There's the billboard in the bay area along 101 that says "Application Extreme Makeovers: 10x HOTTER THAN HELL!", with a giant 10x with flames, and a picture of a hairy pig to the left wearing a scarf and glasses.
What's the deal with this billboard? Are they trying to make the pig look sexy via an extreme makeover, or what? My wife almost drove off the road laughing about how bad that billboard was.
Every time I see one of these articles, I read the fine print just to be sure it wasn't one of my bugs that brought the system down. Looks like I'm probably safe this time.
Anything you can test with a command line, do. Text in, text out, use 'diff result.log official.log' to find if you broke anything. Anything that requires a mouse, though, I hear there's products for that but I've never used them.
I have a pet project (jenny) that generates testcases for cross-functional testing, where the number of testcases grows with the log of the number of interacting features rather than exponentially. It often allows you to do a couple dozen testcases instead of thousands, and it lets you cover more interactions too. It's of no use for pull-down menu interfaces, though, where the goal is pretty much to test every node in the tree and there's no interactions between nodes.
OMG, I write in C and Java, my OS is Windows ME, I'm willing to fix bugs in 20-year-old code, and it's taken me over two weeks so far to write a simple stream cipher + MAC! I must be lousy!
I've long wanted a genealogy wiki. For example, my great-great-great-great-grandfather Benjamin Stinnett has several thousand descendents, about a dozen of which are building family trees, but each researcher maintains their own sets of notes on Benjamin. The wikipedia explicitly does not allow genealogical wikis, except for famous people.
The reason to colonize space is to have more grandchildren. That's plenty motivation.
However, our current mastery of space isn't going to give us any more grandchildren. We need to be able to form colonies that pay for themselves and can build new colonies. They don't have to do everything themselves, but they have to make a profit even after subtracting the cost of the support they get from earth.
Once we can do that, great, colonize space! But in the meantime, we'd be better off developing the needed technology rather than pretending that we already have it. I agree with Van Allen, robot probes are doing us far more good at the moment than our manned space program.
The article claims that since unintentional bugs are so hard to find, it's ridiculous to think that the intentional bugs can be caught. I don't know about you, but I wouldn't be surprised at all if intentional bugs are easier to spot than unintentional ones. I've found that unintentional bugs are often far cleverer than anything I can think up myself.
People are hardwired to be primarily visual. I doubt you can do a hack that will get better throughput than that already existing massively tuned interface. Ears are a close second.
Eyes are actually tuned to analyze tons of info and discard almost all of it. Lots of processing happens before the signal even reaches the brain. I may be wrong; perhaps if the goal is to get lots of info into the brain unmunged the eyes aren't the way to go.
Re:simulate religion in games?
on
Game with God
·
· Score: 1
That's about right. And if there are no recent misdeeds that can be used to explain the misbehaving physics, attribute it to original sin or to God moving in mysterious ways. A very proactive God could make sitting quitely in a chair a challenge.
However, if the rules are tunable, you need to fix all the bugs in your physics simulator in order to deal with the divine_interference=0 setting.
simulate religion in games?
on
Game with God
·
· Score: 4, Insightful
What if you simulate a religion in a game? It would probably come out more like SimCity than Doom, but it seems like a reasonable thing to do. Religion posits that certain things exist, that if x happens you should do y, that if you do y then z will happen. Those could be built into the rules of the game simulator. For example,
God speaks through fortuitous circumstances. When God wants to answer "yes", you happen to find what you're currently looking for.
Murphy's law. If anything can go wrong, it actually does. Always.
The laws of physics aren't reliable because God or Angels keep mucking with things.
The devil can hear your innermost thoughts, but God only hears what you say. Or vice versa.
What goes around comes around. Always.
No good deed goes unpunished.
Cleanliness is next to Godliness.
You can't tell what your standing is until you die. (Might be useful to have several characters and a rewind button, so Iago can kill you off every now and then so you can peek at your standing without going to Hell due to suicide.)
You could configure the game to play by your favorite belief system.
