2$, and they look the same once processed. As for cut and paste, unless you're really quick with your custom emacs/vi macros, that's going to be slow too. I understand the motivation to make cutting and pasting easier, but just how often does one replace a standard-height expression in parentheses with a tall one? --Dfeuer (talk) 22:45, 8 January 2013 (UTC)
Not so much less often than one needs more than two levels of nested bracing (in TeX, that is; regular contributors know I can get pretty deep in my prose (such as this illustrative yet meaningless phrase)), I would say. While I do admit the paradigm is awkward at first, it works quite well (says this one, having been fully internalised into PW). I'm still not inclined to change my view - sorry if that comes across as stubborn. --Lord_Farin (talk) 22:53, 8 January 2013 (UTC)
Let's suppose you're writing $((f(x)+g(y))h(z))^2$ and you make a mistake miscounting the brackets (easy to do). Using your preferred style: ((f(x)+g(y))h(z)^2 you don't miss the missing close bracket in $((f(x)+g(y))h(z)^2$ but with this: \left({ \left({ f \left({x}\right) + g \left({y}\right) }\right) h \left({z}\right) ^ 2 it doesn't render: $\left({ \left({ f \left({x}\right) + g \left({y}\right) }\right) h \left({z}\right) ^ 2$.
Case rests. --prime mover (talk) 23:05, 8 January 2013 (UTC)

I personally get confused/annoyed by quantified expressions, disjunctions, etc., that aren't sufficiently parenthesized, but these become relatively deeply nested once parenthesized enough, at which point I find them unreadable under PW style rules. Take, say $\forall x \in A:((3(x+a)<4) \lor (x>17))$. However: I would frankly be much, much happier if we even had a very slightly laxer rule, not requiring innermost parentheses to be fancified, while requiring it for all others. As for the automatic error detection, as you're clearly using a fancy editor with fancy custom macros I'm sure your editor is quite capable of checking for mismatched parentheses. --Dfeuer (talk) 23:19, 8 January 2013 (UTC)

I'm not using a fancy editor with fancy custom macros. Those are for girls. --prime mover (talk) 23:21, 8 January 2013 (UTC)
On-topic, I think there may be a case for Dfeuer's suggestion. It seemingly combines the best of the two approaches. I pose the slightly adapted suggestion that innermost parentheses with a content of just one entity (like e.g. $x$, $\overline \R$ or $\operatorname{id}_X$) needn't be endowed with lefts and rights. --Lord_Farin (talk) 23:25, 8 January 2013 (UTC)
I disagree because then somebody's going to go through all the existing pages and change them. I don't think that would be helpful. --prime mover (talk) 06:10, 9 January 2013 (UTC)

onlyinclude

Is there any way to have different includeonly "classes"? Sometimes it would be neat to be able to have, say,

Theorem:

(onlyinclude class = theorem)

Detailed setup

(onlyinclude class=theorem-summary)

The good parts

(/onlyinclude)

Some less-good parts

(/onlyinclude)

Proof:

(onlyinclude class=proof)

Good, solid proving.

(/onlyinclude)

Miscellanea. --Dfeuer (talk) 23:10, 10 January 2013 (UTC)

Have you read the MediaWiki documentation? --prime mover (talk) 23:12, 10 January 2013 (UTC)
This annoyance was one of the main reasons for me to start developing an extension tailored to ProofWiki. Among other things, it allows precisely what you want. Cf. User:Lord Farin/Long-Term Projects/Extension. --Lord_Farin (talk) 23:15, 10 January 2013 (UTC)
If I were convinced of a need for such complexity I might be persuaded to come on board with this, but I'm sceptical. --prime mover (talk) 06:19, 11 January 2013 (UTC)
The need is clear. Sometimes you want to put together a list of theorems. Sometimes you want to put together a list of proofs. Say you have three theorems, each of which has two proofs. You want on page that transcludes the three theorem statements. You want three pages that each transclude two proofs. BLAM. --Dfeuer (talk) 06:34, 11 January 2013 (UTC)

Lax definitions

A question about complex analysis: all sources that I have consulted speak about a connected domain which they define to be "a path-connected open subset of $\C$". Some texts even omit the word "connected" and simply call the thing a domain. This definition is not in line with the topological definition of connected, where a connected set need not be path-connected, and certainly not open. Nor does it follow our definition of a domain. Of course, we could simply write "open path-connected set", but that would sound differently from the book texts. My suggested solution is to create Definition:Connected Domain (Complex Analysis), and then explain the lax definition in a note --- similar to what we have done with Definition:Smooth Path (Complex Analysis), which is not smooth. Thoughts? --Anghel (talk) 14:28, 11 January 2013 (UTC)

Sounds like an adequate solution to me. Ideally we would also loosen the assumptions on each theorem as much as possible, and then e.g. put corollaries for common stronger assumptions. This path, however, lies relatively unexplored at the moment, so maybe we needn't make it a priority. Having the results in the first place is good enough for the moment I think. --Lord_Farin (talk) 14:38, 11 January 2013 (UTC)


A Fond Farewell

Hello all,

As has probably been apparent from my gradual decline in editing/general participation over the last year or so, I've been finding it progressively harder to keep up with the shear volume of content and changes that have been going on in ProofWiki. I've finally decided that it isn't really beneficial to the site or to me to try to flip through the hundreds of edits a day (and that's after filtering out the ones I don't think will be important). In light of that, I'm going to step back from ProofWiki and let it move on without me. This doesn't necessarily mean that I won't add a proof here or there if I run across something worthwhile, but I won't be here on a daily basis anymore. If you need to get in touch with me about something, I'll still get email and my talk page will notify me of changes (to address the only specific question I've thought of that might come up: I have no objection to future changes in the specifics of the copyright/licensing stuff, as long as ProofWiki stays available to those who want it).

I've had a lovely four and a half years as a regular part of the ProofWiki community, and I've really enjoyed seeing all of the changes the site has been through and getting to play a part in them. It's almost ridiculous to look back on what the site was like that long ago; I thought just for fun I'd link to what the main page looked like when I joined: [1]. Back then we were excited about putting up proofs one through six, and I never would have thought I'd still be here at proof 7000. I was only the 22 user to join, and I guess of the semi-regular contributors, only Joe has been here longer, and he helped start this whole thing.

I think I've grown up mathematically along with ProofWiki. Back in the day when I first joined, I was still in high school, just starting to learn what math was really like. Now I'm looking towards the all too rapidly approaching end of my time in college, and while I know a lot more, perhaps even more importantly I realize how much more there is that I don't know yet. ProofWiki has helped teach me about the need for rigor and clarity in mathematics, and I'm sure I owe no small part of what academic success I've had to the practice I've gotten as a part of this community. Sorry for the nostalgic mini-essay here, but I wanted to say a last goodbye and thank you to all of you for helping make this site the incredible repository of knowledge that it has become.

If I have one piece of advice I'd like to leave you all with, it's to stay open to new participants and the ideas they bring. Of course they won't always be good ideas, and disagreements will inevitably come up, but if you approach them with an open mind and welcome into the community as warmly as I was welcomed over 1,700 days and 2,000 edits ago, they might become a valuable part of ProofWiki, they might learn something from all of you, and you might just learn something from them. I'll no doubt be stopping back in to ProofWiki occasionally to see how it's coming along, but I'm confident that the group of all of you that I'm leaving behind will keep up the good work for many years and many proofs to come.

