## Pi Day and an Interview With a Pi Researcher 188

Posted
by
timothy

from the much-depends-on-your-date-format dept.

from the much-depends-on-your-date-format dept.

JoshuaInNippon writes

*"In honor of Pi Day, March 14 (or 3.14 for those who may need a hint), readers may be interested in reading an interview with Professor Daisuke Takahashi, the Japanese researcher who found 2.5 trillion digits of Pi back in August, before being apparently being edged out in December by a French computer programmer looking to prove his efficient coding abilities. Professor Takahashi's interview gives some unique insight into one man who truly marvels at the number that has driven people to ever greater lengths to find more digits for centuries."*Plant Kingdom adds*"There have been a number of proposals for alternatives to March 14 (see the Wikipedia page for Pi Day). Here's mine: when the Earth has gone through 1/pi-th of its orbit, as measured from Winter Solstice to Winter Solstice. I've put together a web site to make the case."*
## I don't get it (Score:5, Informative)

Huh? Pi isn't 14.3 or 14/3.

## My suggestion (Score:2)

## Re:My suggestion (Score:4, Funny)

## Re: (Score:3, Funny)

## I just want to celebrate "i" day (Score:2)

Except I don't know when. I could celebrate i^2 day, but then it would always be yesterday , and I'd miss it.

## Re: (Score:2)

What about e^(i pi)+1 [wikipedia.org] day? That would always be today.

## Re: (Score:2)

e, pi, they're both important numbers. Perhaps they both deserve their own days. Now, I'm off to bed. I stayed up too late. But before I go, here's a joke. Please let me know how bad it is.

Dream food: ln(-1)

## Re: (Score:2)

Eye pie? That sounds kind of gross.

## Re: (Score:2)

Imaginary pie.

## Re: (Score:2)

I was thinking the same thing, since pi is 180 degrees and the earth has ~ a round orbit.

1/2 of its orbit makes sense.

## Re: (Score:3, Informative)

Huh? Pi isn't 14.3 or 14/3.No, but it's close to 22nd July.

## Re: (Score:2)

Huh? Pi isn't 14.3 or 14/3.No, but it's close to 22nd July.

I refer to July 22 as Pi Approximation Day, as 22/7 is approximately Pi - it's close enough for most things

It's also nearly my birthday, so another reason to celebrate. :-)

## Re: (Score:3, Funny)

It's also nearly my birthday, so another reason to celebrate. :-)Surely you mean approximately your birthday?

## Re: (Score:2)

No, obviously Pi day is January 3rd. Though to be completely accurate, Pi day won't occur until January 3rd, 4159. A Wednesday, for anyone who's interested.

## Re: (Score:3, Funny)

In Indiana, it's celebrated on March 2.

## Re: (Score:2)

What? 03014159? The digit 0 does not appear in pi until the 32nd position after the decimal point. Surely you're not trying to equate 0 with a decimal point?

Please just let go of this archaic month/day/year standard. Obviously it's not even used consistently - 03/01 vs. January 3? If the month should come first when spelled out, how can one with clear conscience insist that it come second when in numeric form?

Really, why deny any day its connection to pi? Find any reason at all to equate any date with

## 2010-03-14 per ISO 8601 (Score:2)

## Re: (Score:2)

Year-Month-Day as in the ISO standard it's 2010-03-14.

## ISO too (Score:2)

idiot american system of month.day instead of day.month

Actually, since we're ignoring year, the ISO 8601 [wikipedia.org] format (yyyy-mm-dd) makes this Pi Day too--and given that there is an ISO date format, and given that it solves the sorting problem so thoroughly, I'd say that a format which uses day.month is, at best, less stupid. But still pretty stupid.

## Re:I don't get it (Score:5, Insightful)

No, day-month-year is more sensible since the size of the unit is increasing monotonically. The sanest way to do it would be year-month-day, because then you could increase the precision of the time string to whatever you needed just by adding units to the right. The month-day-year system is probably the lease sensible method of the lot.

## Re:I don't get it (Score:4, Interesting)

The month-day-year system is probably the lease sensible method of the lot.

Not to those of us who often work with dates that often land on the next month. As a friend of mine likes to say "six of one, half dozen of the other."

The sanest way to do it would be year-month-day...

This gets my vote.

