User talk:RandomP/archive 1

The other half of the problem with Verifiability

I agree with you about the problem that you mention, but it is only half of the problem. The other half is that still many editors consider that sources are only ways to attribute views such as in "The Sun reports ... [ref to The Sun]." or "Joe says ... [ref to The Sun]". The role of the publisher in checking facts, accuracy, fairness, etc. is ignored, that is, the notion of "reputable publisher" is reduced to the weaker notion of "well known publisher", which is sufficient to verify the attribution. -Lumière 19:14, 14 February 2006 (UTC)

Your comments

Too gloomy, really. I know what you're talking about. But if the policies, taken as a whole, weren't also capable of sensible application, by sensible people, I don't think we'd have made such progress. In mathematics, anyway, fussing about sources is not very common. We have to have a uniform set of rules, to cover the most contentious matters as well as the least. That doesn't mean in practice they have to be applied in the same way. WP can't avoid reflecting real-world tensions, but anyone who stays away from obvious flashpoints can have a relatively peaceful time of it.

Charles Matthews 22:37, 14 February 2006 (UTC)

Hmm, I'm not very happy with rules being arbitrarily enforced based on what's sensible, in particular since quite a few administrators disagree with me about what is. Just to make my motivation a bit clearer, there are two issues I have with the current state of things: I'm violating WP policy, and quite a few articles seem to be more about the references than anything else. RandomP 12:04, 15 February 2006 (UTC)

Well, the logic is that we have a job to do, namely write the encyclopedia. There has to be a framework for that. I'm not too happy myself about over-zealous use of footnotes, for example. On the other hand, it's a big collaborative venture, and I can't expect to be completely in tune with a few thousand other people about everything. Charles Matthews 08:35, 16 February 2006 (UTC)

Axiom of choice

r okay, I can't prove it, but that hardly means the negation is true.

I don't know, but I think that $(\forall x\in X:f(x)\sim g(x))\to (\prod _{x\in X}f(x))\sim (\prod _{x\in X}g(x))$ implies some choice principle which isn't valid in ZF, possibly AC itself. The predicate A ~ B means that there exists a bijection between A and B, but not a unique one. To construct a product bijection, it would be necessary to pick one from each. You can probably prove it. Not "you" as in there exists some person, likely with two or three PhDs in infidelclassical set theory, but as in the person whose talk page I'm defacing who showed some ability and originality and was probably in the right ballpark but gave up after getting shot down all of one time and might not even be all that interested due to having a life or something. Well I think it's interesting anyway. -Dan 05:02, 3 May 2006 (UTC)

Er, thanks, I guess :-) Just to be clear, I wasn't giving up on the problem or declaring it beyond my ability to solve. It just didn't seem the appropiate place, and my time was limited.

I am fairly sure the above (let's call it (D) for short) necessitates the axiom of finite choice (cartesian products of nonempty finite sets are nonempty, AIUI), but that requires several theorems that I haven't had the time to find:

• (ZF) → If x is Dedekind-finite, it is in bijection with a finite von Neumann ordinal n
• (ZF+AFC) and (ZF-AFC) are consistent relative (ZF)

Then work in (ZF-AFC), let $\prod _{x\in X}f(x)$ be an empty cartesian product of nonempty finite sets, let g(x) be the (unique!) finite von Neumann ordinal in bijection with f(x), and observe that $\prod _{x\in X}g(x)$ is nonempty, containing the constant function 0. Therefore, (D) conflicts with (-AFC), and a consistent model for (ZF-AFC) makes (D) false.

However, I'm not sure that "finite" sets in the axiom of finite choice are indeed necessarily in bijection with finite ordinals, and I'm not sure that (AFC) is indeed independent of (ZF). No article on that, I see.

RandomP 14:46, 3 May 2006 (UTC)

Yes, I think everyone watching Talk:Axiom of Choice stopped caring other than you and me, which is why I'm stalking you now I moved the discussion to your talk page in the first place. And no, there is no article on AFC. If we use "usual finite" instead of "Dedekind finite", it might be that AFC is already provable in systems weaker than ZF, so I'm not sure that's quite the way to go either. Haven't thought it quite through though. Hmmm.... -Dan 15:53, 3 May 2006 (UTC)

I'm afraid my copies of my parents' books Equivalents of the Axiom of Choice (two editions), and Consequences of the Axiom of Choice are at my office. I'm sure that both AFC (in fact, A2C) and "If x is Dedekind-finite, it is in bijection with a finite von Neumann ordinal n" are both independent of ZFU, and probably of ZF. (D) is probably named in Consequences, and one can easily look up what is known about it. — Arthur Rubin | (talk) 08:43, 7 May 2006 (UTC)

Hmm. I'd be very interested if the second statement were indeed independent of ZF. Ultimately, though, I hope I'm right in saying that A2C indeed applies to sets that are in bijection with {0, 1}, and thus if it's independent from ZF, so is the statement Dan made above.

RandomP 11:32, 7 May 2006 (UTC)

Exactly what I meant by A2C. — Arthur Rubin | (talk) 15:37, 7 May 2006 (UTC)

Okay, since

$(\forall x\in X:f(x)\sim g(x))\to (\prod _{x\in X}f(x))\sim (\prod _{x\in X}g(x))$

implies A2C, is true under AC, and, you say, A2C is independent of ZF (could you just verify that?), the statement is independent of ZF; in particular, it's not always true under ZF.

That seems to me to be a fairly strong argument against saying the product of cardinalities is always well-defined :-)

RandomP 16:01, 7 May 2006 (UTC)

OK, found my copy of Consequences.

Believe it or not, the statement in question doesn't seem to appear. However, AFC is form 62 $C\left(\infty ,<\aleph _{0}\right)$ and A2C is form 88 $C\left(\infty ,2\right)$, which both fail in model ${\mathcal {M}}2$, among others.

The closest similar form to the one you give is form 20, which I can paraphrase as

$(\forall x\in X:f(x)\sim g(x))\to (\sum _{x\in X}f(x))\sim (\sum _{x\in X}g(x))$

where

$\sum _{x\in X}h(x)=\left\{(u,v)|u\in X\land v\in h(u)\right\}$

Form 20 is not known to be weaker than the axiom of choice.

Does that help? — Arthur Rubin | (talk) 19:14, 8 May 2006 (UTC)

Er, can you just check form 20 for typos? shouldn't it be $v\in h(u)$?