--Alec (talk) 20:24, 12 January 2013 (UTC)

*Applause* You were an excellent editor Alec. You shall be missed :') --Jshflynn (talk) 22:35, 12 January 2013 (UTC)
It's been good. If only there were a simple way to ensure the structural integrity of the site continues to be maintained ... --prime mover (talk) 23:34, 12 January 2013 (UTC)
Although we haven't been involved as much as you and PM undoubtedly have been, I still consider you one of the most valued editors of this web resource. It would appear that at this point, we're hitting the inflationary stage in the growth of the site (considering that I joined only 15 months ago and witnessed the 5000th proof), so that it's understandable that you can't keep up with it as you have the last few years. I hope we all do our best to make you proud and honour your piece of advice. Hope to see you around sometime as well. --Lord_Farin (talk) 00:36, 13 January 2013 (UTC)
So long, and thanks for all the fish! --Joe (talk) 23:56, 17 January 2013 (UTC)

LEM

We seem to have a template to warn about LEM. That's great. However, we also have a slew of pages that rely on LEM without saying so, or that use it without necessarily needing it. If someone wants to try to actually keep track of what does and does not need LEM, may I suggest that we add a "no-LEM" template for theorems known NOT to rely on LEM? That way, we'll be able to tell the difference between "does not rely on LEM" and "no one's checked if this relies on LEM". Personally, I don't care about LEM (unless I'm dealing with a computer language whose type system doesn't support it), but clearly some people do. --Dfeuer (talk) 10:38, 24 January 2013 (UTC)

Far too much work. Results that use LEM are restricted to a) those in propositional logic, and b) those which are specifically dependent upon it like one of the proofs of Gelfond-Schneider. The reason why there are plenty of results which do use LEM that are not specifically flagged as using LEM is purely because the template was created like a couple of days ago, and it is taking me that length of time to go through the entire category and refactor it.
Incidentally I'm interested to know which pages specifically use LEM without actually needing it - I thought I'd made sure that wasn't the case. Please point out all such pages.
Is it just that you want me to do this work to add that template to every single page in PoofWiki that doesn't rely upon LEM (I mean FFS that's thousands) so that I've got plenty to do and it will stop me interfering with what anyone else is up to? If so, then I quite understand and will get on with it straight away.
Oh, and if you don't care, why are you raising this issue? --prime mover (talk) 11:02, 24 January 2013 (UTC)
My view on the point is that it should be taken up, if so at all, by people that *do* care about LEM. I have the impression this group of people is disjoint from the group of active PW contributors. It is thus not likely to happen in the near future. --Lord_Farin (talk) 11:10, 24 January 2013 (UTC)

Template:BookReference

As my current state of mind does not permit expanding the mathematics on ProofWiki, I have taken on the project of template documentation and general documentation guidelines (again?, some may say). I have written a comprehensive technical manual regarding the use of Template:BookReference and intend to start producing pages for other templates and support pages (such as Help:Editing/Creating Pages for Mathematicians or whatever I decide to call it). I hope this effort will lighten the ongoing burden on some of the editors pertaining to documenting sources, and referring to them, as well as allow editors so inclined to learn by themselves how to perform such tasks. I hope the style and presentation are comprehensible (and appreciated). --Lord_Farin (talk) 13:42, 24 January 2013 (UTC)

Incidentally, my work has brought to me the need for an option to indicate which edition of a particular source is used. This will reduce ambiguity and allow for accurate display of how recent a source is. I was thinking of something like replacing the year - currently e.g. "(1976)" - by "(2nd ed., 1976)" if the second edition is referred to. I think however that it is not constructive to cover multiple editions of a work; the main intent is to allow for this added detail that makes it more accurate to refer to year of publication. I envisage this new trinket to gradually be incorporated and do certainly not desire a marathon of changing all sources (that'd be insane). --Lord_Farin (talk) 13:47, 24 January 2013 (UTC)

We will need to work out exactly what to do with multiple editions in due course. For a work like 1978: Roland E. Larson and Robert P. Hostetler: Calculus which has multiple versions written by several people, it is probably worth documenting the different editions, although the reading trail will probably be done (if at all) for only one of these. For most books it is "probably okay" just to include one edition (naturally, the one on the contributor's bookshelf) but cite it with the date as for the first edition. It's not ideal, but it should work well enough; IMO the initial publication date is significant. Why? Because, among other things, it is interesting to see which concepts come into and go out of fashion. Some concepts are not referred to in books after, e.g. year m, while others do not appear until year n. And so on. You can get a glimpse into which works may have been used as source works for others, by how the flow in one directly parallels that in another. And so on. There is fuel for scholarly study here which may not have been gathered anywhere else. --prime mover (talk) 14:10, 24 January 2013 (UTC)
Then I think it's best to try and adhere to your stated rule for the moment and put the suggestion into the fridge until I've come up (if I ever will) with a way of ensuring both accurate reference and availability of year of initial publication. --Lord_Farin (talk) 14:40, 24 January 2013 (UTC)
I have finished the important parts of the manual - that is, the parts we use frequently. Comments are appreciated, before I proceed with replacing the current material. --Lord_Farin (talk) 14:44, 29 January 2013 (UTC)
I've been following it on and off - I'll give it a good workout presently, but it looks good - far more detailed than I would have the patience for ... --prime mover (talk) 16:35, 29 January 2013 (UTC)
Which is why I took several days of reduced contribution to write it up; the expected benefit helped a lot in finding the patience. Also, it's admittedly incomplete at certain points, but as stated, it's complete regarding the features most used and most relevant to newcomers. --Lord_Farin (talk) 18:32, 29 January 2013 (UTC)

Symbolic Logic: paradigm shift

The Axioms of Natural Deduction as defined here are basically taken from 2000: Michael R.A. Huth and Mark D. Ryan: Logic in Computer Science: Modelling and reasoning about systems (it was the first book that came to hand when I was first studying this subject). However, further books I have read, in particular 1965: E.J. Lemmon: Beginning Logic and 1964: Donald Kalish and Richard Montague: Logic: Techniques of Formal Reasoning use a slightly different set. Other books use a (subtly) different set again, and so on.

So specifying a particular statement of PropCalc as "absolutely" axiomatic seems a bit previous.

What I am considering doing is changing the namespaces of all those results whose names are e.g [[Axiom:Proof by Contradiction]] and [[Axiom:Modus Ponendo Ponens]] and so on, merely calling them Proof by Contradiction and so on.

Then, in order to bring us back to the concept of how Natural Deduction works, the various source works which do specify a set of axioms are used as bases for a set of axioms each, and the "Axiom" namespace will thereby contain a master page for each such set, each one containing / transcluding the specific results as defined as "axiomatic" in that source work.

This will inevitably require links to specific proofs (as opposed to specific results) in order to prevent circularity within a particular axiom space, even though e.g. LEM proves DNE and similarly DNE proves LEM - either one can be chosen as axiomatic, or neither (some sources using Modus Tollendo Tollens as an axiom instead).

So basically, this is a heads-up that the good old PropLog section is up for another seismic shift again. The plan is for the all the axiom schemata to be unified in a homogeneous manner.

If any person can show just cause why they may not be joined together – let them speak now or forever hold their peace. --prime mover (talk) 21:42, 28 January 2013 (UTC)

Since I support this kind of organization across the whole site, you won't get any objection from me. --Dfeuer (talk) 21:52, 28 January 2013 (UTC)
This will be a marvellous project once complete. It will be nice and modular (or so I hope) so that we can reasonably easily accommodate new approaches which will undoubtedly arise from covering more sources. Because the task is so monumental, we need to be very careful to ensure a smooth transition from the one state to the other. I would like to use this operation as an excuse to rigorously separate the proof theory and model theory branches of PropLog (the PredLog section will follow once we finish PropLog) because in the future we will not only want different approaches on the proof theory side (which is what you try to allow presently) but also on the model theory side (Kripke semantics, categorical semantics); the latter are important when considering nonstandard logics. I am willing to help out - I fear that we are up for a huge rewriting, almost as if we were to start from scratch. --Lord_Farin (talk) 21:56, 28 January 2013 (UTC)
The exercise to separate out every small result is progressing well. (This is in itself important as it makes sure that we can adequately isolate the specific threads of implication from others, despite the obvious-looking commonality between, say, $p \implies \neg p \vdash \neg p$ and $\left({p \implies \neg p}\right) \implies \neg p$.
I have started by moving "[[Axiom:Proof by Contradiction]] to Proof by Contradiction because that's probably the most work-heavy one, and I can iron out all the technical issues up front. Then the others will be more straightforward to process. --prime mover (talk) 22:49, 28 January 2013 (UTC)
Right, I've burned myself out by refactoring the entire PropLog space by replacing all lines in tableau proofs with template invocations. Those pesky abbreviations have been replaced with full names as appropriate, and a consistent approach has been taken with respect to sequent introduction. I may have missed some, but it's mostly complete.
For my next trick I will start construction of axiom schemata according to the various works I have available (there are a few) and make an attempt to match up where they match and where they differ, with a view to synthesising something out of it. But that may not be today, or even tomorrow. --prime mover (talk) 21:36, 31 January 2013 (UTC)
While at it, and I know I've asked this before but it's important to me for future plans for dashes into Model Theory, could you please see to a rewrite of Boolean Valuations and the whole mess around it? --Lord_Farin (talk) 22:11, 31 January 2013 (UTC)
Point me towards it. Can't find any pages referring to "Boolean Valuation" so it must be under some other name. --prime mover (talk) 23:56, 31 January 2013 (UTC)
In particular, Definition:Boolean Interpretation, Definition:Propositional Function, Definition:Propositional Function and whatever may be around those (results as well). --Lord_Farin (talk) 08:12, 1 February 2013 (UTC)

The problem here is in how to define such an object in terms which do not rely on the language of set theory to do so - as set theory itself requires the definition of a propositional function to get it off the ground in the first place (via Axiom of Replacement or whatever the axiom is). (Yes I know ZFC is not the only game in town, I just haven't had access to any such materials, particularly sources expounding paradigms which do not require this fundamental dependency.) --prime mover (talk) 09:02, 1 February 2013 (UTC)

It is impossible to circumvent the distinction between naive and axiomatic set theory. We simply cannot start with nothing. This does not pose a real problem, since it is intuitively clear that we can talk about "sets" (well at least as long as stuff is non-circular - Cantor's "set" simply doesn't exist because it is self-referent). We can suitably develop model theory in this naive setting, introduce ZFC (or any other approach to set theory) and then exhaustively formalise every construction used in naive set theory (which until now hasn't posed a problem to the mathematical community) by treating the concept "set" as "element of a model of ZFC". This leaves only the development of model theory as subject to discussion and possible confusion.
That said, my main problems with said stuff are due to the rather ad hoc presentational style which appears at best a bit dislodged from the rest of the site. A not worthless distinction would be that between set and meta-set (the latter corresponding to the naive notion), but since I haven't encountered that in any of the literature I've seen I don't consider that a good idea. Of course, once the whole thing is set up one can nest the construction, considering models of ZFC in a model of ZFC, rinse and repeat - this does however not add anything to the rigour IMO. The current Definition:Set is pretty good for a description of a "naive" set; since also Definition:Mapping is accurate, we should be fine. It's really only the presentation of said section that is up for improvement (the other pages could be amended with an elaboration on axiomatic set theory but that's for a later project). --Lord_Farin (talk) 09:51, 1 February 2013 (UTC)
I will make an attempt. Bear in mind that the PropLog section was among the work done on ProofWiki, and as such the style and direction had not completely crystallised. If the issues you have are primarily presentational, then feel free to experiment with those areas you have issues with. The Natural Deduction section will continue to be reskinned independently of the model-theoretic work, although it may yet be appropriate for the "Boolean Interpretation" and "Truth Function" sections of the e.g. Definition:Conjunction pages to be extracted into subpages, with a view to transcluding much as the Definition:Truth Table page has been presented.
I confess to being uncertain still as to exactly how to proceed in that area - so for now will just continue to work on the Natural Deduction stuff. --prime mover (talk) 10:09, 1 February 2013 (UTC)

Complex Analysis: assignment of responsibility

Anghel has now been put completely in charge of developing the Complex Analysis field on this site until further notice. Any amendments made by other collaborators are to be restricted to stylistic and cosmetic amendments only (of which few are expected to be necessary).

If you have any major questions regarding the direction of the documentation of this field of mathematics, particularly if you need the existence of a particular result or definition, place a comment on the relevant Talk page.

Anghel: please feel free to post your thoughts on this arrangement. --prime mover (talk) 14:57, 31 January 2013 (UTC)

I do not want that responsibility. I've stated my reservations here. Also, if I should drop dead tomorrow (or get a 9 to 7 job, which is equally likely), you would have a problem. Please, delete this notification from the talk page. This wiki should remain open. --Anghel (talk) 17:09, 31 January 2013 (UTC)
While I'm all for someone being " in charge" of a section, I think this should only be as a maintainer, not only contributor. Any user should be able to add anything they like. --Joe (talk) 17:14, 31 January 2013 (UTC)
I'm interpreting it like Joe. It is a good idea to let someone have the overview over an entire section so as to ensure consistent notation, terminology, etc. The site benefits from someone who knows the way in a field, what has and hasn't been covered, what inroads are up next and generally up to which level of complexity the whole coverage has advanced (which is a laughable level for most of the site - let's try and do something about that!).--Lord_Farin (talk) 18:34, 31 January 2013 (UTC)
The thing is this. When somebody is working methodically through a section of work, it is usually not a good idea for someone else at the same time to be looking at what is going up and deciding that the structure needs to be changed. There are always different ways of putting these things together, and if one person has a vision of something that works, then it is a bad move for anyone else to change things to suit the way they want it. As this potential danger appeared to exist at the time, I posted up what I did. --prime mover (talk) 20:46, 31 January 2013 (UTC)

Lord_Farin's backup, etc.

I just looked at the double redirects page, and I see that there are a number of redirects from Main into Lord_Farin's backup. Structurally, I don't think there should be links of any kind from Main into User. If a Backup area is necessary for pages moved from Main, I think it should go under Main. On the other hand, what are the chances of external links to material so incomplete or shaky that it had to be taken out of Main? --Dfeuer (talk) 19:58, 31 January 2013 (UTC)

Leave them, please. It's on my to-do list. --prime mover (talk) 20:41, 31 January 2013 (UTC)
Those moves were made because all the terminology and the whole approach to lattices and boolean algebras was being made rigorous. This obsoleted the results, but I wanted to keep them accessible in some way, hence the moves. --Lord_Farin (talk) 22:11, 31 January 2013 (UTC)

A proposed resolution of the naming problem

We are confronted time and time again with the problem that some perfectly valid results suffer from the lack of a properly descriptive name for them. There are usually two solutions practised to this issue: to post the result with a very inadequate or even misleading title; and to not post the result.

Both of these are really unacceptable and (while it be noted that I somehow feel a bit of credit to Jshflynn is due here - however, I couldn't locate the source of this thought) the solution that occurred to me just tonight is to effectively give them a number, something like "Lemma N". And while this is obviously not an elegant solution, it will allow us to post up, and refer to, results that haven't been assigned a proper name yet. Of course, as soon as a good and reasonably concise title is proposed, the page is to be redirected to that proper name.

Furthermore this will instigate the need of an overview page, listing all such "Lemmas" as well as good categorisation both on the Lemma pages and on the overview page (a bit like the Books section perhaps). I repeat: this isn't very elegant but it does take care of it, for now. More thought on how to keep the Lemmas accessible (so that there won't be five lemmas with the same content) would be appreciated. Finally, I'd appreciate any critique to be supported by arguments - rational ones, not purely emotional blubber. --Lord_Farin (talk) 19:38, 1 February 2013 (UTC)

I'm not sure what I think about the idea in general, but if we do it then the titles should have both numbers and a description, such as Lemma 5:Condition for Potato being Fried, even if the actual condition is too hard to describe. The current practice with such seems to be just to call it Condition for Potato being Fried, which is terribly arbitrary in that there might be various other conditions under which the potato is fried. --Dfeuer (talk) 19:46, 1 February 2013 (UTC)
I fail to see that it's a big problem in the first place - but as I've already been a proponent of using numbers in the recent work on PropLog I would be insane to say I objected to the numbering technique. As for providing them with consistent descriptive names I don't think that's important. Anyone who wishes to use a lemma that contributes towards the proof of such-and-such a proof will probably be reading that proof in the first place.
Certain people are rather more than normally concerned with naming, categorising, boxing, labelling and strangling the entities on this website (while being unconcerned about the presentational quality of such entities) and IMO such over-concern is not necessarily healthy for the site.
And if someone comes up with a better name, then good, we can rename. --prime mover (talk) 23:08, 1 February 2013 (UTC)
I have had some issues with results defying description to a point where I simply decided not to put them up. In many cases this applies to e.g. examples satisfying certain conditions subtly stronger than a certain definition or equivalent conditions for a def in some specific, hard to state context. I have generally decided not to cover those things in the past. Admittedly the first of the two current solutions isn't much of a problem. The second manifestly is.
Perhaps "Unnamed Result N" is better than "Lemma N" but that's up for discussion. It's not so much that I want to rename things, it's just that I want to have a means to let people post results they can't make up a title for. And with the lack of a title, the categorisation becomes quite important for already stated reasons. The "descriptive names" could instead be supplied by an explanatory line at the lemma collection page accompanying the link to said lemma. --Lord_Farin (talk) 23:36, 1 February 2013 (UTC)
In my suggestion, the descriptive name is optional but encouraged. Condition for Tent to be Leaky, Consequence of Tent being Leaky, Relationship between Two Notions of Leakyness, etc. --Dfeuer (talk) 23:40, 1 February 2013 (UTC)
But if you have a descriptive name for the page you don't need to number it. L_F: can you provide a solid concrete example so we know what we're talking about? If necessary post up one /some of those results and we can then discuss naming at that point. --prime mover (talk) 06:16, 2 February 2013 (UTC)

From Schilling, Exc. 6.11 (paraphrased):

Let $\Sigma$ be the Sigma-Algebra of Countable Sets on $\R$. Let $\mu$ be the co-countable measure. Let $\mu^*$ and $\Sigma^*$ be the outer measure and $\sigma$-algebra appearing in the proof of Carathéodory's Theorem. Describe $\mu^*$ and show that no $B \in \Sigma^*$ is such that both $B$ and $\R \setminus B$ are uncountable.

I have refrained from covering it. It exhibits non-measurable sets on this $\sigma$-algebra. It clearly defies concise description as $\Sigma^*$ is an involved construct relying on $\mu^*$ as well. --Lord_Farin (talk) 10:53, 2 February 2013 (UTC)

Okay ... thought needed. "Nature of Outer Measure of Co-Countable Measure on Sigma-Algebra of Countable Sets on Reals" and "Sigma-Algebra by Carathéodory's Theorem has no Uncountable Subset whose Complement in Reals is Uncountable"? --prime mover (talk) 11:15, 2 February 2013 (UTC)
Actually, $\Sigma^*$ arises in the proof of C's theorem in Schilling. There are undoubtedly other proofs avoiding it (since this $\sigma$-algebra is not canonical - it is larger than necessary). Still, even now my remark "defies concise description" stands. Also, this was merely an example; I've had a lot of these (also ones that weren't exercises but genuine results). --Lord_Farin (talk) 12:02, 2 February 2013 (UTC)
Maybe conciseness is the criterion which needs to be sacrificed? --prime mover (talk) 12:29, 2 February 2013 (UTC)
Maybe. I just felt that my suggestion could be a solution to a problem I've been dealing with a lot, but apparently this feeling isn't widely recognised. OTOH long titles don't promote memorability and ultimately are prone to the same duplication a numbering system would be - or so I think. --Lord_Farin (talk) 20:05, 2 February 2013 (UTC)
In the past I've preferred shortness over explanatoryness (some of the early proofs are laconic to the point of being gnomic) but I'm coming around to the idea that it may be best to explain all in the title. The argument against is that the contents of the various category pages look scruffy, but it's occurred to me that maybe we can rewrite the software so as to rigorously display category contents in one colum only thereby eliminating that concern. --prime mover (talk) 20:22, 2 February 2013 (UTC)
I oppose long titles for sake of reference. Nobody's going to find PW if many of the titles are long and pedantic. That'd make it no more than a hobby of some contributors allegedly in desperate need of a social life. --Lord_Farin (talk) 20:33, 2 February 2013 (UTC)
I see your point, but some of these results are important and/or relevalt only in the context of other, existing, more easily-referenced and (dare I say it) important results. While you may not directly need some of these obscurely-long-named proofs, one of the proofs someone out there is following the construction of most definitely will. The bigger picture relies on these futile and seemingly inconsequential (and unwieldy) details. The actual unwieldily-named result itself will be linked to from the bigger result that it forms a small part of. Its individual inaccessibility is therefore immaterial.
But that's just my take on it. That's how I envisaged this site working. Simplified for clarity: The important payload is the "named theorems" category - everything else is (ultimately) just a stepping-stone. --prime mover (talk) 21:01, 2 February 2013 (UTC)

That's practically my view. It's just that I don't want to waste time making up long names for futilely small parts and therefore I proposed the numbering system. It saves time. --Lord_Farin (talk) 21:06, 2 February 2013 (UTC)

In which case, as I say, if the numbering technique works for you then feel free to use it. I've been doing so in the context of PropLog as you will see. IMO the titles of things aren't as important as their content, and if someone comes up with a better title, then no problem, we change it. Post up what fills in the gaps towards what you feel important. I have omitted many examples and digressions in the books I've covered - only to go back again and fill them in when they become relevant to something else. --prime mover (talk) 21:42, 2 February 2013 (UTC)
BTW some of our contributors may have physical, social or political constraints so as to make it impossible for them to have a social life (beyond e.g. facebook). A person's got to have a hobby. --prime mover (talk) 21:03, 2 February 2013 (UTC)

True. But it's sometimes more rewarding to know that this hobby may actually end up helping other people out. (Also: Eeewww, FailBook - but that comment may be completely disregarded.) --Lord_Farin (talk) 21:06, 2 February 2013 (UTC)

Consider the following three:
  • The pages on Proofwiki are supposed to be short.
  • All significant results should in principle be capable of being integrated into Proofwiki.
  • Some very significant results require up to a hundred pages to be proved and will not be able to be broken down into a string of neatly named lemmas.
The issue is an inevitable one. I'm very surprised it's reared its head so early. The solutions you have all suggested sound great at this stage to me. In the future when the site reaches $10^5$ proofs (and it will) we will meet again to discuss "dependency spaces". Though we will all have long white and grey beards by then... --Jshflynn (talk) 18:47, 5 February 2013 (UTC)

slow

Someone doing a database dump? It's taking up to a minute to respond. --prime mover (talk) 07:01, 7 February 2013 (UTC)

I noticed. Not me. --Jshflynn (talk) 09:20, 7 February 2013 (UTC)
It's fine the way it is. No need to change. --Dfeuer (talk) 09:32, 7 February 2013 (UTC)
When you reply to a conversation, with whatever childish sassy backchat it may be, it is considered good manners to indent your comment with one more level of indent than the one that goes before. It makes following the conversation easier. You, as a Wikipedia veteran (unless you're not the same Dfeuer) ought to know that.
Indentation is based on threading level, not just chronology. --Dfeuer (talk) 10:04, 7 February 2013 (UTC)
Suboptimal. Better to use the technique based on chronology because its easier for us less intelligent mortals to work out who said what and in what order. Overcomplication is rarely a good idea, despite the fact that we are all supposed to be of above subhuman intelligence on here. --prime mover (talk) 10:30, 7 February 2013 (UTC)
Incidentally it seems to be back to normal now, so whatever glitch it was now seems to have been fixed. --prime mover (talk) 09:45, 7 February 2013 (UTC)
Except response is now sporadic. --prime mover (talk) 09:47, 7 February 2013 (UTC)
Had no problems earlier today. Currently facing the same response problems as PM. --Lord_Farin (talk) 13:23, 7 February 2013 (UTC)
Had an edit conflict with the above while posting: "Still horrendously slow. Been like this most of the day. Anyone else see this behaviour?" --prime mover (talk) 13:25, 7 February 2013 (UTC)
I can't find anything wrong on the server, but I'll keep looking. I've updated MediaWiki in hopes of a bug. --Joe (talk) 13:31, 7 February 2013 (UTC)
May be informative to know that I didn't receive any e-mails regarding my watchlist. This issue started this morning. --Lord_Farin (talk) 13:53, 7 February 2013 (UTC)

I think we've fixed this. It's looks like there was an issue with the mailer, so whenever an edit was made and an email needed to be sent the MediaWiki software hung. I've updated a ton of stuff and rebooted the server. Let me know if there are any more issues. --Joe (talk) 15:08, 7 February 2013 (UTC)

Much quicker now - back to normal. Thx. --prime mover (talk) 17:34, 7 February 2013 (UTC)
In relation to this, there was an incident yesterday where one of the contributors to this site managed to post up the same edit six times (or thereabouts) and expressed bewilderment at how that happened. I will make a suggestion as to why that may have been the case: being impatient at how slow response time was (in direct contradiction to the comment expressed by that user on this thread) it is believed that the "stop loading" button was pressed on the browser and "reload" pressed. It's tempting, I've done it myself when posting an edit which is taking too long (because this sometimes works to kick a slow server into action, don't know how). However, when done with a section of a talk page, it appears to post that section up as many times as you press "reload". I have demonstrated this on a posting of my own just now. So beware of the "reload" button. --prime mover (talk) 07:43, 8 February 2013 (UTC)

Template for ProofWiki

Every so often it is necessary to write things like "we use" or "on this website" or "specific to ProofWiki" and so on.

It occurs to me that, like "LaTeX" has its own presentational format: $\LaTeX$, we might want to do something similar for ProofWiki. Thus we would invoke the template {{ProofWiki}} and it would put in place an appropriately structured and styled (probably graphic-based) in-line "trade mark": e.g. $\mathcal{Pr}\infty\mathcal{fWiki}$ or $\mathsf{Pr}\infty\mathsf{fWiki}$ or $\mathbb{Pr}\infty\mathbb{fWiki}$ or pick your font.