Re:Personally, I would go one step further.
on
Game with God
·
· Score: 3, Interesting
It seems to me that religion should be a natural topic for games. Religions tell you that if you're in situation x, you should do y. Or that if you do y then the world will do z. Religion attempts to model reality. Games... do exactly the same thing.
So, given a religion, the matching game should simulate a world that behaves exactly the way the religion says it should behave. If Christianity can't be made into a game that is believable and enjoyable, um, that says something interesting about Christianity.
There's also the error correction code they're using (Viterbi-encoded convolution codes) which multiplies the amount of data the probe has to send by a factor of 6.
A probe sent today would use Turbo codes or LDPC (Low Density Parity Check) codes, which nearly reach the Shannon limit for channel capacity. The Shannon limit is a theoretical limit -- we could get faster encoders and decoders than LDPC codes (encoders particularly need work), but they're about at the end of the road for channel capacity. (These codes will work on 1s and 0s, but they can also take into account the analog strength of each bit measured to improve their performance. Decoding them involves iteratively forming and testing beliefs about what an encoded codeword should decode to.)
I wasn't able to locate what error rate can be corrected by the Viterbi code they're using now, or how much lower their bitrate could be with the same amount of correction if they were using a code that met the Shannon limit. Help?
I'm not certain about #21, "get the team into ship mode". I prefer the model when development happens continuously, and shipping is done on a branch off of development. That makes shipping a sideshow.
However, a lot of bugs aren't evident until the software is stabler than a development branch usually is. Ship mode is when those bugs are usually caught. Does that make it a good idea to pull all of development off of development and put them on bug fixing in ship mode? I'm not sure.
I do lots of guy-in-a-room development. However, projects that have to meet a tight deadline I give status reports on. The trick to guy-in-a-room development is to have a heap of prioritized things that can be done by a guy in a room, but no deadline for when any one of them will get done. Then the only status that needs to be reported is when something gets done.
I used to invest in Manhattan Technologies (mhtx), which made micro fuel cells that ran off methanol, before they ran out of money and went into hibernation. These things were manufactured using printed circuit technology on plastic sheets, no moving parts, rolled up pretty small. They made prototypes, but never divulged which technical details were keeping them from going production. They've got lots of patents. They'll probably wake up and sue anyone who actually succeeds in manufacturing small fuel cells.
Most people who have computers use them as one tool among many. They don't have to maintain their phone weekly or even monthly, or their hammers, or their sofas. Smoke alarms are supposed to be tested once a month, but who does that?
I have a lot of relatives who used to use computers but have mostly given up on them. What with spam, and viruses, and worms, and trojans, and spyware, I can't blame them. Unless they give you a whole lot in return, they're not worth the hassle.
I asked around recently. Most people have two or three passwords for various purposes (insecure, work, private) that they use for everything. When they can't remember a password, they try all their username x password combinations until something works. But they can remember all their passwords.
I'm experimenting with storing my passwords encrypted on my computer. One file per password. Still two or three passwords that I've committed to memory (for insecure, work, private) for decrypting those files. Now I can use different truly random passwords for every purpose. But I also have to worry about leaving decrypted files around on my computer. I don't worry about temporarily having decrypted files on my computer, on the theory that anyone who could grab those could just as easily watch my keystrokes and catch the passwords as I use them.
I wrote myself a web page (http://burtleburtle.net/bob/crypto/password.html) for choosing passwords by rolling dice or flipping coins. I like flipping coins.
(20*6)^4 = 2x10^8. At 90,000 guesses per second (a number I saw somewhere recently), it would take 38.4 minutes to test all 8-character CVCVCVCV passwords. 22.5 minutes if you treat Y as a consonant instead of a vowel.
Re:EDUCATION IS NOT SCHOOL.
on
The Flickering Mind
·
· Score: 2, Interesting
For most kindergarteners, school and education are and should be about:
reading
getting along with classmates
drawing pictures
playing soccer
unorganized recess
doing arithmetic in your head
And if the kids would prefer not to do these things, they have to be required to do it. They're hard skills that can only be gained by practice. As near as I can tell, a computer won't help with any of them. Maybe after they have these skills a computer would be useful.
If you limit your self-assembling assembler to hydrogen and carbon, there are only a handful of types of bonds it needs to be able to form, there's only two atoms it needs to recognize and consume, plus it needs some way of scavenging power. Throwing in nitrogen and oxygen would square the number of bonds it ought to be able to form.
The industry didn't do a good enough job conveying the benefits. What cropped up in the absence of that public dialog was heightened concern over the risks.
Public dialog = infomercials. Ah-hah! I hadn't made that connection before. Thanks.
Slashdot Mirror
There's the billboard in the bay area along 101 that says "Application Extreme Makeovers: 10x HOTTER THAN HELL!", with a giant 10x with flames, and a picture of a hairy pig to the left wearing a scarf and glasses.
What's the deal with this billboard? Are they trying to make the pig look sexy via an extreme makeover, or what? My wife almost drove off the road laughing about how bad that billboard was.
Every time I see one of these articles, I read the fine print just to be sure it wasn't one of my bugs that brought the system down. Looks like I'm probably safe this time.
Anything you can test with a command line, do. Text in, text out, use 'diff result.log official.log' to find if you broke anything. Anything that requires a mouse, though, I hear there's products for that but I've never used them.
I have a pet project (jenny) that generates testcases for cross-functional testing, where the number of testcases grows with the log of the number of interacting features rather than exponentially. It often allows you to do a couple dozen testcases instead of thousands, and it lets you cover more interactions too. It's of no use for pull-down menu interfaces, though, where the goal is pretty much to test every node in the tree and there's no interactions between nodes.
OMG, I write in C and Java, my OS is Windows ME, I'm willing to fix bugs in 20-year-old code, and it's taken me over two weeks so far to write a simple stream cipher + MAC! I must be lousy!
I've long wanted a genealogy wiki. For example, my great-great-great-great-grandfather Benjamin Stinnett has several thousand descendents, about a dozen of which are building family trees, but each researcher maintains their own sets of notes on Benjamin. The wikipedia explicitly does not allow genealogical wikis, except for famous people.
The reason to colonize space is to have more grandchildren. That's plenty motivation.
However, our current mastery of space isn't going to give us any more grandchildren. We need to be able to form colonies that pay for themselves and can build new colonies. They don't have to do everything themselves, but they have to make a profit even after subtracting the cost of the support they get from earth.
Once we can do that, great, colonize space! But in the meantime, we'd be better off developing the needed technology rather than pretending that we already have it. I agree with Van Allen, robot probes are doing us far more good at the moment than our manned space program.
The article claims that since unintentional bugs are so hard to find, it's ridiculous to think that the intentional bugs can be caught. I don't know about you, but I wouldn't be surprised at all if intentional bugs are easier to spot than unintentional ones. I've found that unintentional bugs are often far cleverer than anything I can think up myself.
People are hardwired to be primarily visual. I doubt you can do a hack that will get better throughput than that already existing massively tuned interface. Ears are a close second.
Eyes are actually tuned to analyze tons of info and discard almost all of it. Lots of processing happens before the signal even reaches the brain. I may be wrong; perhaps if the goal is to get lots of info into the brain unmunged the eyes aren't the way to go.
That's about right. And if there are no recent misdeeds that can be used to explain the misbehaving physics, attribute it to original sin or to God moving in mysterious ways. A very proactive God could make sitting quitely in a chair a challenge.
However, if the rules are tunable, you need to fix all the bugs in your physics simulator in order to deal with the divine_interference=0 setting.
You could configure the game to play by your favorite belief system.
It seems to me that religion should be a natural topic for games. Religions tell you that if you're in situation x, you should do y. Or that if you do y then the world will do z. Religion attempts to model reality. Games ... do exactly the same thing.
So, given a religion, the matching game should simulate a world that behaves exactly the way the religion says it should behave. If Christianity can't be made into a game that is believable and enjoyable, um, that says something interesting about Christianity.
What if John Preskill asks for a set of Wikipedia? That's the best encyclopedia these days. But I don't see how anyone can be given a set of it.
