Time to plug my free pairwise testing tool jenny (it jennyrates tests). 10 on/off dimensions with three constraints (-w1a2b3a -w1a3b5a -w1a5b9a) took 8 or 9 testcases, depending on the seed to the random number generator.
Pairwise testing takes fewer testcases than Taguchi's orthogonal arrays. Orthogonal arrays guarantee every that every dimension setting appears an equal number of times (does it guarantee that for pairs of dimension settings too?) N-way testing just requires that every combination of n settings is covered at least once. Detecting bugs only requires n-way testing, where n is 2 or 3.
I wouldn't be surprised if it's something in their genes that turns off evolution. That there are genes that control how fast mutations can appear. Some species have these genes turned all the way up. Perhaps others have it turned all the way down. What's the mutation rate of blue-green algae?
I just bought a TI-36X II from Costco. And promptly lost the manual. Most of the functions I haven't been able to figure out how to get at despite a fair amount of effort. And those that I can get at are very cumbersome. For example, decimal to hex is: enter number, hit "enter", hit "2nd", hit 8 (which means hex), read the answer, clear, hit "2nd", hit 7 (which means dec), enter next number... . The keys are small and close together too and don't give me a reassuring click to let me know I've pressed them. Even turning it off is tricky enough for me to usually need to attempt it more than once.
I recommend steering clear of the TI-36X II no matter what you're looking for in a calculator.
I also have a 17-year-old TI-55 III, and remain quite happy with it. I can't imagine why TI went in for such a bad user interface as this TI-36X II.
Woah, spoke too soon. Serves me right for believing Slashdot. The business world is reporting this as an outrageous defeat of business by open source advocates: http://biz.yahoo.com/djus/030924/1303000668_2.html
Whether you can issue a patent on a pure software invention isn't as interesting as whether you can sue someone for writing pure software with a patent. If you can get a patent on a linked list implemented in a computer system with a keyboard and monitor, then sue anyone who implements a linked list, that's just as bad as being able to patent a linked list.
The right to think, then express your thoughts in code, is out. Has been in the US for some time. Now in Europe too.
Someone wanted to sue my company once because I wrote a random number generator that involved multiple steps. They had a patent on "multistep random number generation". Turned out the RNG was mine, not my companies, and I wasn't selling it, so nothing happened.
Next question. I have a spiffy new free tool for covering the interactions n features in log(n) testcases instead of exp(n). So I scout groups.google.com for people complaining about the cost of testing feature interactions. When I spot them I post a reply, saying my tool exists, also pointing to several similar tools. Sometimes I just send the author an email. Am I spamming under this law?
I live in California. I got 1276 emails yesterday. One (1) was not spam. Wow, that would be $1,275,000 in penalties due to me alone, just yesterday!
That one, though, was from someone I've never heard of before, asking questions about things discussed on my website. Does that count as solicited or unsolicited?
No no no. You don't compress it with the letters rearranged randomly. You compress it with the intermediate letters rearranged in SORTED ORDER. Now go try it again.
There's this stuff called Public Key Cryptography. See, there's a public key, and a private key. They come in pairs. You keep the private key private, never give it out to anyone, not the government, nobody. The public key is public, you can make phone books out of it. The tricky bit is that a computer can use that private key to do some math that produces a result that shows that that private key really is paired with that public key. Only you, the only one who has the private key, can demonstrate that you have a private key matching that public key.
So you could use a public key instead of a SSN, make it truly public, and identity thiefs could learn that public key but still not impersonate you.
There's still the matter of what happens if you forget your private key or it becomes public, but it's a lot harder for that to happen than what we have today where the thief just has to learn your SSN and date of birth.
Hard SF looks a little way into the future and it sees nanotechnology, AI way beyond human intelligence building its own communities in space, and humanity either wiped out by plagues or at best bypassed by the wave of progress.
From the human perspective, the future looks bleak. We've got 50 to 200 years until obsolescence.
From the AI perspective, well, how is the writer going to simulate someone 1000x smarter than themselves? Probably not well. And even if they succeed, how do you get the reader to relate?
The use I've seen for code generators is performance. They've all been in C, writing C code. One avoids a swarm of if-statements by generating 64 different routines for 64 different cases, then the caller calls one of them based on the current input. Another writes a perfect hash for a tokenizer based on the current set of tokens. Without generated code, we'd have to use slower code instead.
The blurb mentions the camera will supply a Sony RAW format, with software to manipulate it. Will this format be the data straight from the four types of pixels, or will it be preprocessed into RGB?
The blurb also says a linear transformation is used to transform the four input colors into RGB. I believe you can get a lot of gain in accuracy from doing a nonlinear transformation, possibly more than you can get from having a fourth color. The camera's sensors will just have different sensitivities than human eyes, and a curve is likely to describe the mapping from camera sensor to eye better than a straight line. If Sony's RAW format gives you the original four colors, you could do the nonlinear transform yourself. Otherwise you can't.
There's a French saying. If you're young and not a liberal, you have no heart. If you're old and not a conservative, you have no brain. If the trend continues, a voting public of average age 200 would be scary.
For example, if the average voting age today was 200, that's born on average in 1803, there'd be this huge pro-slavery faction.
When I tried installing it a year ago, it couldn't use winmodems, so we couldn't get email with it. We reverted to Windows, since then Linux has just been a mysterious blob that sits on half our hard drive.
I think the real reason for declining tape and CD sales is cell phones. It used to be that driving in the car, or walking from here to there, there wasn't much you could do but listen to music. Now you can talk to people on the cell phone instead. So, out with radios, CD players, and walkmans, and the tapes and CDs that go with them.
Before I learned calculus, I asked my 7th grade teacher what the area of a sphere was. "You're in 7th grade, you don't need to know that", they said. OK, that was a substitute 7th grade math teacher.
By 8th grade I knew it was 4 pi r^2, but didn't know why. I could see you could derive it from the volume, 4/3 pi r^3, by dividing the volume up into a bunch of pyramids, each of volume 1/3 base*r. But I still had no way to derive one or the other from more basic principles. I asked my 8th grade math teacher. "Bob, there are some questions that every man must answer for himself", he said. (He was actually a good teacher, by the way.)
As near as I can tell, what is offensive in price discrimination, is when I get charged more if they know who I am than if they don't know who I am. It's also a offensive if I can pay less total if I purchase a little more, but that's just a matter of ill-tuned formulas.
Getting charged more to preserve my privacy is OK. Getting charged more because of how the product will be used is OK. Getting charged more for a brand name is even OK. Getting charged more per item for purchasing small amounts is OK.
What am I missing? It would be offensive if a net worth of $1000000 meant less purchasing power than a net worth of $1000. What else? What's the simplest definition of what's bad about price differentiation?
It's really really hard to remain anonymous when you have lots of joinable databases. We have lots of joinable databases, and there are more all the time. Outlawing joining of databases to preserve privacy strikes me as a lost cause.
However, if the only goal is to add more public information to databases, namely which databases are being joined and why, that's a good thing. Especially if it can be automated.
My impression is that when companies outsource to India, they don't outsource to individual developers in India. They form a whole group in India, complete with a manager in India. Usually many groups with several layers of management. They have offices, they come into work every day, they have meetings.
Outsourcing teams or departments is something different than individuals telecommuting.
Slashdot Mirror
Time to plug my free pairwise testing tool jenny (it jennyrates tests). 10 on/off dimensions with three constraints (-w1a2b3a -w1a3b5a -w1a5b9a) took 8 or 9 testcases, depending on the seed to the random number generator.
Pairwise testing takes fewer testcases than Taguchi's orthogonal arrays. Orthogonal arrays guarantee every that every dimension setting appears an equal number of times (does it guarantee that for pairs of dimension settings too?) N-way testing just requires that every combination of n settings is covered at least once. Detecting bugs only requires n-way testing, where n is 2 or 3.
I wouldn't be surprised if it's something in their genes that turns off evolution. That there are genes that control how fast mutations can appear. Some species have these genes turned all the way up. Perhaps others have it turned all the way down. What's the mutation rate of blue-green algae?
I just bought a TI-36X II from Costco. And promptly lost the manual. Most of the functions I haven't been able to figure out how to get at despite a fair amount of effort. And those that I can get at are very cumbersome. For example, decimal to hex is: enter number, hit "enter", hit "2nd", hit 8 (which means hex), read the answer, clear, hit "2nd", hit 7 (which means dec), enter next number ... . The keys are small and close together too and don't give me a reassuring click to let me know I've pressed them. Even turning it off is tricky enough for me to usually need to attempt it more than once.
I recommend steering clear of the TI-36X II no matter what you're looking for in a calculator.
I also have a 17-year-old TI-55 III, and remain quite happy with it. I can't imagine why TI went in for such a bad user interface as this TI-36X II.
Woah, spoke too soon. Serves me right for believing Slashdot. The business world is reporting this as an outrageous defeat of business by open source advocates: http://biz.yahoo.com/djus/030924/1303000668_2.html
Whether you can issue a patent on a pure software invention isn't as interesting as whether you can sue someone for writing pure software with a patent. If you can get a patent on a linked list implemented in a computer system with a keyboard and monitor, then sue anyone who implements a linked list, that's just as bad as being able to patent a linked list.
The right to think, then express your thoughts in code, is out. Has been in the US for some time. Now in Europe too.
Someone wanted to sue my company once because I wrote a random number generator that involved multiple steps. They had a patent on "multistep random number generation". Turned out the RNG was mine, not my companies, and I wasn't selling it, so nothing happened.
Not. They're safe.
Next question. I have a spiffy new free tool for covering the interactions n features in log(n) testcases instead of exp(n). So I scout groups.google.com for people complaining about the cost of testing feature interactions. When I spot them I post a reply, saying my tool exists, also pointing to several similar tools. Sometimes I just send the author an email. Am I spamming under this law?
I live in California. I got 1276 emails yesterday. One (1) was not spam. Wow, that would be $1,275,000 in penalties due to me alone, just yesterday!
That one, though, was from someone I've never heard of before, asking questions about things discussed on my website. Does that count as solicited or unsolicited?
No no no. You don't compress it with the letters rearranged randomly. You compress it with the intermediate letters rearranged in SORTED ORDER. Now go try it again.
There's this stuff called Public Key Cryptography. See, there's a public key, and a private key. They come in pairs. You keep the private key private, never give it out to anyone, not the government, nobody. The public key is public, you can make phone books out of it. The tricky bit is that a computer can use that private key to do some math that produces a result that shows that that private key really is paired with that public key. Only you, the only one who has the private key, can demonstrate that you have a private key matching that public key.
So you could use a public key instead of a SSN, make it truly public, and identity thiefs could learn that public key but still not impersonate you.
There's still the matter of what happens if you forget your private key or it becomes public, but it's a lot harder for that to happen than what we have today where the thief just has to learn your SSN and date of birth.
Hard SF looks a little way into the future and it sees nanotechnology, AI way beyond human intelligence building its own communities in space, and humanity either wiped out by plagues or at best bypassed by the wave of progress.
From the human perspective, the future looks bleak. We've got 50 to 200 years until obsolescence.
From the AI perspective, well, how is the writer going to simulate someone 1000x smarter than themselves? Probably not well. And even if they succeed, how do you get the reader to relate?
The use I've seen for code generators is performance. They've all been in C, writing C code. One avoids a swarm of if-statements by generating 64 different routines for 64 different cases, then the caller calls one of them based on the current input. Another writes a perfect hash for a tokenizer based on the current set of tokens. Without generated code, we'd have to use slower code instead.
Hum, I always thought the point of diamonds was how hard they are, not how expensive they are.
But then, I'm not the target market.
Horror! Horror! What's next? Slashdot tracking people's behaviour and sticking them with Karma rating?
My initial reaction was, cool, where can I query this thing to see how they rate my value to Usenet.
The blurb mentions the camera will supply a Sony RAW format, with software to manipulate it. Will this format be the data straight from the four types of pixels, or will it be preprocessed into RGB?
The blurb also says a linear transformation is used to transform the four input colors into RGB. I believe you can get a lot of gain in accuracy from doing a nonlinear transformation, possibly more than you can get from having a fourth color. The camera's sensors will just have different sensitivities than human eyes, and a curve is likely to describe the mapping from camera sensor to eye better than a straight line. If Sony's RAW format gives you the original four colors, you could do the nonlinear transform yourself. Otherwise you can't.
There's a French saying. If you're young and not a liberal, you have no heart. If you're old and not a conservative, you have no brain. If the trend continues, a voting public of average age 200 would be scary.
For example, if the average voting age today was 200, that's born on average in 1803, there'd be this huge pro-slavery faction.
I saw something similar from the BBC when they linked to one of my pages last September.
When I tried installing it a year ago, it couldn't use winmodems, so we couldn't get email with it. We reverted to Windows, since then Linux has just been a mysterious blob that sits on half our hard drive.
I think the real reason for declining tape and CD sales is cell phones. It used to be that driving in the car, or walking from here to there, there wasn't much you could do but listen to music. Now you can talk to people on the cell phone instead. So, out with radios, CD players, and walkmans, and the tapes and CDs that go with them.
Before I learned calculus, I asked my 7th grade teacher what the area of a sphere was. "You're in 7th grade, you don't need to know that", they said. OK, that was a substitute 7th grade math teacher.
By 8th grade I knew it was 4 pi r^2, but didn't know why. I could see you could derive it from the volume, 4/3 pi r^3, by dividing the volume up into a bunch of pyramids, each of volume 1/3 base*r. But I still had no way to derive one or the other from more basic principles. I asked my 8th grade math teacher. "Bob, there are some questions that every man must answer for himself", he said. (He was actually a good teacher, by the way.)
As near as I can tell, what is offensive in price discrimination, is when I get charged more if they know who I am than if they don't know who I am. It's also a offensive if I can pay less total if I purchase a little more, but that's just a matter of ill-tuned formulas.
Getting charged more to preserve my privacy is OK. Getting charged more because of how the product will be used is OK. Getting charged more for a brand name is even OK. Getting charged more per item for purchasing small amounts is OK.
What am I missing? It would be offensive if a net worth of $1000000 meant less purchasing power than a net worth of $1000. What else? What's the simplest definition of what's bad about price differentiation?
It's really really hard to remain anonymous when you have lots of joinable databases. We have lots of joinable databases, and there are more all the time. Outlawing joining of databases to preserve privacy strikes me as a lost cause.
However, if the only goal is to add more public information to databases, namely which databases are being joined and why, that's a good thing. Especially if it can be automated.
My impression is that when companies outsource to India, they don't outsource to individual developers in India. They form a whole group in India, complete with a manager in India. Usually many groups with several layers of management. They have offices, they come into work every day, they have meetings.
Outsourcing teams or departments is something different than individuals telecommuting.
Come nanotech and the colonization of the solar system, that won't be nearly enough.
Very good! I would have fallen for it if /. didn't print out the URLs for links.
What's an environment that does the same thing, but for C? I can see that macros could foul up the works unless the system is extensible.