## Re: (Score:2)

## Re:I don't get it (Score:4, Interesting)

Yup, yyyy-mm-dd is the ISO standard date format for a reason. You get the advantage of easier chronological sorting (ala the US system of month/day), and the unambiguity of the unit size constantly going in one direct (in this case, largest to smallest).

## Re: (Score:2)

!dolc evitisnesni uoy tfel ot thgir daer I

## Re: (Score:2)

The US system is dreadful for chronological sorting. It only holds up if you don't go past 31st Dec, and then in time you end up with a bloody mess that is hard to sift through, which is only marginally better than dd-mm-yyyy.

## Re: (Score:2)

Who sorts dates as if they were strings/numbers?

Doesn't the date type come with its own rules for sorting? If not, why not?

## Re: (Score:2)

Um, firstly the parent mentioned this already.

You get the advantage of easier chronological sorting (ala the US system of month/day)

Secondly, you can't sort a date

alphabetically because dates are comprised of digits, not letters of thealphabet.## Re: (Score:2)

Yeah, and continuing to use Imperial also has a mile-long list of disadvantages, yet here we are.

## Re: (Score:2)

The sanest way to do it would be year-month-day, because then you could increase the precision of the time string to whatever you needed just by adding units to the right

ISO standard date format is YYYY-MM-DD. Everything else is non-standard's compliant gibberish of the kind that under most circumstances /.'ers would be violently opposed to (especially if Microsoft did it.)

ISO 8601 [iso.org] is the full standard, which includes times in the form:

YYYY-MM-DDThh:mm:ss (the letter "T" is often replaced with a space for human-readable but non-standard date/time presentation).

I have no idea if the standard has accommodation for the archaic AM/PM system of representation.

There is really no

## Re: (Score:2)

I have no idea if the standard has accommodation for the archaic AM/PM system of representation.

True UTC has no sense of AM/PM.

## Pi day? (Score:5, Insightful)

Judging by the big hunk of meat in my 'fridge, today is Steak and BJ Day [steakandbjday.com]. Pi day just isn't nearly as fun.

## Re:Pi day? (Score:4, Funny)

Pi day is irrational, but at least it's real.

## Re: (Score:3, Funny)

Judging by the big hunk of meat in my 'fridge, today is Steak and BJ Day

Please tell me the "big hunk of meat" in your fridge is for the steak part of that day. You're doing it wrong if not.

## Re: (Score:2)

I always have to eat pi before a bj.

## Don't forget. (Score:4, Funny)

## Re: (Score:2)

## US-centricity (Score:5, Funny)

## Re: (Score:3, Informative)

TFA suggests Archimedes' approximation of 22/7

## Re: (Score:2)

There is quite likely a plethora of different dates in our calendar year that map to the fist digits of pi in decimal form, or the digits in its improper fraction approximations. So yes, there will likely never be any agreement on what is the "real" pi day.

Not to mention /. isn't where one should seek consensus anyway!

## Re: (Score:3, Funny)

#define PI_VALUE 22/7

## Re:US-centricity (Score:5, Funny)

#ifdef INDIANA

#define PI_VALUE 3

#endif

## Re: (Score:2)

355/113 is better

## (22.0/7) (Score:2)

## Re: (Score:2)

Still, it may be a good idea you posted that, in case someone tries something like that expecting a certain result.

In the case of PI, I would rather suggest checking if the includes define such constant instead, and otherwise fall back to a proper PI define.

## Re: (Score:2)

## Re: (Score:2)

## Re: (Score:2)

Just celebrate on Quadectober 3, duh.

## Re: (Score:2)

I think that 22nd July is more appropriate because we can only ever approximate pi anyway.

## Re: (Score:2)

## Stupid (Score:4, Insightful)

This is stupid, who thought MONTH - DAY - YEAR is a reasonable date format? Do you frequently find yourself asking "Hmm, I wonder what month it is?" And always make people look in the center to find out the date? WTF

It's like throwing away metric and using some crazy-ass divisible by 12 unit.

## Re: (Score:2, Insightful)

This is stupid, who thought MONTH - DAY - YEAR is a reasonable date format?

The person that noticed that we generally say "March 14th, 2010", so it's more intuitive.

It's like throwing away metric and using some crazy-ass divisible by 12 unit.

