Any inaccurate information in M40 news reports?
 

I have been very impressed in the quality of reporting of the latest prime number found by GIMPS.
Even so, I thought it would be fun to collect inaccuracies in the reported information. Also, this might be of some benefit in avoiding similar occurrances in the future.

Here is one that I have found so far:
The [B]Detroit Free Press[/B] ([URL=http://www.freep.com/money/tech/nnumb4_20031204.htm]MSU student finds rare prime number during computer search[/URL]) and [B]The Michigan Daily[/B] ([URL=http://www.michigandaily.com/vnews/display.v/ART/2003/12/04/3fcecab6a1fd9]MSU graduate student uses PC to find largest known prime number[/URL]) have an identical paragraph that describes the sequence of prime numbers as beginning with "1." I do not mean any disparagement whatsoever; I have also already seen a correction in the Detroit Free Press: [URL=http://www.freep.com/news/metro/strait5_20031205.htm]Getting it straight[/URL]

What other mistakes are out there?

I've always thought that it's a little picky to get on people's cases for saying the primes start at 1 rather than 2. I like to think of 1 as prime, but trivially so. I mean, it fits the definition, except that for 1, the "1 and itself" both refer to the same number.

It sounds a little picky to me too, but I think it has some merit anyway. I have some recollection of mathematicians taking some pains to correct this in encyclopedia entries, etc. My feeling was that by not making the distinction, many proofs would require more elaborate description. Also it leads to the more precise definitions that partition nonzero integers into three classes: the units (integers with reciprocals in [B]Z[/B]), the prime numbers, and the composite numbers.

Anyway, it isn't a great example, but I felt good enough to start the ball rolling.

Thinking about this just now (and partially agreeing with the sentiment you mentioned), I often see power series expansions that start with a term of 1 that I, myself, would find prettier if it relied on the formula that generated the subsequent terms. For example:
cos x = 1 - x[sup]2[/sup]/2! + x[sup]4[/sup]/4! - . . .
In my wacky way of looking at things, I find this prettier:
cos x = x[sup]0[/sup]/0! - x[sup]2[/sup]/2! + x[sup]4[/sup]/4! - . . .

The article in the Ann Arbor news (from the Associated Press) said the computer used had 2 GHZ of memory. You'd think [I]someone[/I] would've caught that before it was released.

Good catch Kevin.
Also on the topic of computers, some of the articles went into describing the type of collaboration that GIMPS and/or PrimeNet is in slightly different was. Here is a thread in our forum about that: [URL=http://www.mersenneforum.org/showthread.php?postid=15308#post15308]The difference between P2P and distributed computing and grid computing[/URL]. I've seen articles discussing terms P2P, distributed computing, grid computing, and utility computing -- sometimes with slightly different interpretations.

Also on the topic of prime numbers, the [URL=http://www.lsj.com/news/local/031204_numbers_1a-10a.html]Lansing State Journal[/URL] correctly informs readers that the numbers 3, 5, 7 and 11 are prime numbers. I think that leaving 2 out of the list is a result of uncertainty of the correctness of including it; something like "hey, 2 is an even number... that [i]can't[/i] be right."

I didn't like the impression which german news magazine Focus gave to their readers ([url]http://news.focus.msn.de/G/GN/gn.htm?snr=127350&streamsnr=241&streamsnr=241[/url]).

Unfortunately the article didn't show clearly, that everyone, who's interested, could devote their idle CPU time to GIMPS by simply installing a client although they say, that everyone could join.

FOCUS did also mention, that Michael Shafer is a 26 year old [I]mastermind[/I]. But that does [B]not[/B] - i repeat - [B]not[/B] belong to this thread :grin:

That is another story.. and surely distracted more people.

Maybe some readers even got the impression that they have to sit down with pencil and paper and look for prime numbers - and use their PC for some assistance - with GIMPS being just a coordination portal of all that paper work
:wink:

What? I've been doing this wrong all along... I was wondering why my hand was cramping and I kept running out of pencils and paper... :rolleyes:

ehm.. I just tested 2^7-1 and it is prime!

What is my next number to test? :grin:

[QUOTE][i]Originally posted by Dresdenboy [/i]
[B]ehm.. I just tested 2^7-1 and it is prime!

What is my next number to test? :grin: [/B][/QUOTE]
I'll do the doublecheck.

If some of you are interested in science fiction novels, i can recommend [I]Souls in the Great Machine[/I] from Sean McMullen.

