Keep using what whatever tech you want to use and stop obsessing over what you "should" be using. If a compelling reason to start using something new develops then...start using something new. This isn't rocket science.
I just googled "browser market share" and took the first link, which happened to be this one. It lists FF at 20.3%, Chrome at 16.35% and Safari at 5.38%. That seems to correspond to the "NetApplications" row on the wiki page, which, admittedly, has the least favorable estimate for Chrome.
I wasn't the one who brought up Moore's law; that was the poster to whom I was responding, who claimed an application of Moore's law to current supercomputer performance would lead us to expect ExaFLOP performance in 2028 as opposed to 2020.
If you look at his presentation, he mentions that the current top performer is 17.6 PFLOPs. The list is from November 2012. 1000 / 17.6 is a factor of 56.8. Moore's law predicts a 2x performance improvement every 1.5 years. 56.8 log 2 = 5.83. 5.83 = 1.5 years = 8.7 years. November 2012 + 8.7 years = mid-2021.
Often this is because state legislatures exercise tight control on "tuition" but not "fees".
btw, It's not deceptive to give "tuition and fees" figures (sans room & board) as long as they're clear about what the number means. Most of them are. Also, the original poster to whom I was replying (ButchDeLoria, who introduced the $20k figure) specifically said "excluding supplies, housing and additional fees". My strong suspicion is that he's mistaken.
This is approximately what you'd expect given the following assumptions:
1. All else being equal, there's a positive correlation between general intelligence and ability to get into Harvard.
2. All else being equal, there's a positive correlation between general intelligence and household income.
3. Intelligence is, at least to some degree, heritable.
In other words, if your parents are wealthy then they're more likely to be "smart" and if your parents are "smart" you're correspondingly more likely to be "smart" yourself, meaning you have a leg up on the race to get into an elite school.
We might also add the following:
1. Having parents with a certain style of parenting (supportive, stable, stressing educational achievement, etc.) correlates with the sort of academic achievement that gets one into Harvard.
2. One is predisposed to use the same type of parenting style used by one's parents did and to stress the same things one's parents stressed. If your parents were abusive or negligent then you're more likely to be an abusive or negligent parent. If your parents stressed education then you're more likely to stress education as a parent. If your parents read to you as a child then you're more likely to read to your own children. Etc.
Are you including room and board in that $20,000 figure? Because the original poster claimed his nephew was paying $20k/year excluding "supplies, housing and additional fees". For Oklahoma State I used this estimator to calculate tuition and fees for 30 credit hours (i.e. two semesters) of 100% engineering courses, which carry the highest per-credit-hour fee. Result: $9,229.
Much of it goes to need-based financial aid. You raise the stick price so that those who can afford to pay more do, then turn around and take some portion of that extra money and use it to defray the costs of those who can't. See this piece from NPR on sticker price vs. actual price paid. Inflation-adjusted net price has gone up by about 30% for public schools and 22% for private schools over the past 15 years. That's obviously a growth rate that exceeds inflation, but it's a good sight less than the growth in sticker price over that same period.
assuming virtue is not correlated with parental income
Who said they're admitting kids based on "virtue"? They're admitting them based on expected future performance. In fact, I suspect they've had to lower standards for the sub-65k crowd just to juice the number to 20%. I agree with you, btw, that "virtue" (e.g. kindness, honesty, etc.) isn't correlated with household income.
I'm not saying more wealthy families don't have advantages. They do. But Harvard is somehow managing to fill 20% of its class w/ kids whose families fall under the $65k/year threshold. So some of these families, at least, are doing "what it takes" to get into Harvard.
They (elite schools) seem to be locating and successfully recruiting the lion's share of low-income high-ability students. At least if this article is to be believed:
Low-income high-achieving students at these schools have close to 100 percent odds of attending an Ivy League school or other highly selective college...
"These schools" are "from 15 large metropolitan areas. These areas often have highly regarded public high schools, such as in New York City or in the Washington, D.C., area." It's the 30% of low-income high-ability students outside those metro areas that aren't heading to elite universities. Harvard also claims that 20% of its class falls under the $65k/year threshold and therefore pays nothing.
For instance, if your parents make less than $65k/year (approx. 150% median U.S. household income, or 300% the cutoff for "poverty level") you can attend Harvard for free. Assuming you can get in. Which, in the grand scheme of things, sort of makes it a "merit based" scholarship after all.