Oh please, it isn't done right where you are, either. It should be 2010-03-14 so it sorts chronologically

andintuitively can't get the month or day mixed up. Your preference is simply different, not less stupid. This isn't one of those topics you can use to pose as one of the smarter people.## Re: (Score:2)

But we don't. Americans/Canadians say that. Everyone I know says "14 March" or "the 14th of March". I've heard the "but that's how you say it" argument for the US date format before, but it's simply untrue. Other English speaking countries generally say the date before the month (or, just the date ... "the 14th").

And those accursed cheap electronics (like wristwatches and low-end VCRs and stuff) that ONLY display in month/day (with no year) are terribly annoying because if it's before the 12th of any month,

## Re:Stupid (Score:4, Informative)

Bad form replying to one's self but this is interesting: http://en.wikipedia.org/wiki/File:Date.png [wikipedia.org]

The number of countries using the US system is pretty small. It's basically the US and a few random places like Palau and Micronesia.

But - there are quite a few variations on date format, more than I thought :)

## Re: (Score:2)

Actually I'd argue that the accursed cheap electronics are that way because they were developed in Asia. They're not showing mm-dd-yyyy with the year missing, but rather yyyy-mm-dd with the year missing.

Still bloody annoying and confusing for everywhere else in the world.

## Re: (Score:3, Funny)

Base 12 is an excellent base. Metric/decimal is okay but we're not monkeys anymore. Can't we please get over the fact that we have just 10 fingers? Base 12 makes it a lot easier to work with common fractions. Halves, quarters, AND thirds are all easy to calculate. Assuming you accept that the inch is no more or less arbitrary than the centimetre as a unit of measure, then in base 12, feet and yards become completely sensible. A great gross of yards (12^3 = 1728 yards = 5184 feet) is pretty close to to

## I think not quite right. (Score:2)

## Re: (Score:2)

What is a metre? Traditionally, one ten millionth of a quarter of the circumference of the earth. Now, it is defined as the distance that light travels in a vacuum in 1/299,792,458 of a second. At some point, somebody has to make an arbitrary decision about how to define distance. So what is an inch? Who cares? If we must be pedantic, it's the distance that light travels in a vacuum in 10,000/76,147,284,332 of a second.

## Re: (Score:2)

Meter and the metric system are human attempts to bring some order into the chaos of the imperial system.

As there is a single ruling body that determines the length of a meter they could just as well have had it twice as long it would make nearly no difference.

It would still be more practical and more precise than the imperial system.

Why is it more precise?

Because it is determined by a single ruling body, because its properties can be based on unchangeable measurable properties of atomic particles AND artif

## Re: (Score:2)

What the fuck is an inch?

Unzip and you'll find out.

But you had to bring fucking into the discussion.

## Re: (Score:2)

I prefer writing yyyy-mm-dd, especially for anything programming-related.

In regular writing, I'm fine with seeing dd-mm-yyyy. mmm dd, yyyy is still okay since there's no ambiguity about what which the day and month are, but this leads to the laziness that results in....

mm-dd-yyyy or mm/dd/yyyy. Which are about the stupidest written conventions I've ever come across. Along with imperial measurements, why is the US hell-bent on foisting backwards systems on the rest of the world?

## Re: (Score:2)

"It's like throwing away metric and using some crazy-ass divisible by 12 unit."

Or: Metric is like throwing out base 60 and using some unit that you can't divide evenly by 3, 4, or 6.

## a French computer programmer? (Score:2, Interesting)

## Re: (Score:2)

Yeah even I thought the summary doesn't do justice to him. He found algorithmically faster ways to compute Pi [bellard.org]. I doubt this was just to 'prove his efficient coding abilities.' He's someone to be honored for giving us LZEXE, FFMpeg and QEMU. The summary treats him as some random guy, which is weird.. this being Slashdot and all.

## Re:a French computer programmer? (Score:4, Informative)

OTOH, reading Bellard's FAQ [bellard.org] on his latest result does seem like he was interested more in fast algorithms and not in Pi. So I stand corrected. Still.. he's not some random programmer to us. :P Following links from his FAQ, I found two cool books:

## What's the significance of 1/pi? (Score:3, Interesting)

1/pi is not pi. It's like celebrating the third time something happened by doing something one third of the way through and then stopping!

