Not when the "anonymous reader" uses the word "effectively" to describe Schwartz's actions. That means he knows that the facts of the case are different than what he's describing, and he's trying to substitute his own interpretation for what's already asserted.
How many database servers did you have backing your app?
My whole point was that the engineering concerns of a large-scale app are hugely different than those of a small app. I believe that you could port an app from MySQL to PostgreSQL in a couple weeks, assuming that it was as simple as going through the code and adjusting the queries. For Facebook there's a lot more than that to do, though. There's cluster performance, there's handling replication lag, there's integrating it back with memcached, etc. Cluster engineering is its own discipline.
A lot of people are saying "they'll be fine if they used an ORM layer", but it's highly doubtful they did because the first performance overhead to eliminate is the ORM layer by either writing directly to the database or customizing the ORM layer so much that it's equivalent to writing directly to the DB.
A minor difference that exists in 4,000 instances and who knows how many places in the code that's also distributed across multiple servers, isn't minor, especially when there are hundreds or even thousands of minor differences.
And no, the differences in SQL between Oracle and MySQL aren't minor. It's not just syntax, and it's not MySQL-can-Oracle-can't. It's the performance characteristics of various queries, the logic of how they're implemented, and the incredible investment in configuring a large cluster to work smoothly (which MySQL and Oracle do extremely differently. Large scale systems add a layer of complexity all their own that's a totally separate engineering challenge.
Short version: Switching from MySQL to anything else would be the equivalent to a ground-up rewrite, though this is largely true of any database system. MySQL hasn't somehow uniquely trapped them here.
Not mistaken, but the issue the NoSQL people face when trying to replicate something like a Facebook-sized cluster of relational databases is that they have to build the ACID features back in, which tends to negate the performance advantage.
Facebook isn't really using a relational database system either: they have a gigantic memcached layer on top of a gigantic MySQL layer. They have, effectively, a massive in-memory database that's continually being written back to MySQL for permanence. That's the only way they could get sufficient read performance.
Right, good point, because we all know that wars come with a flat price, so Obama's 3 are obviously much more expensive than Bush's 2. 3 > 2, it's simple math!
You're a freak. You probably try to eat a turkey by putting one end in your mouth and pushing really hard.
Seriously, good on you for picking it up the right way at 12, but you're extremely unique in terms of people, their optimal learning strategies, and what they're likely to keep working at without payoff.
Did you know that Friendster has 125 million active users still?
It's a little nuts because it's dead in North America, but it's going strong in southeast Asia, and is a very viable business still. Frequently we tend to equate Internet popularity with business success, and the reverse, but neither are very strong correlations.
The difference being that in Debt of Honor, a virus stopped recording transactions at noon while trading continued until 5:00 p.m., so no one knew who owned what after the afternoon's trading. They didn't roll back so much as agree to ignore all the non-recorded trading.
What MtGox is proposing is to revert a bunch of known transactions, and a lot of bitcoin traders are rightly pointing out that they profited by the volatility and shouldn't lose their profits, especially since whoever scammed the exchange likely didn't leave their bitcoins in MtGox, so reverting does nothing to address the underlying "crime".
The only way to lower housing prices is to increase the amount of housing on the market. The foreign investors drive prices up, but it's the self-satisfied yuppies and seniors sitting in million dollar shacks who keep prices high by refusing to allow the city to build anything higher than three stories outside of downtown.
They're not housing projects. It's a certain amount of any development that is set aside for low-income or rent-controlled housing. You don't get granite countertops, but you do get the same neighbourhood and development complex as middle and upper class people. You mix with people with jobs and a decent life, and because the well-off live right next door, the police pay attention to calls and crime rates.
Nothing cures homelessness and poverty like access to housing, and apportioned social housing avoids the problems for which housing projects are famous.
I don't dispute the problems with poverty in Vancouver. I see it every day. My wife teaches at a high school in Whalley, Surrey. I go through Chinatown a couple days a week.
What I'm disputing is that the riot was caused by social unrest. The pictures tell an obvious tale: half the crowd is wearing expensive Canucks jerseys and have nice haircuts. The ones who've been identified from the photos are rich kids from the burbs. If you see sunglasses, they're expensive designer sunglasses. This was hooliganism, not the poor rising up.
Foreign investment in real estate is a mixed blessing. It drives up prices, preventing the poor and the young from owning property, but it also gluts the rental market driving rental prices down. A common complaint among investors is that they're unable to finance the purchase solely by renting it out. The only unambiguously good thing is that the demand for more condos allows the city to require an apportionment of new construction to go for social housing.
Vancouver housing prices weren't the cause of the riot. Neither were disaffected youth angry about the cost of living.
While the housing market is grossly overinflated, the rental market is sane. Young people simply rent instead of buying, and rent quite nice places too because the main driver of inflating housing costs are foreign investors buying up all the condo stock. Metro Vancouver's unemployment rate (7.6%) is lower than Canada's overall, and has been pretty constant for the last decade. There's no large, pent up reservoir of anger.
The cause of the riots was 1) corralling 100,000 fans downtown to watch the game on outdoor screens, and 2) a large portion of those fans being drunk suburban kids looking to get their riot on. Blame lackluster police presence if you want. It was hooliganism pure and simple. Look at the photos. Look at their expensive shoes. Those Canucks jerseys they're all wearing aren't cheap. They're young, middle-class drunks having fun.