Thanks, in any case! That might answer the question, and I'll try to remember to check out the book when i make it to the library.

RandomP 21:15, 8 May 2006 (UTC)

Er, just to be clear on this: your right-hand-side

$\left\{(u,v)|u\in X\land v\in h(x)\right\}$

seems to have no quantifier for x.

RandomP 21:18, 8 May 2006 (UTC)

Thanks -- fixed. Please note that updates and partial form tables are available online (with permission from the publishers) under Paul Howard's web site (see the Axiom of choice article) and my late mother's web site at http://www.math.purdue.edu/~jer , and that the book has a floppy disk (!) with software to generate the form tables, which may have been removed in library copies. — Arthur Rubin | (talk) 21:41, 8 May 2006 (UTC)

Hmm, interesting. I'd suspect that's because it's equivalent to (AC), but I do not know a valid proof. It's certainly implied by (AC), and it implies (A2C), so it seems to be in the interesting range. RandomP 23:24, 8 May 2006 (UTC)

A little more study shows that (D) implies form 60 $C(\infty ,WO)$.

Proof: Let X be a set of non-empty well-orderable sets. Define f to be the identity function on X, and define

$g(x)=\min \left\lbrace \alpha \in On\mid x\sim \alpha \right\rbrace .$

(The minimum exists because it's a set of ordinals.) This clearly satifies the hypothesis of (D), and the RHS of the conclusion contains the 0 function, so the LHS is nonempty. QED

Now, 60 is known not to imply 1 (the axiom of choice), but there may be other consequences of (D).

— Arthur Rubin | (talk) 00:17, 9 May 2006 (UTC)

Thank you. Well done!

Still, I ask myself if we can't go all the way. For instance, in the notation of what sounds like a handy book, we can do $C(\infty ,\geq \aleph _{0})$. If there is an injection P from omega to a set X, then there is a bijection Q from X-plus-one-element to X: let Q(u) = P(0), where u is the extra element not in X; let Q(x) = P(1 + n) if there is an n such that P(n) = x; and let Q(x) = x otherwise (law of excluded middle used: either there is an n such that P(n) = x, or there is no such n; it would likewise be needed to make Q^-1).

Now this doesn't give us a unique Q since P was not unique to begin with. But the precondition of (D) doesn't require a specific bijection, only specific sets f(x) and g(x) and the fact that bijections exist. Let f(x) = x union {x} and g(x) = x, which exist, are unique, but by foundation are not equal, and as shown bijections do exist between them. Then feed it into (D), then take the resulting bijection and feed it the identity function as an element of $\prod _{x\in X}x\cup \{x\}$, and get out an element of $\prod _{x\in X}x$, which is your choice function.

This gives us $C(\infty ,\geq \aleph _{0})$ and $C(\infty ,<\aleph _{0})$. Ironically, in ZFC, all sets are finite or have a countable subset, but not so in ZF. Unless that happens to be a consequence of (D) as well. -Dan 21:56, 10 May 2006 (UTC)

Ooooh, nice. So now we know where to look for counterexamples to (D) $\leftrightarrow$ (AC): Dedekind-finite infinite sets.

Of which, I'm afraid, I know virtually nothing, but one can always learn :-)

(I've also looked through the book, and nothing beyond what we've talked about struck me as particularly applicable -- I might just have missed it, of course)

RandomP 22:21, 10 May 2006 (UTC)

This paper [1] seems relevant; in particular, it introduces the notation DP for (D) and DP(n) for the special case where all f(x) are of cardinality n. The implication DP(n+1) → DP(n) was apparently proved by Paul Howard.

RandomP 23:00, 10 May 2006 (UTC)

Well, Blass's DP is (D) with the index set restricted to be $\mathbf {N}$, so it's not quite the same....

Oh, my. I think (D) does imply AC, after all.

Lemma. $C(\infty ,\geq \aleph _{0})\to C(\infty ,\infty )$

Proof: Let X be a set of non-empty sets.

Define $Y=\left\{\omega \times x\mid x\in X\right\}$

By $C(\infty ,\geq \aleph _{0})$, let g be a choice function on Y.

Finally, for $x\in X$, define f(x) = second component of $g(\omega \times x)$.

Clearly, f is a choice function on X. QED

— Arthur Rubin | (talk) 23:54, 10 May 2006 (UTC)

Very neat. I'm extremely impressed :-)

As far as I can see, the two last proofs are correct and prove the equivalence of (D) and (AC). Hope you two (I'd offer to do the TeXing, but really I haven't contributed anything) put it on the arxiv or make it accessible otherwise?

I, at least, think (D) is a wonderful way to state (AC) (and so obviously true :-) ), and I hope it's not lost again.

Also note that both proofs work without modifying the index set, so $D(\infty ,\infty )$ (i.e. (D)) is equivalent to AC, and $D(\mathbb {N} ,\infty )$ ((D) where X is countable) is equivalent to countable choice.

Thank you, wonderful proofs.

RandomP 00:28, 11 May 2006 (UTC)

Thanks for the DP note, too. As far as I can see, in the paper I'd linked to, DP isn't defined at all without a cardinal number to go in it - I'd derived a (wrong, apparently) definition purely from the explanation that it abbreviates "defined product" -- thanks for catching that.

RandomP 00:33, 11 May 2006 (UTC)

OOPS — I should have done more research. Form AC7 of Equivalents of the Axiom of Choice, II is proved equivalent to the axiom of choice, and clearly follows from (D).

AC 7: The Cartesian product of a set of non-empty sets of the same cardinality is non-empty.

So much for a paper, even in arxiv ... — Arthur Rubin | (talk) 01:27, 16 May 2006 (UTC)

Uhm, I don't see how it "clearly follows". Maybe I'm just being dense ... RandomP 09:56, 16 May 2006 (UTC)

Oh, sorry. Got it now. Yes, that's a much shorter proof. RandomP 10:35, 16 May 2006 (UTC)

Still, well done everyone! And thank you again! --Dan 17:56, 17 May 2006 (UTC)

User:Sisalto

Thanks. I hate it when that happens. I've moved it. Cheers! -- Mwanner | Talk 22:27, 20 May 2006 (UTC)

Wikipedia's zenith of usefulness

I took a look at your User page, and I'm curious-- what would you say is the ratio of articles-that-are-deteriorating to articles-that-are-improving? I'll grant you, some go through some rocky periods, but I personally don't see a lot of articles that are going steadily down hill. How many can you list that seem to you to be in steady decline? -- Mwanner | Talk 00:01, 21 May 2006 (UTC)

Okay, I'm going to answer both of your question, but let me take off on a tiny little detour first:

It is entirely possible for 99% of wikipedia's articles to be improving and the remaining percent to be "deteriorating" (in quality, not in length! article bloat is in fact a problem) so badly that overall utility decreases. (of course, the opposite situation is possible, too). Considering that articles are constantly being created, it would even be possible for the average article (what you get when you hit "Random article") to deteriorate even though every single article is improving.

I've just performed a little experiment: I've selected ten articles from the bottom of the 500 recent changes, waited for a while (20 minutes), then compared the new (and stable! minor vandalism should have been reverted after the 20 minutes, and i would have seen it and excluded the page) version to the one 24 hours prior.

The result was that 3 articles had had no overall change; one article had had link improvements and a small copyedit; one article had changes that I think did not influence its quality; one article had minor edits that I thought made it slightly worse; three articles had major changes adding essentially no new information but moving significantly further from the NPOV; and one article, heartbreakingly, added some valid and arguably notable information, but had its opening paragraph ruined in the process, as well as losing its NPOV.

Overall, I would have been better off reading those ten articles yesterday. I realise I must have gotten a bad batch, but I'd guess (I'm basing this, mostly, on intuition, and that one experiment) that it is only two or three articles actively being edited that are actually improving, and three or four that are deteriorating, with no way to tell for the others.

To answer your second question: Yes, I could mention articles that have been in steady decline; however, I'd really prefer not to. I do not want to attack anyone who identifies with an article (and put hard work into it), and ultimately I'm a bit ashamed I'm no longer trying to fix every single article I come across.

Ultimately (sorry this got a bit long), I'd like to retract some of the "zenith of usefulness" statement. I still think it's very snappy and provocative and all that, but I should have made it clearer that I was referring, well, to the web site Wikipedia. The usefulness of finding out about something by typing it into the article box and hitting go might be declining, but the enormous database that I have begun to think of as Wikipedia since installing MediaWiki here will continue becoming more useful — it is trivially more useful than it was at any prior point, because you can just forget about all late edits.

Just a really short note to end this: I think these problems are, partly, of a technical nature. That's why I'm playing with the MediaWiki software, and trying to improve it.

RandomP 01:04, 21 May 2006 (UTC)

Thank you for the thoughtful answer. I know that it is unreasonable to expect you to have anything like a definitive defense of your provocative statement, and I'm certainly not ready to make a defense of a contrary position. I'm curious, though, what the results of your test would have been if you had looked at changes to the articles over a longer time scale-- a month, or even six months. I suspect that daily changes amount to little more than noise, greatly aggravating the problem of the small sample size. I know that many articles drift along for long periods with scores of insignificant edits, and then someone sits down and does some serious work (not always for the better, of course).

It is certainly true that the growing popularity of Wikipedia is somewhat unnerving-- one needn't be much of an elitist to worry that as the number of participants grows, the average quality of the contributions might be expected to fall. That, at any rate, would make a reasonable explanation of a peak-and-decline model. Still, in the articles I watch, the relatively small number that have suffered what I consider to be declines in quality have done so at the hands of editors who are recognisable experts-- otherwise their work would have been reverted pretty painlessly. It is true, though, that I only pay attention to a few small corners of the whole, and I am certainly not polymath enough to confidently judge all that I see. It would be nice to see a study similar to Nature's Wikipedia-EB comparison applied to the Wikipedia change-over-time question.

In any case, I'll be curious to see if your technical approach to the problem pans out. It would be wonderful if the quality could be sustained without adopting the elitist methods called for by so many over the years. -- Mwanner | Talk 14:06, 21 May 2006 (UTC)

To answer things in a last-in-first-out fashion, I don't believe there is anything elitist about signing off on an article or edit, essentially saying "this is okay, it's not vandalism". One-click reverts might help a lot, as would letting random readers help by marking recent changes and giving them a "click here if this looks suspicious to you" link, without all the scary warnings and IP logging.

I might do the other experiment you suggested, but I'll wait for a weekday ...

And I think the problems might be the articles people watch aren't actually the ones that should be watched in a perfect world ...

RandomP 18:38, 21 May 2006 (UTC)

No, there's nothing elitist about signing off on an article or edit-- the elitism (potentially) comes in when you decide who can do that. I like the one-click revert, and the click-here-if-suspicious. I would also like to see Wikipedia:General_complaints/Archive_4#Automatic_Edit_Summaries. -- Mwanner | Talk 13:21, 22 May 2006 (UTC)

Everyone, of course. Whether you trust another editor who signed off on something is, ultimately, your decision, and the server-side wikipedia software shouldn't make it for you. Not a terribly easy thing to implement, of course, since the server-side software pretty much does everything the way things are set up now, but maybe a guideline to keep in mind.

One of the things I'm thinking about isn't whether edits of articles shouldn't themselves be first-class citizens - with a talk page and the ability to carry additional metadata (think Category:Suspicious edits), and, ultimately, an existence outside of that shady "history" tab. The main problem with that is that there are way too many of them, so (in particular) talk pages should probably be shared whenever possible.

Just things I'm thinking about, really. Metadata is good (even though I don't think the category system is a terribly good way of doing it), keeping metadata in the article works great for "encyclopedic" metadata but sounds like the wrong thing to do for administrative metadata, which is currently stored in the history tables, entirely differently from the article data, category metadata, and article discussion.

Note that this would also make clear the difference between reverting an edit (creating another edit that reverts all changes) and deleting it (which wouldn't necessarily delete all traces of it, but would remove the material that was added from wikipedia's data bases, something eminently useful for copyvios or huge edits ....)

RandomP 17:32, 25 May 2006 (UTC)

Hello again, turns out I was still stalking youwatching your talk page, when I noticed this section. Having now had a good read of your user page:

I disagree with Wikipedia policy. For certain edits, I feel it's worth it to adhere to it anyway, 
and make them, and for certain others, I feel it's not, so I don't make them. I also have no idea, 
to be honest, how WP policy is made.

Heh. Suffice it to say WP:IAR still survives from ancient time. -Dan 16:54, 25 May 2006 (UTC)

Hello!

Thanks. I had forgotten about that, might have to use it in the future :-)

FWIW, I don't mind being watched at all. I certainly couldn't trust any encyclopedia that would let someone as untrustworthy as me edit unwatched.

RandomP 17:32, 25 May 2006 (UTC)

Informal mediation over Inflation

Carbonate has requested the involvement of the Mediation Cabal at Mediation Cabal/Cases/2006-05-27 Gold as a measure of inflation. From Talk:Inflation you seem to be willing to be involved in mediation, but Carbonate doesn't seem to have informed you about the request yet. If you would like to take part, I will be happy to act as mediator. --Sam Blanning^(talk) 12:35, 31 May 2006 (UTC)

I have gone ahead and accepted the case. It would be quite helpful if you added a summary of your perspective of the dispute under the 'discussion' section at the bottom - nothing too heavy, just to confirm your interest and help us get a clear idea of how both sides view this. --Sam Blanning^(talk) 13:17, 31 May 2006 (UTC)

Thanks for accepting the mediation request! I had indeed not known where to find the Mediation case. RandomP 16:56, 31 May 2006 (UTC)

Tenebrae

I didn't remove anything that I'm aware of. I added two comments to User:Facto's 3RR request since it contained inaccuracies. Please let me know what you think I removed and Ill fix whatever I might have done inadvertently. Thanks. -- Tenebrae 12:27, 2 June 2006 (UTC)

Uhm, the problem has been fixed now. From what I can tell, you must have accidentally reused my template instead of creating a new one, or something. No big deal, though others on your talk page seem to think it was ...

RandomP 12:31, 2 June 2006 (UTC)

Thanks. Yeah, I just saw that what I'd done at "01:38, 2 June 2006 Tenebrae (comment)" was replace the whole page's content, which I'd accidentally blanked out. Thanks for getting back to me about it. Would you mind leaving a message on User:Facto's talk page? We're in mediation and he's been making all sorts of unfounded, bad-faith accusations. I'm sure he's already taking this to the mediation page.

Thanks for any help clearing this up, and again my apologies for the inadvertent thing. Best regards, -- Tenebrae 12:43, 2 June 2006 (UTC)

Big thanks (not sarcastic this time)

I wanted to thank you for making the two topics you did on the Hubbert Peak Theory article (may be more by the time you read this. The issue is really messy, and proponents are hard to pin down on what they mean. The article (and its branches) has devolved into speculation and addition of unreliable sources, as it's mainly doomsayers editing it. So, the topics you made are appreciated. This is proof that your powers of pedantry can be used for good, not just evil ;-) MrVoluntarist 02:55, 6 June 2006 (UTC)

That's the impression I got. It seems a bit weird to have an article about a mathematical model that doesn't actually describe the mathematical model. Whether or not it's due to Hubbert is ultimately immaterial, but it seems to me the mathematical breakthrough he is credited with largely consists of "uh, and I can't draw a curve that falls off steeper than this without violating my intuition". I might be wrong, of course.

It's great to see someone similarly critical of the article as myself. Let's see what happens ...

RandomP 03:18, 6 June 2006 (UTC)

Spelling error

Hi

I use a programme to point out spelling errors on wikipedia. I was just using that. -- Funky Monkey  (talk)  21:34, 6 June 2006 (UTC)

Ah. As far as I concerned, it's a matter between Carbonate and you, but you might to be careful not to correct spelling errors in quotes, on talk pages, and the like :-)

Happy Editing!

RandomP 21:36, 6 June 2006 (UTC)

A soliloquy about the Hubbert peak theory

This is a monologue, but feel free to chip in. I don't want to place it on the article's talk page.

I had a bit of a shock yesterday.

I had read the Hubbert peak theory article before, found it a bit heavy on the speculation, a bit light on the theory (so he uses a logistic curve .. hmm, I'm sure he had his reasons to), but generally acceptable. Surely there was an established mathematical theory that explained everything. There had to be, the article said so, and besides, I'd read it on the internet many times before.

Then, yesterday, I got a bit bored. Something about the Hubbert peak theory had been nagging me (it's this: I'd read in several places that a logistic curve was used both to model the production in a single country, and to model the global production. But there are only finitely many countries, and there is no nontrivial positive linear combination of logistic curves that would, again, result in a logistic curve. Hmm. Of course, somewhere in the model might be something to explain it, and treating the whole world as one country doesn't seem that illogical. However, all things considered, that means that even in the model, the logistic curve would merely be an approximation), so I decided to read the article again, and read the references.

It seemed an easy endeavour. Hubbert, I read, had published in 1956 a most influential paper, Nuclear Energy and the Fossil Fuels. A bit of background research verified to me that "Nuclear Energy" was the buzzword of the time, so this seemed to be what I was looking for. I skimmed through the paper, and it seemed good. A couple of integral signs, nothing so much like a differential equation, but an odd knack for writing out in words what would today be formalised. When, on page 21, he said

"Utilizing the method of extrapolation described earlier ..."

I was pretty convinced that I merely missed that "method of extropolation", that "mathematical model" everyone and their cat seem to believe in these days.

So I read it again.

And again.

There is no mathematical model.

So, I did the next best thing, and checked the other references.

None of them, that I have found, include a mathematical model.

I did the next best thing, and tried to come up with one myself:

Let's see, by the analogy to population dynamics, the total amount of oil that has been mined so far is the population. It's something like the "industrialisation" state of society, I guess, or something like the "oilification of society". How many machines and cars and planes and trained surveyors there are. How much is known about oil, too.

Now, for every unit of "oilification" (measured in barrels), that society will attempt to get more oil, every so-many years (this so-many constant is the r of population dynamics). Furthermore, if we expand the analogy a bit, society will stay as olified as it has once been, because that r is already corrected to account for dead barrels: barrels whose direct effects have long since ceased influencing society.

But this is only an attempt, and its chance of success is proportional to the oil remaining. If the attempt fails, well, ... nothing happens. The society doesn't get de-oilified, there's no overflow where extra effort is put into finding oil that an oilified society needs, nothing.

This model is, to put it mildly, simplistic.

To put it less mildly, had I set out to come up with something from scratch, there would be more intuitive reason to believe it would work.

And as I said above, it's not even additive with regard to two populations: for most of the oil-producing era, humanity was divided into two blocks, largely without contact. Both depended on oil. Interactions between them seem insufficient to postulate a common degree of oilification. But the "theory" doesn't even have the .. er .. theory to deal with that.