"When a circle's diameter is one unit, then the cirmcumference is pi units." [http://en.wikipedia.org/wiki/Pi]

So if a year is "one unit", we should celebrate pi every 3.14 years or something.

## Re: (Score:2)

``1/pi is not pi.''

pi r square!

No, pies are round!

## Re: (Score:3, Informative)

For every pi proportioality the reciprocal relation is a 1/pi proportionality, so it's every bit a significant. For a near-circular Earth orbit 1/pi of a year is the time when the earth has passed the equivalent of the diameter of it's orbit and swept an area the size of of a rectangle made from it's major and minor axis.

## RE: yeah, okay (Score:5, Insightful)

Okay, I'm about to troll, but...

First, do we really need a holiday for every fucking thing out there? Where the hell is dung beetle day? Aardwolf week? Permian Extinction Day?

Secondly, you invested enough energy into worrying about WHICH day should be Pi day that you created a website over it?

Finally, if Pi gets its own day, I think its entirely fair that 1.618 [wikipedia.org] get its own celebration. Phi is easily as fascinating a number as Pi, so why didn't you get your panties in a twist over not having Phi day?

Sometimes, you CAN be too much [gocomics.com] of a geek.

--------------

End of Troll.

## Re: (Score:2)

## Re: (Score:2)

April 15, in the US, if my pronunciation is correct.

## Did he find a message? (Score:4, Interesting)

I understand that in the book "Contact" by Carl Sagan, when the scientists meets the aliens he asks them a question:

Scientist: Do you believe in God?

Aliens: Yes.

(Astonished) Scientist: Really?! Why?

Aliens: We have proof.

Scientist: Proof?!!!

Alien: Yes, when we decoded Pi to (a very large number) we found a Message...

Of course this idea was exploited in a different way by the movie "Pi". (Sorry didn't see it either.). In any case, if Pi is truly Random (it is isn't it?) won't every possible message occur? Just like those monkeys with their typewriters (if you don't know what a typewriter is look it up).

## Re:Did he find a message? (Score:5, Funny)

What the fuck is a monkey?

## Re:Did he find a message? (Score:4, Interesting)

In any case, if Pi is truly Random (it is isn't it?) won't every possible message occur?Hmmm ... You must be using an unusual definition of "random", which usually means that the value is unpredictable. Pi is the opposite of random. It's precisely defined and always has the same value for anyone who calculates it correctly. (Which leaves out the religious folks, but that's to be expected for anything requiring validity. ;-) Pi would even be the same in a different universe with different physical laws, because its value isn't dependent on anything physical.

As for every message occurring, I think you're thinking of normal numbers [wikipedia.org]. There is a conjecture that pi is normal, but it hasn't been proved. So far, statistics of the digits of pi are consistent with it being normal to as many digits as have been tested. A normal number does contain every possible message, in every possible encoding. If pi is normal, then so is e. [The proof is trivial for anyone who knows the well-known equation relating e and pi.]

And yes, this mathematical (ab)use of the word "normal" is one of the silliest things that mathematicians have ever done. But there is a long tradition of such silly misuse of common words as mathematical terms.

## Re: (Score:2)

Ok, you're right, I wasn't using the right definition of random. Thanks for telling me about "normal" numbers!

## Re: (Score:2)

No, it doesn't. What does have a different value is the ratio of a circumference to it's diameter, that can be grater than pi (hyperbolic surfaces), or smaller (spherical surfaces). But that ain't the definition of pi, 'cos even in a given surface of constant curvature, this value varies.

## Re: (Score:2)

A little knowledge is a dangerous thing, and that "umm, no" comes across as a little arrogant.

The ratio of the circumference of a circle to its diameter differs from Pi in a non-Euclidean geometry, but that's not to say that the value of Pi intrinsically depends on the geometry.

The *definition* of Pi is the length of the circumference of a circle with diameter 1 in *in the Euclidean geometry*.

We could just as easily take the definition of Pi to be the limit of the sequence

4 (1 - 1/3 + 1/5 - 1/7 + 1/9 -...)

w

## Re: (Score:2)

Actually, I have heard of Buffon's needle. We might note that if one looks at the equations that turn up in probability and statistics, one sees a lot of instances of pi (and e).

Pi might have been originally defined in terms of the diameter and circumference of a circle (in the Euclidean plane), but that's not the only way it can be defined. Any of the many other equations containing a related "magic number" would do as well. The Euclidean circle is just one of the simplest things that uses this number,

## Re: (Score:2)

## Re: (Score:2)

It makes a great story, yet one requiring a large dose of suspension of disbelief, since we know that pi is not arbitrary, but logically necessary. Even God could not insert a message into pi.

## Re: (Score:2)

Yes, I've seen those decodings. In base 8 it's a glowing testament to the crystalline clarity of the signal/noise ratio of the future universe when it cools to micro-degrees Kelvin. Decoded in base 6, it's a screech of satanic positrons travelling backward in time to the original inferno.

Decoded in base seven, it contains the message "these digits intentionally left blank".

## new date... (Score:2)

I vote for July 22nd.

## Pi *Researcher* (Score:2)

What exactly is there to research about pi? Its digits are random. They change simply by switching to a different base so no pattern can be truly meaningful just an artifact of the chosen representation.

## Re: (Score:2)

It is primarily a way to test computational algorithms.

## Re: (Score:2)

No, they are perfectly deterministic.

To calculate the numerical value of the digits I could understand as useless, but there are plenty of properties that are basis-independent, like it's irrationality, transcendence and normality.

Also, there are some quite beautiful representations, look for the continued fractions one.

## Re: (Score:2)

Well, it's still an unsolved problem as to whether Pi is "random" (that is, whether it's a Normal number [wikipedia.org] in base 10. Even if it is, it may not be normal in another base.

## And now our Pi Day anthem! (Score:2)

pi - Kate Bush [youtube.com]

This version only takes six minutes to sing. The extended version? You don't even want to know.

## Party Time??? (Score:2)

I hope they're already planning for a really, really big Pi Day blow-out in 2016.

## 1/pi of the orbit? shwa? (Score:2)

Um, if you wanted to use a circle as a metaphor for the year, then 2 pi radians would be a full circle, so wouldn't it make more sense to make it the day half way through the year (pi radians)?

## Re:Ellipse != Circle (Score:4, Informative)

Pi is relevant to the circumference of circles. The earth has an elliptical orbit. http://en.wikipedia.org/wiki/Circumference [wikipedia.org]

Um...pi is relevant to a lot of things, including ellipses. And besides, the orbit of the Earth has very low eccentricity, meaning it is very close to a circle. Who modded the parent "informative"?

## Re: (Score:3, Funny)

Clearly it was a case of circular logic.

## Re: (Score:2)

Clearly it was a case of circular logic.

I've always found it peculiar that circular logic is not a degenerate case of elliptical logic. :)

## Re: (Score:3, Informative)

Who modded the parent "informative"?This is Slashdot, in case you hadn't noticed. Only on Slashdot is pointing out the trivially obvious considered "Informative."

## Re: (Score:2)

## Re: (Score:2)

## Re: (Score:2)

## Re: (Score:3, Informative)

If the doubling time of computational power is N hours (a bastardization of Moore's law, approximately 17500 hours.. or two years), then it never makes sense to start a calculation that will take more than 2N hours.

For easy visualization of this concept, lets suppose you have a program that will take 6 years to complete if run on todays fastest hardware, and you begin it March 14th, 2010

But you have an adversary who wants

## Re: (Score:2)

Of course, if you never start the computation

at allthen youneverget results.The terminating condition is knowing when Moore's law will fail.... far enough in advance to know when it is optimal to begin computation.

## Re: (Score:2)

Of course, if you never start the computation at all then you never get results. The terminating condition is knowing when Moore's law will fail.... far enough in advance to know when it is optimal to begin computation.

Sure..

AFAIK, nobody has yet done any 4+ year computations on silicon, or 2+ year computations for that matter. The record holder for Pi is I believe 2.7 trillion digits and took 131 days, and this wasnt even a super computer.

## You just got pwnd by Captain Literal. (Score:2, Funny)

I've got 0 through 9 here on my keyboard, I'd have thought with a trivial bit of rearrangement they'd just about do it.

## Re: (Score:2)

Came here looking for radians. Now disappointed that they're only mentioned this far down and by and AC.

One "pi-th" is 2 radians.