In this novel, a huge calculator is build with thousands of humans performing the primitive operations. Even an error correction concept is included.

[QUOTE][i]Originally posted by Dresdenboy [/i]
[B]I didn't like the impression which german news magazine Focus gave to their readers ([url]http://news.focus.msn.de/G/GN/gn.htm?snr=127350&streamsnr=241&streamsnr=241[/url]).

Unfortunately the article didn't show clearly, that everyone, who's interested, could devote their idle CPU time to GIMPS by simply installing a client although they say, that everyone could join. [/B][/QUOTE]

Another way that readers might get a mistaken impression follows from the use some articles are making of this factoid found within the Mersenne.org press release title: "220996011-1 is Found with 25,000 Years of Computer Time"

For example the BBC News UK Edition article ([URL=http://news.bbc.co.uk/1/hi/sci/tech/3256822.stm]Prime position for large number[/URL]) says that GIMPS [I]spent[/I] 25,000 years of computer time to find this prime. This seems to me to imply that the time used was devoted in lieu of other activities. A minor quibble, especially since (although not mentioning idle time) the article mentions elsewhere that the software doesn't affect computer use for other activities at the same time.

Do these 25000 years mean P90 time? It would be difficult to calculate the actually spent calculation time of todays computers because of the many differences.

Actually it would have been better for me to say the press release hosted on mersenne.org. Described thus:
Scott Kurowski, who runs the server that controls GIMPS' massive CPU power, wrote this [URL=http://www.mersenne.org/20996011.htm]press release[/URL].
The title is:
Mersenne Project Discovers Largest Known Prime Number
on World-Wide Volunteer Computer Grid
220996011-1 is Found with 25,000 Years of Computer Time

I think it must be using P90 years but don't really know.

Anyway, I hope no one has taken this thread I started badly. I find it amazing how well the articles I have read cover so many points about the numbers, mathematics, people, machines, software and organizations involved. That is a lot of ground to touch upon. The care taken in preparing for the announcement of this discovery has obviously paid off.

[url]http://www.gridcomputingplanet.com/news/article.php/3287041[/url]

New Mersenne Prime Found

Michael Shafer, a 26-year-old volunteer in the Mersenne.org research project called the Great Internet Mersenne Prime Search, has discovered the largest known prime number.

Shafer used a Michigan State University lab PC and free software as part of an international Grid of 211,000 networked computers.

The new number, expressed as 2 to the 20,996,011th power minus 1, has 6,320,430 decimal digits and was discovered Nov. 17. It is more than two million digits larger than the previous largest known prime number, and belongs to a special class of rare prime numbers called Mersenne primes. The discovery is the 40th known Mersenne prime, named after Marin Mersenne, a 17th century French monk who first studied the numbers.

The new prime is 63% of the qualifying size for the $100,000 Electronic Frontier Foundation award for the first 10-million-digit prime, also being sought by the project's volunteers. In May 2000, a previous participant won the foundation's $50,000 award for discovering the first million-digit prime.

The new prime was independently verified by Guillermo Ballester Valor of Granada, Spain using twelve days of time on a 1.4GHz quad Itanium II server at the HP Test Drive center, and by Ernst Mayer of Cupertino, Calif., using three weeks of time on a 1 GHz HP Alpha workstation. The discovery is the sixth record prime found by the GIMPS project, and the fourth discovered using distributed computing software from Entropia Inc.



"The discovery is the 40th known Mersenne prime, named after Marin Mersenne, a 17th century French monk who first studied the numbers."

I don't think Mersenne was the first to study primes of the form (2^p)-1.

"The new prime is 63% of the qualifying size for the $100,000 Electronic Frontier Foundation award for the first 10-million-digit prime, also being sought by the project's volunteers."

I may seem a little picky here but, this seems to give the impression that a prime big enough to win the $100K is 1.582 (1/0.63) times larger than M(40). In fact, a ten million digit prime would be about (10^3679570.9) times larger than M(40). That is, M10million digits >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>M(40).

When we talk about the sizes of large primes we always use the number of digits (or for the Mersennes, the exponent, which is equivalent), i.e. we talk about the prime in the logarithmic sense. So it's not inaccurate to say M#40 is 63% of the size needed for the 10Mdigit prize.

Re: Any inaccurate information in M40 news reports?
 