Well, it's Sunday and I'm allowed to kick back at least once a week ... --prime mover (talk) 12:09, 10 February 2013 (UTC)

Sounds like a great idea! I like the calligraphic and sans-serif versions. --Lord_Farin (talk) 08:13, 11 February 2013 (UTC)
... or we could mix the fonts: $\mathcal{Pr} \infty \mathcal f \mathbb W \mathsf{iki}$ or something ... but I fear mathcal versions of lowercase are just standard italic font. --prime mover (talk) 12:53, 11 February 2013 (UTC)
The template has been implemented: $\mathsf{Pr} \infty \mathsf{fWiki}$ ... feel free to play with it, if anyone is of a mind. --prime mover (talk) 12:53, 11 February 2013 (UTC)

Users Online

Anyone else not able to bring up the "users online" page, http://www.proofwiki.org/wiki/Special:WhosOnline ? --GFauxPas (talk) 22:43, 11 February 2013 (UTC)

Been that way since Joe implemented a few changes to fix the load time problem of a few days back. I've tried to ask him but as of yet, no response. --Lord_Farin (talk) 22:46, 11 February 2013 (UTC)
If it compromises the efficiency of the site, I recommend we leave it non-functional. --prime mover (talk) 22:56, 11 February 2013 (UTC)
I have no strong opinion either way, but it's a bad idea to have the link extant on the side-bar thing if the functionality isn't there. --GFauxPas (talk) 13:30, 13 February 2013 (UTC)

Foldable template update

It turned out the foldable template Template:Begin-foldable had a flaw in it, causing the math inside hidden sections to be rendered awkwardly. This has now been fixed. One may need to clear the cache (Ctrl+F5 on most systems) to let the new version take effect. --Lord_Farin (talk) 14:38, 14 February 2013 (UTC)

Terminology question

Are there names for totally ordered sets with the following properties?

1 Every element has an immediate successor (predecessor)
2 Every element with a successor (predecessor) has an immediate successor (predecessor).
3 Every element has both an immediate successor and an immediate predecessor.
4 Every element with a successor has an immediate successor and every element with a predecessor has an immediate predecessor.

--Dfeuer (talk) 19:39, 14 February 2013 (UTC)

Are they needed? Are you posting up some immediate pages such that they need to be defined? If not, then walk away. --prime mover (talk) 21:04, 14 February 2013 (UTC)
I know of none such. --Lord_Farin (talk) 21:30, 14 February 2013 (UTC)

slow again

Response time is sluggish again. Anyone else notice this? --prime mover (talk) 22:02, 15 February 2013 (UTC)

All day. Ever since the "Users Online" feature was re-enabled this afternoon. We'll have to ditch it. --Lord_Farin (talk) 22:04, 15 February 2013 (UTC)
The extension providing the "Users Online" functionality was bugged and throwing errors all the time. It's ditched for now; in due time, we may craft a new version. --Lord_Farin (talk) 12:19, 16 February 2013 (UTC)
How much is it used? By me: never. I find out who's on line by reading the angry postings from people whose work I've just reverted. :-) --prime mover (talk) 12:33, 16 February 2013 (UTC)
I usually check it whenever I'm online and after I check the recent changes. Gives me an idea how the current activity. --Joe (talk) 16:47, 16 February 2013 (UTC)

Physics

The view has been expressed (again) that there is no place for physics on this site. This has been inspired, at least in part, by the fact that the use of mathematical abstractions of certain physical objects require the need for inconvenient disambiguations which have been described as being "obnoxious".

Outside of the unpleasantly fascistic overtones of the attitude expressed here, my own objection to this point of view is that without the underpinning of the basic laws of physics, the majority (no, let's make no bones about it, the entirety) of applied mathematics is thereby discarded as somehow irrelevant to the abstract musings and vapid bletherings of self-styled pure mathematicians.

What does anyone else think? --prime mover (talk) 09:32, 17 February 2013 (UTC)


I think one of the better possible definitions of mathematics is something like "any theorem in the language of set theory is a piece of mathematics".
In particular I think there is a case for restricting to theorems, definitions written in set theoretic terms. Thus "Angular Momentum" would be a valid definition since this reads (missing details obviously):
Let $(\mathbf x,\mathbf v)$ be a solution to Newtons equations. The angular momentum of the system is $\mathbf L = \mathbf x \times (m\mathbf v)$.
As would most physical definitions since physics is largely expressed in mathematical terms. So long as wikipedia-style essays are omitted this is well justified. Just my thoughts fwtw --Linus44 (talk) 10:55, 17 February 2013 (UTC)


So what's your philosophical attitude towards such entries as Definition:Chain (Physics) as a precursor to the derivation of the catenary? While the latter can just be defined as a description of the shape of the graph of the Definition:Hyperbolic Cosine, to discard it as "not mathematical" seems to me to be deliberately throwing away something fascinating for purely nugatory ideological reasons. --prime mover (talk) 11:25, 17 February 2013 (UTC)
Philosophically, I propose a Physics namespace (or perhaps several, as physics has several mutually incompatible formulations, some historical and others current), a Chemistry namespace, an Economics namespace, and probably a Biology namespace as well. All of these sciences inspire mathematics. None of these are mathematics. As an alternative, we could try to come to an understanding with a like-minded physics wiki to cooperate with cross-wiki links on matters of mutual interest. --Dfeuer (talk) 11:50, 17 February 2013 (UTC)
That list was not exhaustive. Electrical engineering, computer engineering, mechanical engineering, signal theory, sociology, etc., could all have their own namespaces if someone wants to use their languages for mathematical purposes. --Dfeuer (talk) 11:54, 17 February 2013 (UTC)


Regarding Definition:Chain (Physics), as it stands, I'm not sure this is strictly speaking a definition, unless the word "thread" is defined also.

It should be, it's just that this is work in progress. --prime mover (talk) 13:51, 17 February 2013 (UTC)

If this is just supposed to be intuitive motivation that's fair enough, but if the intention is that Definition:Chain (Physics) is a basis for the derivation of a catenary, then the definition needs specifying. For example, it might follow these lines:

Let $\mathbb E = \mathbb R^3$ be real Euclidean space. A chain is a mapping $C : \R \times [0,1] \to \mathbb E$ such that $C(t,x)$ ($x$ = position along the chain, $t$ = pt. in time) is differentiable w.r.t. $x$, continuous w.r.t. $t$, and with the properties:
1. The function $t \mapsto \operatorname{length}(C(t,\cdot))$ is constant (inelasticity)
2. etc...

..the precise wording depending on the formulation of the definitions on inelastic, stiffness. Then it becomes a precise mathematical statement; the grey area of interpolating between the intuitive notion and mathematical rigor is avoided (and surely that is work for physicists?). --Linus44 (talk) 13:37, 17 February 2013 (UTC)

Eventually, yes maybe, once we have developed the language appropriately, and add it as supplementary to the (so far) intuitive notion. But IMO this makes it unnecessarily impenetrable to the student of applied mathematics who requires a more intuitive notion. --prime mover (talk) 13:51, 17 February 2013 (UTC)
I agree that motivation by physical intuition is important in giving meaning to the theorem or definition; I don't necessarily suggest that it be omitted. Instead I suggest that whether a notion or property is proofwiki-appropriate could be determined by it's set-theoretic credentials. Thus something is included if and only if following a physical description there is a hard definition to work from.
I also agree that this aspect of the site might be underdeveloped to roll in a purely mathematical description of physics. Probably it is more important to have something written down, and the definitions can be rewritten if and when more precise reformulations/extensions are needed --Linus44 (talk) 14:15, 17 February 2013 (UTC)
I disagree that mathematics must be couched in the language of set theory, as I consider 0th and 1st order logic mathematics. If physical laws are treated as axioms or premises, then one can manipulate data within the frameworks of mathematics. More tricky are arguments such as:
objects are observed to approach the center of the earth in proportion to units of time squared $\vdash$
the rate of acceleration due to gravity at any point on the surface of the earth is a constant
which I do not know how to deal with. But something like First-Order Reaction I feel surely belongs on $\mathsf{Pr} \infty \mathsf{fWiki}$. --GFauxPas (talk) 13:44, 17 February 2013 (UTC)
... that said, Linus44's approach of something like "let $s: (\text{time} \ t) \mapsto \frac 1 2 a t^2 + vt + s_0$ seems like a decent approach, from which the argument about gravity I presented above would just be the derivation of $\dfrac {\mathrm d^2s} {\mathrm dt^2} = \text{constant}$. --GFauxPas (talk) 13:54, 17 February 2013 (UTC)
GFauxPas, that equation is only a very rough approximation of how gravity behaves on fairly short time scales. It doesn't even take Newton's inverse square law into account. Does anyone else think a second/third/seventh namespace for mathematical sciences would be good? --Dfeuer (talk) 17:57, 17 February 2013 (UTC)
It's close enough. Give him a break. --prime mover (talk) 20:35, 17 February 2013 (UTC)

Source citations and documentation flow

We have recently facilitated the documentation of the contents of multiple editions of the various publications used as source works. The question arises: in how much detail do we provide the process flow?