I'm now a bit at a loss what to do. Discussion on Hubbert peak theory suggests, contrary to everything I've been able to find, that I'm just blind, that it's somewhere in the references.

My oilification theory does explain the Hubbert curve, but it seems counterintuitive, and frankly preposterous: the logistic curve might or might not apply to a single deposit, where more oil is only found after less oil has been, but on a global scale, there frankly isn't anything to make such a model make sense. The difference between the Hubbert curve and that of the standard normal distribution is small enough to make the two indistinguishable in data, and whatever the result of Hubbert's model is must surely be convoluted with a normal distribution to account for mere chance delays in finding and producing the oil?

So, the problems that are making me contemplate the Wikipedia equivalent of suicide by cop (by jumping in and demanding primary sources for any mathematical model ascribed to Hubbert, or changing the article to make plain that no evidence of such mathematical models survives to the present day):

• There is no mathematical model. Not in Hubbert's paper, not in the "Beyond peak oil" book (not that amazon's search inside feature would find, anyway), not in pages and pages of google results.
• There can be no mathematical model. It should behave additively with regard to a world split in two, as Hubbert's world was, but it doesn't. That assumption needs to be spelled out.
• The mathematical model absolutely belongs in the article. No compromise on that. It is not a "guesstimate". It is not an "empiric model". It's not a qualitative description, because it includes specific predictions for the peak to occur at the point where half of the oil available has been produced. It's mathematics. Mathematics belongs out here in the public domain (not in the copyright sense of the word, but in the displaying-it-to-everyone sense of the word). There is no excuse to hide it.

There are some (very few) mathematical models that are actually too complicated to include in Wikipedia. But there aren't that many. I'm convinced that whatever mathematics was used to model oil exploration is something we can include, at least.

The problem is that I sincerely doubt there is a mathematical model, that works any better than what I described above. The whole area seems to live off citations of that 1956 paper, and happily churning out prediction after prediction .. they might be right. But then, the normal distribution is all that might be needed, or the symmetry claim might turn out just to be completely off base. Or the whole thing might be a self-fulfilling prophecy, in that everyone "knows" today is the time to stop looking towards increasing production.

There might be symmetry. There might be a plateau at the top, of five or ten or fifteen years. Or exponential increase, again, until shortly before the end, when suddenly it's no longer worth it. There might be a sudden decline, as everyone realises that starting new oil exploration is unlikely to pay off before replacement technologies take off.

One of the most startling assumptions, not spelled out explicitly but implicitly used in Hubbert's paper, is this:

The total production curve is continuosly differentiable, and not too steep.

That's something physicists do a lot: continuity makes things easier. It just isn't always the right thing. It makes sense when you consider individual oil fields, their wells producing less and less oil as they decline exponentially .. until, *whoops*, the energy efficiency critical point is reached. Once whatever comes out of that old well has so much water that you need more oil to separate the oil from the water than you get out of it, effective production will crash. The total production curve is essentially something like max(.95, logistic curve), once you ignore oil that it would be a waste to produce ...

Well, if I can't put a long rant here, where can I put it?

RandomP 01:56, 7 June 2006 (UTC)

Sorry if my posts to you get kind of schizo, but I wanted to express my appreciation for what you're doing on this. It seems you've put a lot of work into this, and, more importantly, useful work. I had always regarded the Hubbert "theory" as "yeah, that's nice, I'm going to go back to the real world now where people can't suddenly declare themselves experts outside their discipline". Your work seems to confirm what I was too lazy to confirm myself -- that there is no basis for it. Whether oil is extracted depends on much more than the technical difficulty of doing so, among other reasons. I remember reading a propaganda piece before about Peak Oil whose *only* response to suggestion of market-driven reaction and recovery was, essentially, "Uh. Right. The market will take care of it. Tell that to your kid when he's starving." Whateva. MrVoluntarist 02:34, 7 June 2006 (UTC)

If you need help, let me know. (I think I qualify as enough of a mathemtician to be able to recognize if there's a mathematical theory present.) I wonder about:

The total production curve is continuosly differentiable, and not too steep.

as well. I think I could justify "piecewise continuously differentiable", but what does "not too steep" mean? (For what it's worth, I think I saw a version of this article chasing spam from someone named "Lavigne" on the usenet groups misc.taxes and can.taxes. The article seemed to be better then than it is now — it didn't support his rants then, and it still doesn't but it's closer. — Arthur Rubin | (talk) 16:51, 7 June 2006 (UTC)

---

Just in the interest of completeness, there is something like a reason for the "not too steep" bit, at least in one direction: essentially, some sources seem to suggest that oil wells do indeed see their production drop off exponentially when they're depleted, rather than just going from "oil" to "no oil". If there's a maximum rate at which that happens (the Ghawar field article mentions 8% p.a.), then that does mean that the production curve can't fall by more than that.

(Note that in a real oil field, the curve would look more like max(exp(-t) - C, 0), but for P the oil production, the inequality

$P'(t)\geq -8\%{\rm {p.a.}}P(t)$

(assuming there is some reason in the model for the 8% p.a. number)

would still hold.

However, that's just a guess. No mathematical model has been appearing so far, though ... I wouldn't be surprised if that were the reason for Hubbert assuming a "necessary long decline", though.

Purely as a note of interest, I've also stumbled across Image:Technocracy graph1.jpg, apparently from the Technocracy Study Guide, to which (or a later or earlier version of which) Hubbert contributed ... I spot a certain similarity, though, again, I don't want to go as far as attributing that to Hubbert :-)

RandomP 20:42, 7 June 2006 (UTC)

---

Incidentally, if that was indeed Hubbert's motivation (I'll reparametrise everything so production for a single oil well is $Cexp(-t)$ above a certain t and 0 below it), then we should be able to find a function D(t), the newly discovered oil wells at time t, such that

$P(t)=\int _{-\infty }^{t}D(\tau )e^{(\tau -t)}\mathrm {d} \tau$

Assuming the Hubbert curve production, by a short calculation (for which I still used Maxima):

$D(t)={\frac {2e^{-2t}}{(e^{-t}+1)^{3}}}$

which is recognizably (gnuplot: plot exp(-x)/(1+exp(-x))**2, 2*exp(-2*x)/(exp(-x) + 1)**3;) not the Hubbert curve. It peaks at -log(2), when production is at 8/9 of the maximum (that last statement is independent of the growth rate).