[QUOTE][i]Originally posted by only_human [/i]
[B]

What other mistakes are out there? [/B][/QUOTE]

On a portuguese online newspaper there was a nice article about the finding, except that it was stated that Mersenne primes are of the form 2*p-1:shock: :sick: Dooohhh!...

Yeah, lycorn, I know, I've sent them an e-mail about that error last Sunday, asking the webmaster to change the HTML code to 2^20996011-1, and 2^p-1, but so far, [i]nada[/i]... :sad:

I've sent them that e-mail again, rephrased, repeating the request, just now.


The same problem (no superscript for the exponent) occurs in several other online articles. Perhaps they were asleep during math classes? :sleep:

Received a friendly reply a few hours ago, that they've now corrected the article. [i]Sím![/i]

[url]http://www.diariodigital.pt/news.asp?section_id=60&id_news=72674[/url]

While I have not sent any corrections to any of these papers, I have sent corrections to two math related web sites.
One was for a minor spelling error. It was quickly fixed and I received two thank you emails;
The other was for a minor formatting error of 2[sup]n-1[/sup] when 2[sup]n[/sup]-1 was intended. I received a thank you for that and it was also fixed.

I felt very encouraged that they saw and personally responded to my emails especially since I have never contacted either of them before. I feel sure they must be receiving torrents of posts about misperceived problems and others offering quasimathematical [i]proofs[/i] of various things.

[QUOTE][i]Originally posted by Kevin [/i]
[B]The article in the Ann Arbor news (from the Associated Press) said the computer used had 2 GHZ of memory. You'd think [I]someone[/I] would've caught that before it was released. [/B][/QUOTE]

The question becomes:
Was it a 2 GHz processer, 2Gb of memory, or both?

[QUOTE][i]Originally posted by tom11784 [/i]
[B]The question becomes:
Was it a 2 GHz processer, 2Gb of memory, or both? [/B][/QUOTE]
The Associated Press news was probably constructed using the facts from this news release:[url]http://www.mersenne.org/20996011.htm[/url]
In that release is this line:[QUOTE]He used a 2 GHz Pentium 4 Dell Dimension PC running for 19 days to prove the number prime[/QUOTE]
So it was a 2 GigaHertz processor.
Here is another newspaper article drawing from the AP source and containing the same error: (The Michigan Daily) [url]http://www.michigandaily.com/vnews/display.v/ART/2003/12/04/3fcecab6a1fd9[/url]
It would be quite a reasonable mistake to make in this context. Someone would only need to see "2 GHz Pentium 4" and decide to expand it to be more understandable to non-technical readers (unfortunately inserting an error).

Here's another classic error from Ananova: [URL=http://www.ananova.com/news/story/sm_846050.html]Ananova story[/URL]

This is the bit at fault:
[QUOTE]The new number can be represented as 2^20996011[/QUOTE]Well it made me laugh, but it's sad that they clearly don't understand.

Did we have someone that volunteered to contact organisations and quietly put them right?

[QUOTE][i]Originally posted by Reboot It [/i]
[B]Here's another classic error from Ananova: [URL=http://www.ananova.com/news/story/sm_846050.html]Ananova story[/URL]

This is the bit at fault:
Well it made me laugh, but it's sad that they clearly don't understand.

Did we have someone that volunteered to contact organisations and quietly put them right? [/B][/QUOTE]
I do not think anyone has taken on that task. Actually, I am very pleased with the reporting accuracy that has occurred and think that things have gone very well. Many times upon reading technical information in any area that I happen to be well informed I will come across slight inaccuracies and ommissions or even serious blunders. For the most part that doesn't disturb me, especially in the cases that I know what meaning was intended.

I have been looking upon these errors we are collecting in a forensic sense. Sort of just taking the pulse of how mistakes have occurred; how well the information about GIMPS is being disseminated; whether reporting is helping the goals of the project or not.

That specific article covered GIMPS, Mersenne Prime numbers, the discovery of the 40th Mersenne Prime, the discoverer and his background. Furthermore it concludes with a wonderful quote from the discoverer, Michael Shafer: "People don't have to be computer whizzes or math whizzes. Anybody can do this project. That's what makes it exciting. We can get so much more work done with networking projects such as this."

So, while I understand your sentiment, I feel much more encouraged than disheartened.

I've sent Ananova an e-mail with the correction just now. Thanks for spotting that one, Reboot It... :smile:

... and the news editor sent me a friendly reply, telling me the story has now been changed; he furthermore informed me that I am not alone in pointing out this error.:lol: