Forgot your password?
typodupeerror

Comment: Error in summary? (Score 1) 133

by Strilanc (#39773665) Attached to: Brain Scan Can Predict Math Mistakes

The summary says the prediction is made 20 seconds in advance. But the source says it takes 20 seconds to make the prediction. That's a pretty significant difference.

From PhysOrg:

[...] Cirett was able to detect with 80 percent accuracy whether a student [...] would answer a question incorrectly about 20 seconds after they began the question.

Comment: Re:Quantum Internet (Score 1) 156

by Strilanc (#39659845) Attached to: The First Universal Quantum Network

Entanglement allows you to instantaneously share a bit of information, but doesn't let you control what the bit will be. This makes it useless for most communication tasks.

That being said, entanglement can allow coordination in a way that is similar to communication. Check out the Wikipedia article on "quantum pseudo telepathy".

Comment: Re:Wishful thinking (Score 1) 249

by Strilanc (#30201794) Attached to: After 35 Years, Another Message Sent From Arecibo

It's really just a matter of trying to make things obvious. You send simple messages at first, allowing the receiver to guess at your format and confirm it with later messages. Then you start sending more complicated messages in the same format.

For example, use a character set of {00,01,10,11}, which I'm going to simplify as {0, 1, space, next} for display purposes. You receive:

0 0 0 0
0 1 0 1
0 0 1 1
0 1 1 10
0 10 1 11
0 10 110 1000 ...

1 0 0 0
1 1 1 1
1 1 0 0
1 0 1 0
1 10 10 100
1 11 11 1001 ...

Get it? Atoms would be done the same way. They even have an implicit order for sending!

Comment: Re:Great on paper - but in real life? (Score 1) 227

by Strilanc (#29989120) Attached to: Maryland Town Tests New Cryptographic Voting System

Not everyone has to verify their vote. An attacker will have to throw away a large number of ballots in order to sway an election. If each voter has a 5% probability of checking their vote and only 100 votes are thrown away, the probability that the attacker is at least detected is greater than 99%.

There's also no need for perfection. The number of reports will be higher when the election is attacked. Apply basic statistics to figure out how likely it is the election was stolen instead of just people making mistakes.

Comment: Re:Programmer Thinking (Score 1) 121

by Strilanc (#29857093) Attached to: Open Source Voting Software Concept Released

I am constantly amazed at the cynicism from Slashdot about electronic voting. Yes, the existing systems have generally sucked. A lot. But that doesn't mean it's impossible to do.

The fundamental problem which must be addressed is verifiability. In order for the election to be secure, you must have a process which guarantees that tampering will be detected with high probability, **even if a malicious company designs a large portion of the voting machines**. This is not an impossible problem!

For example, suppose you want to be sure your vote is not being flipped. A system which meets that criteria is for the voting process to go as normal, then the machine encrypts and publishes the vote (accessible from the internet), which you verify. Then you are prompted to either check or commit the vote. If you 'check' the vote the machine publishes the encryption keys for the vote, and if you 'commit' then the machine publishes a commitment. All of these actions can be verified by you.

Now, that system isn't perfect (it can compromise the 'secret' part of the ballot). But there are verifiable systems which meet all the criteria for an election. They are not less secure than paper ballots, they are *more* secure. Don't trust the machine, trust the verifiable actions the machine must perform.

Comment: Re:Too early yet (Score 1) 334

by Strilanc (#29626263) Attached to: Legal Code In a Version Control System?

That might be interpreted as apathy. It would be better to vote for a *random* candidate in that case. It's unlikely to change the outcome of the election, even if a bunch of people do it, but it sure makes it look close!

In cases where it does swing the election... well, then its likely so many people were misinformed that random chance is better at selecting and "putting the fear" in leaders anyways.

Comment: Re:Simple (Score 1) 365

by Strilanc (#28785867) Attached to: Visualizing False Positives In Broad Screening

That's overkill. We just need a more accurate measurement to communicate. For example we could define "trustability-ishness" to be Min(P(true | tested true), P(false | tested false)).

So, for example, a 99.9% accurate filter for terrorists when the terrorist rate is 1 in a million:
true positives = 0.000001*0.999 = 0.000000999
true negatives = 0.999999*0.999 = 0.998999001
false positives = 0.999999*0.001 = 0.000999999
false negatives = 0.000001*0.001 = 0.00000001

P(true | tested true) = 0.000000999 / (0.000000999 + 0.000999999) = ~0.000998
P(false | tested false) = 0.998999001 / (0.998999001 + 0.0000001) = ~0.999999

"trustability-ishness" = ~0.000998 = ~0.1% out of 100%

Wow! That 0.1% sure makes things a whole lot clearer. I'm sure there are better definitions, but you get the idea.

Comment: Re:Ok, so... (Score 1) 895

by Strilanc (#28603019) Attached to: Researcher Trolls MMO, Surprised When Players Hate Him

That's what I said in the first paragraph. I agree with you. We invent rules to stop games from degenerating into spam-fests. They are useful because the game is funner that way.

My point is that the article contradicts itself. It says all those extra rules are for scrubs, then it says some aren't. My issue is with the start, not the end.

Comment: Re:Ok, so... (Score 1) 895

by Strilanc (#28602077) Attached to: Researcher Trolls MMO, Surprised When Players Hate Him

Wow. That's a very poorly structured article. I get his point. He thinks 'scrubs' have too many arbitrary rules, that those types of rules take away from the depth of the game, but that there are extreme cases where such rules are required. But, as written, the article contradicts itself.

He starts by saying: if it's possible you USE it, or you LOSE. All or nothing. If it's overpowered, then the other guy should stop crying and use it right back. Arbitrary rules are for sissies. Then he talks about how we need arbitrary rules in extreme cases like game-freezing glitches and codes to play as boss characters. So which is it!? Stop crying and pick the boss character, or implement an arbitrary rule?

Can't open /usr/fortunes. Lid stuck on cookie jar.

Working...