typodupeerror

## Comment: Re:Numerical instability (Score 1)226

by kevinatilusa (#45488661) Attached to: Ask Slashdot: How Reproducible Is Arithmetic In the Cloud?

Reminds me of one of Lloyd Trefethen's maxims about numerical mathematics (http://people.maths.ox.ac.uk/trefethen/maxims.html ):

"If the answer is highly sensitive to perturbations, you have probably asked the wrong question."

## Comment: Ends? (Score 4, Insightful)999

by kevinatilusa (#45149755) Attached to: US Government Shutdown Ends

Maybe a more accurate headline is "US Government Shutdown on temporary hiatus"? It's only a few months funding, and there's no guarantee we won't go through the entire thing again come January 15th...

## Comment: Re:haha (Score 2)119

by kevinatilusa (#42269229) Attached to: Facebook Changes Privacy Policies, Scraps User Voting

But any competent marketing department would get the hint when 589,141 out of 668,872 people disliked a proposed change.
You need to poll far less than 30% to get a statistically significant result representing the wishes of those 1,000,000,000 idiots.

"Statistically Significant" doesn't really make sense here...that sort of computation assumes that the people being surveyed are a representative sample of all users.

In this case we've got a pretty strong selection bias going on where people who are most upset about the new policy are the most likely to vote.

## Comment: Clever Coffee Dripper (Score 1)584

by kevinatilusa (#41193277) Attached to: What's your usual coffee-making method?

It's an ingenious little device that's sort of a cross between a French Press and a pour-over filter. You pour the coffee and hot water in a paper filter at the top and let it infuse for a few minutes. Once your coffee's sufficiently strong, you place it on top of your cup, which lifts the stopper and lets the coffee drip out the bottom. For more on this. See http://www.sweetmarias.com/clevercoffeedripperpictorial.php for more information.

## Comment: Re:The Best Advertising... (Score 1)716

For me its just the opposite. An advertisement is an attempt to get me to trust the advertiser's word on their product. If they want to convince me, the way to start is by being honest about what they're doing and not try and disguise it as something else.

## Comment: Re:Amazon's search quality is so appalling (Score 4, Insightful)129

by kevinatilusa (#41039105) Attached to: Why Amazon Is Google's Real Competition

If the balance right now is Google's superior search vs. Amazon's superior convenience/prime shipping, I think that still gives the advantage to Amazon.

Amazon can improve their search mechanism over time, but it's much harder for Google to match Amazon's advantages.

## Comment: Re:Astroturfing on Amazon? (Score 4, Informative)129

by kevinatilusa (#41039051) Attached to: Why Amazon Is Google's Real Competition

Of the 35 five star reviews, about 30 were posted in a 1 week period by people who have no other reviews. Of course, each of those reviewers carefully voted up all the previous other 5 star reviews to promote them in the review rankings (so

## Comment: Re:Sport specific -- fencing (Score 1)82

For an epeeist, that's really terrific coverage. I know what I'm looking for, and the announcer/color commentary are just a distraction. For a non-fencer, it must have been terrible.

As a non-fencer, I actually found the epee much easier to follow than the other events (mainly because there was no need to worry about right of way). The other events were enjoyable to watch, but I did a lot of taking the scoring on faith/outright ignoring the scoring and just watching the fencing.

## Comment: There WERE computers involved, indirectly. (Score 1)170

by kevinatilusa (#39990895) Attached to: Goldbach Conjecture: Closer To Solved?

From the abstract of Tao's paper: Our argument relies on some previous numerical work, namely the verification of Richstein of the even Goldbach conjecture up to $4 \times 10^{14}$, and the verification of van de Lune and (independently) of Wedeniwski of the Riemann hypothesis up to height $3.29 \times 10^9$.

Richstein's work (available at http://www.ams.org/journals/mcom/2001-70-236/S0025-5718-00-01290-4/S0025-5718-00-01290-4.pdf ) definitely involves a computer, and I assume the Riemann hypothesis verification does as well.

## + - John Nash's declassified 1955 letter to the NSA->

Submitted by Anonymous Coward
An anonymous reader writes "In 1955, John Nash sends an amazing letter to the NSA in order to support an encryption design that he suggested. In it he no less than anticipates computational complexity theory as well as modern cryptography.

In the letter he proposes that the security of encryption can be based on computational hardness and makes the distinction between polynomial time and exponential time: "So a logical way to classify enciphering processes is by the way in which the computation length for the computation of the key increases with increasing length of the key. This is at best exponential and at worst probably at most a relatively small power of r, ar^2 or ar^3, as in substitution ciphers."

Submitted by Anonymous Coward

## Comment: Ted Chiang's remarkable streak continues (Score 3, Interesting)162

by kevinatilusa (#37163516) Attached to: The 2011 Hugo Awards

Of Ted Chiang's six stories written since 2001, four have won the Hugo award, one was nominated for the Hugo award before Chiang withdrew it from consideration (saying "The story that was published isn’t the story I wanted it to be."), and the sixth was a 1 page speculation for Nature magazine.

## Comment: The app's a little beside the point (Score 4, Informative)36

by kevinatilusa (#35765188) Attached to: Kyoto Prize Laureate Unsnarls Electronic Networks

After all, it deals with a graph whose nodes and connections are already known exactly.

The more interesting part comes when you move to a graph like the link structure or underlying router structure of the internet, which is both orders of magnitude larger and changing rapidly -- even if you could take a perfect snapshot of it, by the time you finished analyzing that snapshot the network would have changed quite a bit in the meantime.

What Lovasz has been doing recently with his work on "graph limits" is providing a framework for analyzing such graphs. You can imagine global properties of the network approaching some sort of fixed equilibrium and hope to analyze that equilibrium without actually knowing the details of how the network is changing. I don't actually know if the work has been used in practical applications yet, but the concept goes far beyond just redrawing planar graphs.

"People should have access to the data which you have about them. There should be a process for them to challenge any inaccuracies." -- Arthur Miller

Working...