Time went from ~25s to ~3s when I used eight threads. That's when the files weren't already cached on the client. They rarely change, so once they're cached it's just a matter of calling the server to (most of the time) get a 304 response. Even when the files aren't downloaded the client pays the latency penalty to check for a possible change.. We're on a high-latency link, so the round-trip could be something like 400ms. Do thirty of those in serial and you're up to 12 seconds.
Okay. How about Simon Phipps who covers OSS at InfoWorld. He seems to have made around 1300 posts since March 2007. Google +"Simon Phipps" +site:slashdot.org and you get 8,230 hits. Google +"Steven Vaughan-Nichols" +site:slashdot.org and you get 101,000 hits.
It's not part of my job description either, per se, but it is occasionally warranted. We have an application that needs to download ~20 files at launch, pulling them over a fairly high-latency network. None of them depend on the others. Apparently it never occurred to anyone who had worked on this application before that, hey, maybe we could transfer these things in parallel instead of in serial?
I could be wrong, but I suspect the OP was referencing "anything involving more than just a single thread" and not the more specific definition of concurrent programming.
I work with a developer who is 10 years my senior, but still doesn't understand how to write concurrent code
Concurrent code isn't new. If this guy doesn't understand it then his problem isn't that he has neglected to stay current, but that he was never very skilled to begin with.
There original article certainly seems to imply that the latter thing ("launching their pro-techie immigration push") runs counter to their formerly stated goal of boosting C.S. education in the U.S.
Eh. I'm getting compensated quite well by a company that wouldn't exist w/o immigrants. Ditto for my last job. I'm quite happy to see more competent tech workers enter this country and put down roots. It's not a zero sum game. Or, if it is, the competing parties aren't "me and the Indian guy who moves here" but "The U.S. and India". I'd rather the U.S. benefit from that guy's talents than India. If nothing else, the taxes he pays while working here will subsidize the education of American children.
Slashdot Mirror
Keep using what whatever tech you want to use and stop obsessing over what you "should" be using. If a compelling reason to start using something new develops then...start using something new. This isn't rocket science.
I just googled "browser market share" and took the first link, which happened to be this one. It lists FF at 20.3%, Chrome at 16.35% and Safari at 5.38%. That seems to correspond to the "NetApplications" row on the wiki page, which, admittedly, has the least favorable estimate for Chrome.
I wasn't the one who brought up Moore's law; that was the poster to whom I was responding, who claimed an application of Moore's law to current supercomputer performance would lead us to expect ExaFLOP performance in 2028 as opposed to 2020.
You might consider starting, since FF's 20% market share is approximately equal to the combined share of Safari and Chrome.
If you look at his presentation, he mentions that the current top performer is 17.6 PFLOPs. The list is from November 2012. 1000 / 17.6 is a factor of 56.8. Moore's law predicts a 2x performance improvement every 1.5 years. 56.8 log 2 = 5.83. 5.83 = 1.5 years = 8.7 years. November 2012 + 8.7 years = mid-2021.
Often this is because state legislatures exercise tight control on "tuition" but not "fees". btw, It's not deceptive to give "tuition and fees" figures (sans room & board) as long as they're clear about what the number means. Most of them are. Also, the original poster to whom I was replying (ButchDeLoria, who introduced the $20k figure) specifically said "excluding supplies, housing and additional fees". My strong suspicion is that he's mistaken.
This is approximately what you'd expect given the following assumptions:
1. All else being equal, there's a positive correlation between general intelligence and ability to get into Harvard.
2. All else being equal, there's a positive correlation between general intelligence and household income.
3. Intelligence is, at least to some degree, heritable.
In other words, if your parents are wealthy then they're more likely to be "smart" and if your parents are "smart" you're correspondingly more likely to be "smart" yourself, meaning you have a leg up on the race to get into an elite school.
We might also add the following:
1. Having parents with a certain style of parenting (supportive, stable, stressing educational achievement, etc.) correlates with the sort of academic achievement that gets one into Harvard.
2. One is predisposed to use the same type of parenting style used by one's parents did and to stress the same things one's parents stressed. If your parents were abusive or negligent then you're more likely to be an abusive or negligent parent. If your parents stressed education then you're more likely to stress education as a parent. If your parents read to you as a child then you're more likely to read to your own children. Etc.
Are you including room and board in that $20,000 figure? Because the original poster claimed his nephew was paying $20k/year excluding "supplies, housing and additional fees". For Oklahoma State I used this estimator to calculate tuition and fees for 30 credit hours (i.e. two semesters) of 100% engineering courses, which carry the highest per-credit-hour fee. Result: $9,229.