If we have access to more than one edition of a given source work, then we can apply the "previous ... next" technique to each of those editions. The disadvantage is that this makes more work. The advantage is that it is then possible to follow (through the chronology of the citations) the evolution of a particular idea: if we only register the flow from a late edition, it does not reflect the fact that a concept was used considerably earlier, which would be the case if the earlier edition were used.

I am coming to wonder whether, whenever we provide a source citation work flow, then unless we need to differentiate between multiple editions, we always use the date of the first publication of that work. There are instances of publication dates used in the sources section which are no more than the latest date that the contributor can see on the publication history page; no research has been undertaken even to note when the work originally came out. Does it really make sense to say, e.g. "Euclid: The Elements (2001)" (to use an extreme example for obvious rhetorical effect), just because the particular contributor had in his hand a reprint dating from that date? --prime mover (talk) 11:39, 17 February 2013 (UTC)

If texts have been reorganized at all (e.g., sections renumbered or even theorems swapped around), and if we only have access to a recent edition, how are we to number things according to the way they were in a previous version? And if we don't, someone actually trying to follow along using the edition they see us claiming to use will be mighty confused. --Dfeuer (talk) 11:58, 17 February 2013 (UTC)
You just don't bother to state which edition you are using. It's usually the case that a newer edition contains all the same stuff in the earlier one, with stuff added. A simple bit of research can easily determine what the differences are, if any. If there is a need to differentiate, then that's when edition-specific stuff is invoked.
We might want to merge the flow when different editions have the same material in them, and fork it as and when they differ. --prime mover (talk) 12:57, 17 February 2013 (UTC)
I'm not sure if I'm a fan of that approach. Although it may come across as excessively source-wanking when we cover multiple editions of the same book in full detail, I think it has some value in it to do that. Of course, it is not very likely that many instances of this practice will occur (for there are incredibly many other books that could be covered instead). This inclusionist practice will however lead to more accurate references. After all, when we cover a later edition of a work, we can't really be sure that it actually occurred in the first edition of that work as well. IMHO this is a minor point that needn't really be attended to. The current practice seems sufficient (which is what PM refers to: date of first publication, or an edition reference - then again, first publication date of that edition). This is IMO the best we can do at this point.
It is tempting to give way to utopian dreams requiring a lot of demanding technical work. These are however not presently relevant and the system is quite good atm. It can undoubtedly be improved, but such improvements probably require a radically different approach from how the sources section is currently used. I repeat, this is not worth the thought nor the effort at the moment. Those discrepancies in the displayed years need to be resolved (on an ad hoc basis) to the guideline set out above. --Lord_Farin (talk) 14:04, 17 February 2013 (UTC)

Mathjax stuff

The matter of upper and lower closures (along with some other things) got me thinking, so I went over to the MathJax documentation and had a look around. It looks like they offer some things we're not using that could make our work easier and better. The things I'm thinking about in particular are:

  • Site-wide custom macros
  • HTML links inside equations
  • Tool tips inside equations

So if I'm reading this right, we could have site-wide \supset, \sdownset, \upclose, and \downclose macros that we could set to whatever we want, and make them clickable links, so the reader can hover their mouse over it, see "upper closure", realize they don't know what that means, and click through to the definition.

Similarly, in the logic areas, an \xor may show the same symbol, but provide a different tooltip and click through to a different definition page. --Dfeuer (talk) 01:42, 21 February 2013 (UTC)

If we can get it to work without (a) breaking everything and (b) requiring that every page needs to be amended in order to handle it, then that could be a useful way forward to impose a site-wide notational standard. What's held us back in the past is that (a) had not been addressed, and the various techniques tried out did not work on e.g. transclusions. --prime mover (talk) 06:33, 21 February 2013 (UTC)
Since I'm planning to go through a hundred pages replacing $\uparrow$, $\downarrow$, $\bar\uparrow$, and $\bar\downarrow$, I think it would be pretty sweet if that project didn't have to be repeated if the choice of symbols is revisited again in the future. --Dfeuer (talk) 06:55, 21 February 2013 (UTC)
Before you do that, read the latest in the Strict Upper Closure talk - IMO this fiddly fussy notation needs to be burned with fire. --prime mover (talk) 07:09, 21 February 2013 (UTC)
I read it and disagree, but the point here stands: if we make site-wide macros we can try things out and change our minds without having to make ourselves crazy. --Dfeuer (talk) 07:15, 21 February 2013 (UTC)
We also have to decide on the appropriate $\LaTeX$ invocation code. \supset is of course unavailable: $\supset$ means something subtly different ... --prime mover (talk) 07:21, 21 February 2013 (UTC)
Whoops! Mind fart. strupset, perhaps? I think I'd want separate macros for the (strict) up/down set of an element than that of a set, so as to have appropriate links (if that turns out to work) and to allow for the future possibility of different notations. --Dfeuer (talk) 07:26, 21 February 2013 (UTC)
The problem with tooltips is that they don't work on SVG (possibly MathML too) rendering, but only with HTML/CSS. It could however be valuable to introduce macros. Other case in point is the usage of symbology in links. This won't work with your intended click-to-see-definition paradigm. The macros are to be looked into, because they can be valuable (particularly when a symbol is used outside its standard context so that the TeX name is misleading). I vote against the links on the grounds mentioned above. --Lord_Farin (talk) 09:00, 21 February 2013 (UTC)
Furthermore, when you right-click a TeX expression, you will not get your browser's context menu, but MathJax's. This is an even bigger problem than the links not showing up as such in SVG mode. --Lord_Farin (talk) 09:10, 21 February 2013 (UTC)


Milestone approaching again

It only seems an eyeblink since we hit 7000 proofs, but we're a mere 50 away from 8000. Most of that activity is probably due to the extensive refactoring work that has happened in the last few months, so a lot of what has been added is not actually new work.

As always, it is a conceit of ours that the round-numbered proofs are in some way "significant", so take care when we get up into the 7990s not to make proof 8000 the result of some mundane refactoring. It does not matter at this stage whether the proof you post up is actually complete (leave it as a stub for all it matters) just so long as the proof statement is there as a placeholder. --prime mover (talk) 08:04, 21 February 2013 (UTC)

Posted the 8000th one: Faltings' Theorem (aka Mordell's Conjecture). — Lord_Farin (talk) 13:37, 25 February 2013 (UTC)
Congratulations! May I suggest it be tweeted? --Jshflynn (talk) 16:35, 25 February 2013 (UTC)
I would suggest we wait for a proof. --Dfeuer (talk) 17:05, 25 February 2013 (UTC)
Okay. It makes me wish from the beginning that two metrics one counting theorems and the other counting proofs was implemented. Because:
* Every proof has an associated theorem statement.
* A theorem statement potentially has multiple proofs.
I am thinking the database could be modeled by a one-to-many right-total relation on $\mathsf{Theorem Statement} \times \mathsf{Proof}$, no? :) --Jshflynn (talk) 17:34, 25 February 2013 (UTC)

Category requests

I'd really like to add Category:Generalized Ordered Spaces as a subcategory of Category:Examples of Topologies. I'd also like a category for results on order-convex sets. --Dfeuer (talk) 04:43, 23 February 2013 (UTC)

Go for it. --prime mover (talk) 07:57, 23 February 2013 (UTC)
Except I'm in 2 minds as to whether it should be Category:Generalized Ordered Space consistent with the others. OTOH there are more than one, Or whether it should be Category:Generalized Order Topology in keeping with the other pages, and the wording of the space itself amended to match. --prime mover (talk) 08:00, 23 February 2013 (UTC)

I have not seen "generalized order topology" anywhere, personally. It's not completely determined by the ordering the way the order topology is. --Dfeuer (talk) 14:02, 23 February 2013 (UTC)

To put it as a subcategory of Category:Topological Spaces and Category:Examples of Topologies would be fine with me. I prefer the plural form. — Lord_Farin (talk) 14:13, 23 February 2013 (UTC)
I don't see the value in separate categories for those. There won't be rhyme or reason to what goes where. --Dfeuer (talk) 14:27, 23 February 2013 (UTC)
I wasn't suggesting two categories. Only that the category Category:Generalized Ordered Spaces be a subcategory of two other categories (of which one doesn't exist; I blame my memory). Thus, just having it as a subcategory of Category:Examples of Topologies would be fine with me. — Lord_Farin (talk) 14:44, 23 February 2013 (UTC)
Ah, I read that as you suggesting that the non-existent category should be created, and I opposed that idea. --Dfeuer (talk) 15:28, 23 February 2013 (UTC)

Exercises and examples

If I start adding previous and next links, should those travel through the exercises and examples? Many texts offer certain results only in those forms. --Dfeuer (talk) 00:51, 2 March 2013 (UTC)

That sounds great :) Have a look at the following pages to see what I believe is the House Style:
Exercises: Left Operation is Associative
Examples: Complex Plane is Metric Space
I predict a good source contains about 1000 "chunks". If you process them in batches of 10 it will just be 100 sessions of pain. Enjoy :D --Jshflynn (talk) 01:27, 2 March 2013 (UTC)
There is no requirement for everything in any particular book to be documented. If a bunch of exercises (and indeed, set of examples) consists of tedious applications of the principles in the text then they won't be included. If you have a mind, as suggested by Jshflynn, click through the "Next" (or "Prev") links of e.g. Warner (pretty well plundered up to approx. chapter 30) and see the big gaps in the Exercises documented.
Take for an example exercise 3.1: "Let $A$ be the set of words occurring in the first sentence of Chapter 1, $B$ the set of words occurring in the first sentence of the second paragraph of Chapter 1. What is $A \cup B$? $A \cap B$? $A - B?$ $B - A$?"
In such circumstances the prev / next links just go from Relative Complement of Relative Complement to Union Distributes over Intersection. If anybody wants to go back and fill in the gaps later, then all that needs to be done is to amend the links appropriately.
However, it is recommended that (unless you are seriously out of your depth) you make the best effort to document the full contents of the text itself. And if you get bogged down, simply abandon it and move on to your next book. --prime mover (talk) 09:11, 2 March 2013 (UTC)
I have always found it profitable to have a page residing in user space that keeps track of where I was, what was skipped in terms of proofs, paragraphs etc. and what still needs to be done. I'd also like to give the warning that for certain texts, there is a considerable amount of detail left to the reader (some of it even tacitly); the site philosophy is that these stepping stones also be covered — moreover, that their covering precedes the results (possibly tacitly) depending on it. Thoughtful consideration beforehand on how to present these "gaps" the book leaves (usually mostly for reasons of space) is necessary to ensure maximal coherence. (Note: This last step can be very hard and it got me to the "bogged down" status at least three times, so don't feel to bad if it happens to you :).)
In case you are working in an area that has already been (partially) covered, you might find the methods I provided somewhere on this page (or in the archives) to search through $\mathsf{Pr} \infty \mathsf{fWiki}$ with the common search engines very useful. I find them more accurate than internal search; one has to reckon with the fact that search engines' indexing will naturally be some days behind. — Lord_Farin (talk) 09:29, 2 March 2013 (UTC)
I was going to ask if it was okay to skip bits, but I have my answer now. I see no need to cover Birkhoff's I.1: Example 3 (of "partly ordered sets"):
Let $F$ consist of all real single-valued functions $f(x)$ defined on $-1 \leqq x \leqq 1$; and let $f \geqq g$ mean that $f(x) \geqq g(x)$ for every $x$ with $-1 \leqq x \leqq 1$.
--Dfeuer (talk) 14:51, 2 March 2013 (UTC)

Renaming after refactoring

I have been tidying up some of the double redirects (has to be done occasionally) and consequently issue this directive: when you have renamed a page, for whatever reason, it is good manners do a "What links here" on the page you have renamed and make some attempt to change the links in pages which reference that page. In the case where there are hundreds of them, then this task can be deferred or done piecemeal - except in the case where a link is itself a redirect, it is important that the redirect is amended to point to the new page, and especially important when there are already several pages which reference this double-redirect.

I note there are three categories of double-redirects at the moment which need to be addressed:

  1. A whole bunch of prev/next links in the area of Huntingdon algebras. This is my responsibility to sort out, and when I get round to it, it will be done.
  2. A partly-completed refactoring job on transitive closures, which will follow in due course.
  3. A considerable quantity of work that is needed to be done in the definitional structure of "Neighborhood", which is under discussion currently.

I have cleared up all the others, but please note that any contributor who starts a job which involved renaming pages is expected to complete it. If this does not happen, then bozo bits are in danger of being flipped. --prime mover (talk) 09:06, 7 March 2013 (UTC)

For part 2, it may be of interest to read my comment on Talk:Equivalence of Definitions of Transitive Closure (Relation Theory)/Intersection is Smallest. — Lord_Farin (talk) 09:11, 7 March 2013 (UTC)
Yes I see that, but I've lost track of the history of it and I fear it may have lost its direction. I still believe there is a place for "transitive closure always exists" and perhaps rename it "existence of transitive closure" etc., even if all it does is reference by link existing pages (e.g. the page in question). --prime mover (talk) 09:14, 7 March 2013 (UTC)

image uploads

I've been trying to upload files but I get this:

  • Could not open lock file for "mwstore://local-backend/local-public/d/db".
  • Could not open lock file for "mwstore://local-backend/local-public/d/db/Books.jpg".

I expect it's because I'm banned from uploading any more pictures.

The same thing happens when I try to restore pictures I've deleted.

If I email a file to someone, would they upload it for me? If the answer is "no, because your pictures are not supposed to be uploaded" I will understand. --prime mover (talk) 18:58, 7 March 2013 (UTC)

Sure, but that kind of sounds more like a database problem than anything else. --Dfeuer (talk) 19:03, 7 March 2013 (UTC)
Yep, I'm looking into this. --Joe (talk) 19:09, 7 March 2013 (UTC)

Can someone try again. --Joe (talk) 19:10, 7 March 2013 (UTC)

[79da9e5f] 2013-03-07 19:11:59: Fatal exception of type UploadStashFileException --prime mover (talk) 19:12, 7 March 2013 (UTC)

Another attempt if you would be so kind. --Joe (talk) 19:15, 7 March 2013 (UTC)

Upload warning
Could not create directory "mwstore://local-backend/local-public/archive/d/db". --prime mover (talk) 19:18, 7 March 2013 (UTC)


I also tried to restore an old existing file, and I got:


View and restore deleted pages
[[File:PrimeMoversOffice.jpg]] has been restored
Consult the deletion log for a record of recent deletions and restorations.
Error undeleting file: Error deleting unused archive file "mwstore://local-backend/local-deleted/s/t/u/stui6fumakk0pnirz9oe36ovl4cbq6o.jpg".
When I then tried to click on that [[File:PrimeMoversOffice.jpg]] link, I get:


404 Not Found
nginx/1.0.15
Does that help? --prime mover (talk) 19:20, 7 March 2013 (UTC)


but bloody hell - it seems to have worked! I will clean this up in a bit once you've done your debugging.

I think I'm good now. Thanks! --Joe (talk) 19:28, 7 March 2013 (UTC)

Nope! Still cannot get past:

Upload warning

Could not create directory "mwstore://local-backend/local-public/archive/d/db".

... when trying to upload "books.jpg" which is, admittedly, the same as "File:PrimeMoversOffice.jpg". (Might that have something do to with it?) --prime mover (talk) 19:36, 7 March 2013 (UTC)

SourceReview template

Note the new Template:SourceReview and how it has been invoked on the Definition:Subset Product page.

If you see an instance of the WIP template with the message "Sort out sources" or some such at the bottom of the page, and you are not in a position to do this (e.g. through not having the books) then please change it to the SourceReview template to give the user of the books in question a heads-up as to work that needs to be done in this specific area. --prime mover (talk) 19:49, 7 March 2013 (UTC)

image uploads again

I still can't get image uploads to work. I'm getting:

Upload warning
Could not create directory "mwstore://local-backend/local-public/archive/d/db".

it's obvious I'm either doing it wrong, or I'm still blocked from posting up images (the latter probably), because nobody else has reported any trouble with this tool. Can someone who isn't blocked give me their email address so they can do it for me? But clear it with the admins first or you'll be blocked too. --prime mover (talk) 20:52, 7 March 2013 (UTC)

It's a problem with the server.--Joe (talk) 21:16, 7 March 2013 (UTC)

Can you try again. --Joe (talk) 21:28, 7 March 2013 (UTC)

Gotcha. That worked. --prime mover (talk) 22:13, 7 March 2013 (UTC)
Excellent. In case you're wondering I've changed the web server from apache to nginx. There should be a noticeable difference in speed. --Joe (talk) 22:17, 7 March 2013 (UTC)

lost it

i cant make sense of anything anybody writes on this website any more --prime mover (talk) 21:46, 12 March 2013 (UTC)

Tell me more PM. Its important to say. --Jshflynn (talk) 22:40, 12 March 2013 (UTC)

Period of Minimal Contribution

Dear colleagues, it is with regret that I have to announce a considerable period of minimal contribution. To be honest, it has started already, and some of you may have noticed some symptoms. It is to do with the fact that I will have to, no, must, direct my energy to finishing my MSc. thesis. Anticipated ending of this period should be somewhere in June. I hope to contribute a bit in the weekends, mostly in discussions.

