National Public Radio

[Prime95]

GIMPS will be featured on National Public Radio the morning of April 10. Tune in if you can at about 6:20 and 8:20AM on most east coast stations.

Their web site article: http://www.npr.org/templates/story/s...ryId=102876903

[TimSorbet]

"Not only do you have to multiply a 13 million-digit number by a 13 million-digit number, but you have to do that about 13 million times," Caldwell says. "And that just takes a tremendous amount of computation."

About the "by": he actually said "times", not "by".
And the other part (doing it 13 million times): This is incorrect. You have to do it about 43 million times. I'm surprised Caldwell would make such a mistake (at first I thought maybe it was just a misprint in the transcription, but the audio confirms that he said that) - I thought he'd be well acquainted with the basics of the LL test after writing the Prime Pages and such.
Also, in the audio linked on the page the reporter says:

The largest prime is actually 12,978,189 digits.

I think we can all spot the mistake there...
And the text of the page here is incorrect:

To get an idea of just how big this prime number is, if you write 10 digits per inch — all 12,978,189 of them — the number would extend for 20.45 miles.

The number should be 20.48 miles (12978189/10/12/5280=20.48325284090909...; the audio has this correct).

Makes you wonder how many mistakes there are in the news about subjects you're not familiar with...

[Flatlander]

Excuse the off-topic but:

Originally Posted by Mini-Geek

...
Makes you wonder how many mistakes there are in the news about subjects you're not familiar with...

My dad said something very similar about 30 yrs ago. He was a keen metal-detect/ive/or/ist and bought a magazine about a local find. It was full of errors. I've learned to be very cautious.

[Prime95]

Originally Posted by Mini-Geek

You have to do it about 43 million times. I'm surprised Caldwell would make such a mistake

Actually, it is easy to see how this slip happens. Caldwell (and I) think in terms of bits. That is, he knows instinctively that an LL test is "square X-bit number X times". But now he is questioned by a layman using digits and it is easy in an interview to mistakenly substitute digits for bits. I'm sure he would never make such a mistake in a published paper where he has more time to review his statements. Note that he was quick enough to substitute "multiply" for "square" to simplify his answer for his target audience.

P.S. This is why I hate interviews. I am not particularly good at quick thinking. Maybe we should appoint quick-witted Ernst as the official GIMPS spokesman!

[Jeff Gilchrist]

Freaky, I was just going to post that link here. My father (who also lives in Canada) found that and forwarded it on to me. Some nice national (and slightly International) coverage.

[Spherical Cow]

Despite the minor error, the interview and segment sounded very good. I was pleasantly surprised to hear it while riding in to work this morning, and it's prominently positioned on the NPR home page right now. Good job. It might add a few processors to the effort.

Norm

(But Ernst would make a fascinating spokesman...)

[ewmayer]

Originally Posted by Prime95

P.S. This is why I hate interviews. I am not particularly good at quick thinking. Maybe we should appoint quick-witted Ernst as the official GIMPS spokesman!

Ah, but that assumes (quick-witted == glib) ... and I'm not sure I have the necessary level of "glibido" for a drive-by-shooting-speed discussion of M-primes. But you do flatter me extremely, sir.

Anyway, I'd probably break in the midst of the tech-babble and start ragging on Alan Greenspan and U.S.-style Ponzi-conomics...

[jasonp]

Originally Posted by Spherical Cow

(But Ernst would make a fascinating spokesman...)

"So Ernst, do you think about the economy when you're not crunching prime numbers?"

[philmoore]

Nice article, I enjoyed hearing the voice of the mysterious George Woltman, after corresponding with him via email for 10 years! And I liked Chris Caldwell's characterization of Mersennes as "the jewels of number theory." I also thought that the follow-on article at the NPR website by Andrew Prince was good.

Purely coincidentally, I was interviewed live on our local public radio station KLCC this past Wednesday with regard to our recent probable prime records ("Five or Bust".) I made sure to mention the current Mersenne record as well. The show was a local one, Northwest Passage, and the setting was informal as the station was wrapping up its fund drive. (Unfortunately, it was not archived for that reason.) The hosts were very quick on their feet. As they returned to the drive, Tripp said, "We're not talking about numbers quite that large, but feel free to add a few zeros to your pledge amount!" Then Angela added "If you contribute online, we will make sure we use prime numbers to keep your credit card information secure." They also publicized a public lecture I gave yesterday (Thursday) entitled "Prime, Perfect, and Sierpinski Numbers". The powerpoint slides will soon be hosted, and I'll post a link for anyone who would like to see them.

[philmoore]

Here's the link to my presentation slides:

http://www.lanecc.edu/fpd/documents/Moorecolloquium.ppt

I'll post this in the "Five or Bust" subforum as well.

[xorbe]

This link was featured somewhere on Yahoo prominently today:
http://www.npr.org/templates/story/s...ryId=102876903