It went over within the last hour.
There's also the error correction code they're using (Viterbi-encoded convolution codes) which multiplies the amount of data the probe has to send by a factor of 6.
A probe sent today would use Turbo codes or LDPC (Low Density Parity Check) codes, which nearly reach the Shannon limit for channel capacity. The Shannon limit is a theoretical limit -- we could get faster encoders and decoders than LDPC codes (encoders particularly need work), but they're about at the end of the road for channel capacity. (These codes will work on 1s and 0s, but they can also take into account the analog strength of each bit measured to improve their performance. Decoding them involves iteratively forming and testing beliefs about what an encoded codeword should decode to.)
I wasn't able to locate what error rate can be corrected by the Viterbi code they're using now, or how much lower their bitrate could be with the same amount of correction if they were using a code that met the Shannon limit. Help?
I'm not certain about #21, "get the team into ship mode". I prefer the model when development happens continuously, and shipping is done on a branch off of development. That makes shipping a sideshow.
However, a lot of bugs aren't evident until the software is stabler than a development branch usually is. Ship mode is when those bugs are usually caught. Does that make it a good idea to pull all of development off of development and put them on bug fixing in ship mode? I'm not sure.
I do lots of guy-in-a-room development. However, projects that have to meet a tight deadline I give status reports on. The trick to guy-in-a-room development is to have a heap of prioritized things that can be done by a guy in a room, but no deadline for when any one of them will get done. Then the only status that needs to be reported is when something gets done.
I used to invest in Manhattan Technologies (mhtx), which made micro fuel cells that ran off methanol, before they ran out of money and went into hibernation. These things were manufactured using printed circuit technology on plastic sheets, no moving parts, rolled up pretty small. They made prototypes, but never divulged which technical details were keeping them from going production. They've got lots of patents. They'll probably wake up and sue anyone who actually succeeds in manufacturing small fuel cells.
What a good idea. In retrospect, it's obvious. I wish I had thought of it.
Most people who have computers use them as one tool among many. They don't have to maintain their phone weekly or even monthly, or their hammers, or their sofas. Smoke alarms are supposed to be tested once a month, but who does that?
I have a lot of relatives who used to use computers but have mostly given up on them. What with spam, and viruses, and worms, and trojans, and spyware, I can't blame them. Unless they give you a whole lot in return, they're not worth the hassle.
Ah? I thought Los Angeles beat Silicon Valley handily on its percentage of autistic children. References?
19 passwords, at last count, just at work.
) for choosing passwords by rolling dice or flipping coins. I like flipping coins.
I asked around recently. Most people have two or three passwords for various purposes (insecure, work, private) that they use for everything. When they can't remember a password, they try all their username x password combinations until something works. But they can remember all their passwords.
I'm experimenting with storing my passwords encrypted on my computer. One file per password. Still two or three passwords that I've committed to memory (for insecure, work, private) for decrypting those files. Now I can use different truly random passwords for every purpose. But I also have to worry about leaving decrypted files around on my computer. I don't worry about temporarily having decrypted files on my computer, on the theory that anyone who could grab those could just as easily watch my keystrokes and catch the passwords as I use them.
I wrote myself a web page (http://burtleburtle.net/bob/crypto/password.html
(20*6)^4 = 2x10^8. At 90,000 guesses per second (a number I saw somewhere recently), it would take 38.4 minutes to test all 8-character CVCVCVCV passwords. 22.5 minutes if you treat Y as a consonant instead of a vowel.
And if the kids would prefer not to do these things, they have to be required to do it. They're hard skills that can only be gained by practice. As near as I can tell, a computer won't help with any of them. Maybe after they have these skills a computer would be useful.
If you limit your self-assembling assembler to hydrogen and carbon, there are only a handful of types of bonds it needs to be able to form, there's only two atoms it needs to recognize and consume, plus it needs some way of scavenging power. Throwing in nitrogen and oxygen would square the number of bonds it ought to be able to form.
Public dialog = infomercials. Ah-hah! I hadn't made that connection before. Thanks.