I may drop into the occasional interesting (major) discussion; if my input is valued, a note on my talk (User talk:Lord_Farin) will likely do the trick. Please follow the instructions there.

While this curfew (effectively that's what it is) is in effect, please feel free to finish any page with a User:Lord_Farin/Later or User:Lord_Farin/ShortLater marker. I would appreciate an invitation to comment on any site-wide paradigm shift. I will regularly check up on the macro discussion Talk:Main Page/Macros and Left-Right Pairs.

Should there remain any further questions (please keep in mind where this announcement originates before asking) I refer to my talk page. I hope to be back soon, and to find $\mathsf{Pr} \infty \mathsf{fWiki}$ in good shape. — Lord_Farin (talk) 22:14, 12 March 2013 (UTC)

Good luck. One will do what one can. --prime mover (talk) 22:29, 12 March 2013 (UTC)
That sounds like a promise to type up the proofs here when you're finished. What subject are you writing on? --Linus44 (talk) 23:05, 12 March 2013 (UTC)

Convention shift

Currently we have an "Also see" section in most pages where the opportunity exists to show pages that have a close conceptual connection with the page in question. Trouble is, it's not immediately obvious which are proofs and which are definitions.

Suggest that "also see" pages which are definitions should be entered as [[Definition:Thing Defined]] rather than the at-present [[Definition:Thing Defined|Thing Defined]]. In that way you can see at a glance what's a definition and what's a proof.

This would apply only in the "Also see" section - elsewhere the usual convention of providing a display link would prevail. Comments? --prime mover (talk) 22:36, 13 March 2013 (UTC)

Seems acceptable. It should be stressed in the documentation that the entries of the "Also see" section are to be referred in this way — as pointers to internal pages, rather than to their content. — Lord_Farin (talk) 22:39, 13 March 2013 (UTC)
Aesthetics: No. Utility: Yes. Utility wins. --Jshflynn (talk) 23:18, 13 March 2013 (UTC)

Axiom of Foundation

Do we have a template for this? Apparently systems with "anti-foundation" axioms are somewhat in vogue these days. --Dfeuer (talk) 01:02, 30 March 2013 (UTC)

What do you mean "a template"? --prime mover (talk) 07:29, 30 March 2013 (UTC)

Transclusion of symbol pages

When transcluding a subpage of the Symbols namespace using as an example Symbols:General/Ellipsis, there is no link at the top of that page back to its parent. On similarly structured subpages of e.g. the Definitions namespace, for example Definition:Set/Implicit Set Definition, there is a link at the top to Definition:Set Is this a MediaWiki bug, or does something need to be specifically configured? --prime mover (talk) 08:25, 14 April 2013 (UTC)

Joe, can you add the Symbols namespace to the list of ones admitting subpages? — Lord_Farin (talk) 06:52, 15 April 2013 (UTC)
Done. --Joe (talk) 16:36, 15 April 2013 (UTC)
Thx - that's perfect. --prime mover (talk) 19:45, 15 April 2013 (UTC)

editing slow

Editing pages is painfully slow at the moment: having to wait several seconds before the page loads. Can the infrastructure be checked to see if something's come loose again?

Thx. --prime mover (talk) 11:46, 21 April 2013 (UTC)

I've changed the fastcgi buffer size to be larger, it looks like it was running out. Hopefully this will fix your issue. --Joe (talk) 12:11, 21 April 2013 (UTC)
Excellent - it's back up to speed now. Mthx. --prime mover (talk) 12:20, 21 April 2013 (UTC)

Captcha

I've just added a new captcha method for account creation. Hopefully this will deter bots. --Joe (talk) 16:41, 21 April 2013 (UTC)

Naming: restriction vs. subset

It seems mostly arbitrary at the moment. I just duplicated Subset of Well-Ordered Set is Well-Ordered at Restriction of Well-Ordering is Well-Ordering (though my version is slightly more general in a fairly trivial way). Shall we make a decision about how all of these should go? --Dfeuer (talk) 20:06, 21 April 2013 (UTC)

Your stuff will all be merged with the existing stuff, leaving the page as a redirect. --prime mover (talk) 20:34, 21 April 2013 (UTC)
If it were just that page I wouldn't be asking here. There are many pages that were here before I was that are inconsistent in this precise manner. --Dfeuer (talk) 20:36, 21 April 2013 (UTC)
Shrug. Same treatment for all, then, I suppose. Does it really matter?
Can you go and work on an area of maths we haven't covered, rather than chew over existing stuff which is all perfectly adequately covered? We understand that there was some work done last year in the area of ordinals which needs to be revisited, and (as you may have noticed) I am now getting round to revisiting this area. By adding random variant definitions found by accident on the web by googling you are not really helping. --prime mover (talk) 21:26, 21 April 2013 (UTC)
Kunen, Kelley, and S & F are in my hot little hands, so there will be no complaining about those. Most of the other definitions fill in little gaps and there's no cause for concern. I know that this material has been touched upon, but it is far from "covered". Plenty of the proofs are shady in one substantial way or another. --Dfeuer (talk) 21:51, 21 April 2013 (UTC)
As for the matter at hand, the question is what treatment they will all get. We have a whole page of things about restrictions of relations (transitive, antitransitive, etc., etc.), foundational relations, well-founded orderings, orderings, strict orderings, total orderings, strict total orderings, well-orderings, strict well-orderings, one-to-many relations, many-to-one relations, and maybe more I'm forgetting. The advantage of the "restriction" approach over the "subset" approach is that it works for relations between two sets/classes, rather than just endorelations. --Dfeuer (talk) 22:24, 21 April 2013 (UTC)


Mathjax rendering

Does anyone else have problems with this? Specifically brackets are not re-sized by \left( \right) and sums are vertically misaligned, and where it's shifted outside the usual "space" of the equation it gets cut off. I've mostly ignored it, but while reading parts of the PNT it's so bad I had to check out the source code to figure out what it was supposed to say. Presumably a problem in this browser (firefox in fedora), but I don't have another installed to try it in. Is there anything to be done? Chrome creeps me out with its browsing-history based adverts.

Unrelatedly, image uploads don't seem to work, I get the message 'Could not create directory "mwstore://local-backend/local-public/3/34"'. --Linus44 (talk) 11:45, 24 April 2013 (UTC)

Not sure about MathJax, but I think I know how to fix the image upload issue. Maybe you could try Konqueror, Epiphany, or Opera? --Joe (talk) 16:00, 24 April 2013 (UTC)
Opera works good. --Linus44 (talk) 17:15, 24 April 2013 (UTC)

spam salad

Well that was an amusing little bunch of spam posts. I know mathematicians are all knuckle-dragging subhuman mouth-breathers, but I didn't think we needed a lecture on how to hold our cutlery. $\smile$ --prime mover (talk) 07:57, 18 May 2013 (UTC)

Haha, it seems like the spam is coming in waves. --Joe (talk) 14:33, 18 May 2013 (UTC)
Can't we put in some mathematical CAPTCHAs? Some Euclidean constructions, some problems on small finite groups, some basic calculus problems, etc.? --Dfeuer (talk) 16:05, 18 May 2013 (UTC)
Seriously though, there used to be an arithmetical captcha which was removed a while ago, at the same time anonymous edits were disabled (too much vandalism). I'm not a fan as it would mean an increase in tedium. --prime mover (talk) 19:06, 18 May 2013 (UTC)
That's a small price if it's only used on registration. --Dfeuer (talk) 19:10, 18 May 2013 (UTC)
There's already a pretty sophisticated captcha on account creation. --prime mover (talk) 19:50, 18 May 2013 (UTC)
Also, I think it would be excluding people who may not know how to solve such things. Compared to some other wikis I've visited, ours holds up pretty well to automated spam bots. I think that the majority of what we're seeing are human assisted bots, or just humans. --Joe (talk) 12:06, 19 May 2013 (UTC)
Someone who can't do basic algebra, geometry, or calculus is unlikely to be able to make any useful contributions here. I'm not suggesting that each person should have to do all of these, just one or another. I'm also not suggesting anything terribly complicated, but something a stupid spammer human might not want to bother with. --Dfeuer (talk) 13:11, 19 May 2013 (UTC)
I could not disagree more. I agree with Joe - spam is not a bad enough problem on this site. Chillax, man. --prime mover (talk) 13:31, 19 May 2013 (UTC)

Problem uploading graphics

Back to this old thing again:

Could not create directory "mwstore://local-backend/local-public/6/67"

when I try to upload a graphic. --prime mover (talk) 19:48, 18 May 2013 (UTC)