Is there any reason Hubbert predicted, if indeed he did so, that oil "discovery" (i.e. oil wells coming online) would not happen along a logistic curve, but have a sharp peak before declining significantly faster than it rose?

I'll try to work out next how many of the oil wells have been "discovered" by the time discovery peaks, but I know now that number is going to be more than .5. Why? Wouldn't it make more sense for that to be symmetrical?

RandomP 21:08, 7 June 2006 (UTC)

Progress

I have now found several secondary sources that state that Hubbert himself apparently didn't decide the logistic curve was the one to fit his needs until 1982, when he wrote

Techniques of Prediction as Applied to Production of Oil and Gas, US Department of Commerce, NBS Special Publication 631, May 1982

Secondary sources claim that that paper contains actual mathematics, beyond coming up with the logistic curve without any justification. However, for a typical example of what those "secondary sources" look like, look at this. The relevant bit is:

Hubbert's theory is simply the assumption that the relation between P/Q and Q follows a straight line

(where P is the derivative of Q, and the straight line goes through P/Q = r and Q = K, in population dynamics terms).

That is not a mathematical theory. It's a restatement of "oil depletion follows a logistic curve".

One of the authors then goes on to ignore all data prior to 1958; it doesn't match his straight-line approximation. Hubbert's original paper was written in 1956.

So, it's essentially down to that paper. Again, that's

M. King Hubbert, "Techniques of Predictions as Applied to the Production of Oil and Gas" in S.I. Gass, ed, Oil and Gas Supply Modeling, Special Publication 631 (Washington, D.C.: National Bureau of Standards 1982), pp 16—141.

(Some references claim a very similar 1980 publication. I have no idea whether they're different).

The closest library where I've found it is 400 miles away; it's on microfiche there, and I'm unable to make the trip and unlikely to be in that area for other reasons.

So, pretty much stuck here. I can read more google hits that all pretty much rehash the same thing, but I do fear that reading the article will show that there is no mathematical model beyond the use of a logistic curve based on empirical coincidence.

I feel tempted to revise the article based on the rule for non-mathematical articles that anything obscure enough not to have secondary literature isn't notable enough for WP, but that would be a borderline call at best — there is secondary literature, it does claim a mathematical model, it just doesn't explain it.

Or am I being overly harsh here? Is "assume that for the function f, which we shall interpret as cumulative oil production up to time t, we have ${\frac {f'}{f}}+f=1$. Then f is the logistic curve, and it's symmetric" a mathematical model?!

RandomP 13:53, 8 June 2006 (UTC)

I thought a mathematical model was saying that "assume this system holds these conditions, then this conclusion follows". That's what the a mathematical model article says. In that case, saying "assume oil extraction behaves like this. Then it will peak like this" is a mathematical model, just unscientific. (That's why it's so hard to know if this counts as a scientific theory, and if so, if it's obsolete.) Or am I way off on that one?

I want to help however I can. If you can think of anything on which I might have a comparative advantage, please let me know. MrVoluntarist 14:16, 8 June 2006 (UTC)

I just want to be sure that you understand that if there indeed is a mathematical model (one that makes reasonable assumptions from which the logistic curve follows), all I really want to change about the article is source the statements to the right papers, essentially, and include a description of the model.

As far as I know, and I'm just going to go by this as long as it's undisputed, the 1982 paper is the only reference that actually might contain such a model, and there are also a couple of sources claiming that it indeed does.

So, I'll try to get access to the paper, read it, and if it doesn't have a mathematical model, I'll change the article to reflect that.

Again: It's entirely possible that that 1982 paper contains an argument that convinces me, at least, that there is a mathematical model. If that's the case, well, I'd try to put it in the article.

RandomP 00:05, 9 June 2006 (UTC)

Thanks for clearing that up, it makes sense now. I didn't mean to imply you were on an anti-peak crusade, but any work you do toward pinning down precisely what the theory says will help me in find relevant critiques of (or supporting evidence for) it. MrVoluntarist 00:18, 9 June 2006 (UTC)

United States article

I agree with the points you made on the opening paragraph:

• America is the world's sole superpower, a nuclear power, the world's largest single-country economy, the large country with the highest per-capita GDP, the country most important to scientific and industrial research, seat to the United Nations, dominant NATO member, largest Western country by population (as well as economy), the country in control of the world's reserve currency, ... not all of this needs to be in the opening paragraph, but I find "oldest existing constitutional republic" a bit weak.
• America has a democratic form of government. While this is nowadays seen to be implied by "constitutional republic", I would think it important enough to point out specifically.

I'd be willing to work with you to improve the opening paragraph to include your observations.

The preceding comment was moved from my user space, where Northmeister put it with the following signature :--Northmeister 02:26, 8 June 2006 (UTC)

Please replace this paragraph with a new signature.

My mistake, your user page looks similar to a talk page. Sorry about that. --Northmeister 03:16, 8 June 2006 (UTC)

---

Not a problem at all. I've been wanting to change it to be one of those fancy box-rich layouts, but for now haven't found the time :-)

Sorry about being so pedantic, but I've noticed how quickly things get confused on Wikipedia.

Yes, I'm willing to work on an opening paragraph. However, I don't think I'm half as good at actually writing things as I am at finding things to criticise about them :-)

As a suggestion, who I'm really writing for is a (hypothetical) kid outside of North America, maybe a smart fourth- or fifth-grader, who did not grow up with reliable information about America. If there's one paragraph you could tell her about it, what would it be?

I keep being tempted to write "The United States of America is the world's most important country" because, well, it's true. But even with a "widely considered" in there, it's too much of a judgment call.

My personal selection would be:

• "among the largest countries" economically, by area, and by population
• seat to the UN
• reserve currency
• democracy

But that's my personal bias.

So, here's my shot, with all those caveats:

The United States of America is a country in North America. It is among the world's largest countries economically, by area, and by population, and, today, the world's sole superpower. It hosts the United Nations headquarters and provides the world with its reserve currency, the US dollar. Its founding in 1776 was the first of a series of revolutions and reforms establishing democratic forms of government that have today reached all but a handful of states.

I realise that last sentence isn't brilliant, but I think it's an important point.

I tried to make another version with the not-totally-PC points clarified, but it was so terrible I had to delete it.