Much of it goes to need-based financial aid. You raise the stick price so that those who can afford to pay more do, then turn around and take some portion of that extra money and use it to defray the costs of those who can't. See this piece from NPR on sticker price vs. actual price paid. Inflation-adjusted net price has gone up by about 30% for public schools and 22% for private schools over the past 15 years. That's obviously a growth rate that exceeds inflation, but it's a good sight less than the growth in sticker price over that same period.
Guy to whom I was responding specifically said: "$20,000 a year, supplies, housing, and additional fees not included.
Who said they're admitting kids based on "virtue"? They're admitting them based on expected future performance. In fact, I suspect they've had to lower standards for the sub-65k crowd just to juice the number to 20%. I agree with you, btw, that "virtue" (e.g. kindness, honesty, etc.) isn't correlated with household income.
Which state? And what percentage of their students pay full sticker price? (Hint: probably only the wealthy ones.)
For fun, here's a list of top public universities and their in-state costs (from US News):
1. UC-Berkeley, $11,767
2. UCLA, $12,692
3. UVA, $12,006
4. Michigan, $13,437
5. UNC, $7,694
6. Wm. and Mary, $13,570
7. Georgia Tech, $10,098
8. UC-Davis, $13,877
9. UC-San Diego, $12,128
10. UC-Santa Barbara, $13,671
11. Wisconsin, $10,384
12. UC-Irvine, $14,090
13. Penn State, $16,444
14. Illinois, $14,428
15. UT-Austin, $9,792
16. Washington, $10,574
17. Florida, $5,656
18. Ohio State, $10,037
19. Maryland, $8,908
20. Pitt, $16,590
So which state's two major state universities are both $20k+?
Perhaps only marginally socialist. My alma mater (large, well regarded state school) gets on the order of 13% of its budget from the state.
I'm not saying more wealthy families don't have advantages. They do. But Harvard is somehow managing to fill 20% of its class w/ kids whose families fall under the $65k/year threshold. So some of these families, at least, are doing "what it takes" to get into Harvard.
"These schools" are "from 15 large metropolitan areas. These areas often have highly regarded public high schools, such as in New York City or in the Washington, D.C., area." It's the 30% of low-income high-ability students outside those metro areas that aren't heading to elite universities. Harvard also claims that 20% of its class falls under the $65k/year threshold and therefore pays nothing.
For instance, if your parents make less than $65k/year (approx. 150% median U.S. household income, or 300% the cutoff for "poverty level") you can attend Harvard for free. Assuming you can get in. Which, in the grand scheme of things, sort of makes it a "merit based" scholarship after all.
I wish it were latter. But, sadly, the former.
Time went from ~25s to ~3s when I used eight threads. That's when the files weren't already cached on the client. They rarely change, so once they're cached it's just a matter of calling the server to (most of the time) get a 304 response. Even when the files aren't downloaded the client pays the latency penalty to check for a possible change.. We're on a high-latency link, so the round-trip could be something like 400ms. Do thirty of those in serial and you're up to 12 seconds.
Maybe the definition of "concurrent programming" is somewhat fluid? All I can say is that I used a java.util.concurrent.ThreadPooExecutor.
Okay. How about Simon Phipps who covers OSS at InfoWorld. He seems to have made around 1300 posts since March 2007. Google +"Simon Phipps" +site:slashdot.org and you get 8,230 hits. Google +"Steven Vaughan-Nichols" +site:slashdot.org and you get 101,000 hits.
It's not part of my job description either, per se, but it is occasionally warranted. We have an application that needs to download ~20 files at launch, pulling them over a fairly high-latency network. None of them depend on the others. Apparently it never occurred to anyone who had worked on this application before that, hey, maybe we could transfer these things in parallel instead of in serial?
I could be wrong, but I suspect the OP was referencing "anything involving more than just a single thread" and not the more specific definition of concurrent programming.
I work with a developer who is 10 years my senior, but still doesn't understand how to write concurrent code
Concurrent code isn't new. If this guy doesn't understand it then his problem isn't that he has neglected to stay current, but that he was never very skilled to begin with.
There original article certainly seems to imply that the latter thing ("launching their pro-techie immigration push") runs counter to their formerly stated goal of boosting C.S. education in the U.S.
Eh. I'm getting compensated quite well by a company that wouldn't exist w/o immigrants. Ditto for my last job. I'm quite happy to see more competent tech workers enter this country and put down roots. It's not a zero sum game. Or, if it is, the competing parties aren't "me and the Indian guy who moves here" but "The U.S. and India". I'd rather the U.S. benefit from that guy's talents than India. If nothing else, the taxes he pays while working here will subsidize the education of American children.