From a global POV, for all but a few fringe disciplines, the United States is the most important country in the world. No two ways about that. The opening section should make that point clear, but avoid the term "fringe", of course.

Would

In many respects, the United States is seen to be leading or dominating the world today.

do the trick? Oh, well, I'll try to think of something better :-)

RandomP 04:30, 8 June 2006 (UTC)

Fermi paradox

The first sentence of this was edited. Anything else, let me know. I'm getting some vague comments ("tone is wrong") but the more specificity the better. Marskell 11:04, 8 June 2006 (UTC)

Ruber chiken

I'd appreciate it if you'd keep an eye on edits by ruber chiken (talk · contribs) in the oil area. He's generating considerable cleanup work, as you've noticed. Some of his info is questionable; that big table of OPEC oil reserve estimates doesn't quite match other published tables of that info, and the source he cites is from 1995, yet the table goes to 2004.

I agree with your deletion of his comment on per-capita oil production. That's a real issue that deserves more attention, though. One of the biggest issues in oil today is that China's oil demand is increasing much more rapidly than expected a decade ago, as China switches from bicycles to cars. This is happening just as worldwide oil production seems to be flat. This is just starting to become a source of political tension between the US and China. --John Nagle 19:39, 12 June 2006 (UTC)

Thats nice talling things in poepols back.It was only one section that didn't existed before.From french wiki that sited ther for more than nine months.Whats more important spelling or content.The info is not questionable,on serch engines you run on thies very fast.The big table thing,i didn't check every number,but from out side sources the suspicious hikes are on the same years.If ther are diferences why are you keeping it a secret,and don't notice it on the articles talk page,it's the same table than in french,i just copy paste it.The 1995,obviously he didn't gess the results 10 years in advance,i think they got mixt up,the numbers stops at 2004,so i supose that they ment 2005,but i didn't see into it yet,this isn't important since the hike dates are corect anyway,that's the goal of the table.The per capita comment is very important for me,it will gate back in the article,but in what form,is negotiable.--Ruber chiken 04:45, 13 June 2006 (UTC)

in fact the inicial source run intil 95,from campels.Then somebody added the 2004 period,but didn't mencioned it in the title.The section is more than a year old in french,with no edit wars,so if it survived that long,it can't be rubish.--Ruber chiken 05:12, 13 June 2006 (UTC)

---

Keep in mind that it's Wikipedia policy for editors to watch each other - there's no systematic editing/review process, so we rely on self-organisation.

Having per capita data sounds like a good idea to me, too, though the relation to the Hubbert peak theory is spurious - population just doesn't come into that one.

I'd suggest discussing this at the relevant talk pages, though.

RandomP 08:44, 13 June 2006 (UTC)

It seems to a conspiracy to me "keep an eye".Combind to overexagerations "generating considerable cleanup work".And accusations "info is questionable".Yes he dones everything wrong,he is massing the article.The article is beter now.You didn't anserd,whats rong in the table numbers,and whats more important content or spelling? --Ruber chiken 10:22, 13 June 2006 (UTC)

---

I assure you there's no conspiracy. Please assume good faith when dealing with other Wikipedia editors - it's not just policy, it's a great way to get along with them, too!

I'm not going to comment on your spelling.

RandomP 13:17, 13 June 2006 (UTC)

---

Well on the gold isue,combined with your critics on the hubert peak theory a asumed that.Well if it's not then what are you thinking on the isue.If your persone takes so mach atention,is because of the fact that you are beening a dick.If you want to be a beter wikipedian,begin by agnoleg that you are beening a dick.This is no insalt,it's wikipedia's terms.--Ruber chiken 15:00, 15 June 2006 (UTC)

---

You might find it helpful to read Wikipedia:No personal attacks. If you want to discuss or comment on my edits or comments, there are plenty of ways to do so without resorting to name-calling. I think a good way to do that is to avoid commenting on the contributor, and commenting on specific edits instead.

what the titel of this section?--Ruber chiken 19:02, 15 June 2006 (UTC)

RandomP 18:28, 15 June 2006 (UTC)

You didn't read me carefauly,read this [2].You schould also read this Wikipedia:Civility,you aparently,think,that by just stiking to the rules is enauf,if you stick to the rules religusly,your pissing everybody.When you deleat a contribution,it's an attack to the purson that made it in the first place,just because it's the rule,or just because you think that it wouldn't or schould't be taken that way,is miningless.You have to be diplomatic,if something is unsourced,discuse it first.Because thers no source it don't mean that it's wrogn.When you deleat a corect information,that is unsourced,it's taken even more badly.--Ruber chiken 18:51, 15 June 2006 (UTC)

---

I think I explained why I removed that statement, and I was assuming it had been readded in a more appropriate place than the opening paragraph.

RandomP 19:11, 15 June 2006 (UTC)

Because you asumed that it schouldn't mater,that don't mean that it don't.Your uncivic again.Just go to Carbonate talk page.Read Wikipedia:Civility carefuly.--Ruber chiken 19:45, 15 June 2006 (UTC)

Proving a Negative

This is amuzing but wrong [3].

Carbonate 15:29, 16 June 2006 (UTC)

breton woods

breton woods,was abandoned by the usa because of ther commercial deficits.Not because they don't beleave in gold.Because they where printing more maney,then the gold they had--Ruber chiken 17:27, 16 June 2006 (UTC)

And,stop editing,articles on subjects that you don't poses.--Ruber chiken 18:51, 16 June 2006 (UTC)

don't respond

We'll i was rogn,on the symetry of logistic curvers.Depending on what you throu in the equation,you get what a hell you whant.Logistic curve is the simlest inconstraint model.So the theory produces what a hell you whant depending on the asumtions.You have the hubbert theory right on this page,under you nose;but you faild to understand it.The model makes asumptions,the rest is math,this is thrue for every model.It's an extrapolation.show it to a physicist.--Ruber chiken 17:37, 17 June 2006 (UTC)

Ruber chiken

I merely asked him to make more constructive edits in his native language, as he obviously can't make them in English. I see above that he has written 'The article is beter now.You didn't anserd,whats rong in the table numbers,and whats more important content or spelling?'. The answer is obviously content, but that changes to spelling when the spelling in question is so bad that the article can't even be read, especially by someone who is unfamiliar with the subject (that's why they're reading the article). Some of the edits are completely unreadable, cause ideas to be misconstrued, and helps to cement Wikipedia's unhealthy press reputation as a load of vandalised, unreliable, POV crap. —Vanderdecken∴ ∫ξφ 12:21, 18 June 2006 (UTC)

I agree that those of the edits I've looked at were problematic. However, until a policy is in place to deal with bad spelling, I think we should welcome all newcomers; I must admit that I've managed if not to read, then to decipher all of the edits (or at least, I think so). My personal hope was if ruber chiken stays for a couple of weeks, his ability to write english will improve drastically :-)

RandomP 12:40, 18 June 2006 (UTC)

If you still feel like defending Ruber chiken, I think you may be very interested to see his reply on my talk page: User talk:Vanderdecken#Vander... —Vanderdecken∴ ∫ξφ 10:17, 19 June 2006 (UTC)

I don't feel like defending him at all, for clear policy violations. It's not the bad spelling that annoys me, is all.

Is there any way to get an admin to look into this? He appears to be, so far, a problem editor.

RandomP 11:18, 19 June 2006 (UTC)

I think an RfC might be in order. I don't think an RfAr would be accepted at this time, although there is some evidence that discussion has failed (and mediation was refused by you — albeit, in the absence of agreement from the fowl.) — Arthur Rubin | (talk) 14:15, 19 June 2006 (UTC)

He now appears to have moved on to other articles and, er, currently appears to be introducing new typos. I was looking for a friendly neighbourhood admin to remind him of the NPA policy, in particular.

But I'm not going to spend more time on this, for now

RandomP 15:59, 19 June 2006 (UTC)

"illegal" versus "undocumented" immigrants

Hello,

I noticed you commented insightfully on the illegal immigration talk page about this issue. Wikipedia's policy on whether to use "undocumented" or "illegal" in places where referring to people is unavoidable is inconsistent between and within articles. The issue needs to be resolved somehow. But how?

Personally, I agree that the article's title is probably as good as it will get, but that "illegal immigrant" is a term to be avoided. I wrote a short essay at User:DKalkin/undocumented on the subject, meant to serve as a naming proposal, before realizing that I don't know where to post it in order to try to come up with a consensus. Do you have any suggestions? Where might some sort of central policy on this issue be established? Kalkin 04:06, 26 June 2006 (UTC)

---

Hmmm. As far as I'm concerned, the best way to refer to this group of people is as "people who are in the country illegally". That states a fact, that they're in violation of a law, however much someone might disagree with that law.

My main concern is not to avoid the use of the term "illegal", it's the "illegal immigrant" construction: not only is it potentially confusing, because it might not have been the act of immigration that's illegal for those who overstay a visa; I've also been trying hard to find another analogous construction, and coming up mostly empty.

When there's a curfew for young people in a town, to cite another controversial case, those who stay out are "teenagers on the streets illegally", but they're hardly "illegal teenagers", and shortening that to "illegals" would, to me, seem totally ridiculous.

Is a Christian minister who practices in a country where there are laws against that activity an "illegal priest"? The term appears to be used once, here, in a different meaning.

When the offence is driving faster than the speed limit, taking a bribe, engaging in acts of free speech beyond the limits of law (easily possible in most of the world), or shooting someone in cold blood, the agent, at least when it is a human being, does not take on the adjective illegal: there are no illegal drivers (or illegal speeders), illegal authors, or illegal killers on Wikipedia.

When the offence is, to take the most clear-cut case, entering a country without that country's knowledge or approval, the agent does take on the adjective illegal.

I think that inconsistency is weird, and that's why I would like to use another term.

As to where to go next, I think a discussion not associated to any given article would be the best next step: Set up a page neutrally listing the various proposals, and discuss the problems with each.

I am just not sure whether there even is Wikipedia policy on this. Encyclopedias, of course, strive to use neutral language; in the process, I believe they have historically created neutral expressions where none were available. Today, for the vast majority of topics, neutral expressions are available (even for issues as ticklish as the China/Taiwan/ROC/PRC issue, or the Israel/Palestine one). I believe Wikipedia should attempt to continue that tradition where necessary, rather than giving up in the face of allegedly common usage.

That is something I think needs to be central to any discussion that might happen: that it be a discussion about terminology only.

If any discussion is going to see anything but a minimal response, we're going to see plenty of responses along the lines of

They're doing something illegal, so the term "illegal immigrant" is okay.

The point is that the problem, as far as I'm concerned, is in the "so", not in the first sentence, which should be beyond the scope of a terminology discussion.

I might get around to drafting a proposal for a discussion/RFC page later, but feel free to preempt me :-)

RandomP 13:54, 26 June 2006 (UTC)

---

Thanks for your response. I do not have a problem with the use of "illegal" in the context of immigration without documents in general, as I said. It's a problem if my proposal doesn't make that clear. My problem with "illegal immigrant" is pretty much the same as yours, although I am less exercised by the grammatical issues and more exercised by the dehumanizing impact.

I agree that the ideal solution would be to find some new term which is neutral.

One alternative I've considered is "unlawful." The word at least doesn't have quite the same ugly connotations. But I'm not sure that it's any better gramatically (and therefore in real meaning). Dictionary.com lists "illegal" as the first definition of "unlawful." Law.com's legal dictionary defines unlawful as "referring to any action which is in violation of a statute, federal or state constitution, or established legal precedents," and illegal as "1) adj. in violation of statute, regulation or ordinance, which may be criminal or merely not in conformity. Thus, an armed robbery is illegal, and so is an access road which is narrower than the county allows, but the violation is not criminal. 2) status of a person residing in a country of which he/she is not a citizen and who has no official permission to be there." I don't think there's any substantive difference there - because it's also the case that the law often uses "unlawful" to describe immigrants.

Therefore I'd still say "undocumented" is the least worst. The objections to it are all based on the idea that it doesn't clearly enough expose the wrongness of the action it describes - Wikipedia has to take them into account because them's the rules, but this is not a general kind of objection that Wikipedia can afford to pay too much attention to.

I'm writing a proposal for a naming convention at Wikipedia:Naming conventions (immigration). I'd appreciate your comments on it before I put it up on the RFC list. Kalkin 19:59, 26 June 2006 (UTC)

I've stuck it up on the RFC list and various article talk pages. I got impatient. :0 Kalkin 19:02, 27 June 2006 (UTC)

---

oh, oops. got distracted by something shiny while reading your message. I'll check out the proposal.

RandomP 19:05, 27 June 2006 (UTC)