Logic tells us that there's no way that the U.S. government would announce bin Laden's death without being 100% sure that he was dead, because it would be too easy for bin Laden to release a video reading the headlines from today's newspaper.
Maybe they know he died a year ago from kidney failure and the U.S. government wants to take credit, but you can bet that bin Laden is dead right now.
You get the reliability you select. All you need to do to avoid this issue with brtfs is 1) use ext4 instead, or 2) use a hard disk that doesn't lie about whether or not it's buffering.
Fedora is not production ready even in the sense of a daily use desktop with data that can't be recovered from backup, even without this problem. It's explicitly a testbed distro, and always has been.
Because it's Fedora, which is explicitly a testbed distro for newer technologies, and the problem is limited to disks that don't handle flush requests correctly. If you care about a 100% reliable system, you shouldn't be using Fedora in the first place. If you're willing to accept a very small risk of losing everything and having to restore from backup, you get to play with the shiny new toys.
Interesting stats. For smoking I was sure it was the other way.
Also, claiming 28% of ALL health care dollars goes to administrative costs makes me really skeptical. Unless you're counting stuff like doctor salaries as administrative
I drew my numbers from memory, but they're backed up here: http://www.ncbi.nlm.nih.gov/pubmed/12930930, and slightly worse: 31% to 16%. If I read this correctly, it does NOT include salaries for doctors, only for employees engaged in administrative roles:
For the United States and Canada, we calculated the administrative costs of health insurers, employers' health benefit programs, hospitals, practitioners' offices, nursing homes, and home care agencies in 1999. We analyzed published data, surveys of physicians, employment data, and detailed cost reports filed by hospitals, nursing homes, and home care agencies. In calculating the administrative share of health care spending, we excluded retail pharmacy sales and a few other categories for which data on administrative costs were unavailable. We used census surveys to explore trends over time in administrative employment in health care settings. Costs are reported in U.S. dollars.
You're correct that closing that gap wouldn't close the larger gap in overall health care spending. There are many areas of financial inefficiency in American health care. However, even if I grant that the difference in average life span is due entirely to non-medical factors, and assume that equalizing the non-medical factors would yield comparable life expectancies, you still have the fact that basically equivalent overall health care is vastly cheaper under a government run system.
Really, this is what perplexes me about American health care. In Medicare/Medicaid and the Veterans Administration, you effectively have UHC for certain segments of the population. While health care costs are growing for Medicare, they're growing faster in the private health care sector.
The per enrollee cost growth in Medicaid (6.1 percent) is lower than the per enrollee cost growth in comparable coverage under Medicare (6.9), private health insurance (10.6), and monthly premiums for employer-sponsored insurance (12.6).(source)
You have your own working models that demonstrate that UHC style programs are at least as effective as your overall system, and they're much cheaper.
Canada has a life expectancy 2.5 years greater than the U.S. We have more smokers per capita, five times the number of donut shops per capita, and a province representing a quarter of our population who thinks a good meal is a plate french fries covered in cheese curds and gravy. I don't think you can say that Canadian lifestyles are significantly different from American ones in a way that accounts for that big a gap.
And there's more to health care costs than the efficiency of health care.
Yes: There's the fact that as a portion of health care dollars spent, the U.S. spends almost twice as much as Canada on administrative costs (28% vs. 16%, last time I checked).
So explain why the first world countries with universal health care run by the government in some fashion all have life expectancies two years greater than Americans, and do it at 55% the cost.
You have to turn off your non-wifi ebook reader because when the flight attendants are getting things going, "turn off all portable electronic devices" is a lot easier and faster than learning to tell which are and which aren't, and checking that those passengers who are still poking a screen are using something non-wifi enabled.
And really, it's not that big a hardship to turn off your device for ten minutes during takeoff.
No, you can't just use MS's.NET. If you write C# for linux to run on Mono, you have to avoid the proprietary bits of the.NET platform and use swap-ins. This makes your code non-portable to MS's.NET, which lacks the swap-ins but includes the proprietary bits.
In other words, thanks to the proprietary stuff, a Mono platform simply cannot be a complete implementation of.NET. When Novell had cut a deal with MS, it could be because it could include the proprietary parts, but since Novell shed Mono, you're back to a separate and incompatible library for the virtual machine.
I cited a government official verifying that the certificate of live birth was real, was given by them, and accurately represented underlying paperwork. You said "you think we should believe everything the government or people in the government says at face value and never question it". I offered a data point; you claimed I made a categorical statement about the trustworthiness of government officials. That's a straw man.
The only difference is you supported one and likely didn't like or believe the other.
No, the difference is that one has multiple, independently verified pieces of evidence directly supporting it; the other had a lot of self-interested government officials saying "trust me".
There are obscured documents- check, the same with WMDs
The certificate of live birth was released generally in obscured form at one point, but members of the media were invited to see and photograph the document. Several did, including staffers from factcheck.org; Here's their photograph of it, unobscured, so go crazy verifying the serial number that was available back in 2008.
News papers and websites reported it as true
More than that, members of the media were able to visit the archives of the newspapers that published the birth announcement and view the archived newspapers with the birth announcement in them, proving pretty conclusively that the announcement (which was placed by the hospital as a matter of procedure) was made at the correct time, in the correct place, providing independent corroboration of the certificate of live birth.
Did I miss something?
That the serial number was available in 2008 from independent sources in the media, who found (i.e., saw and touched) corroborating evidence of his birth in the form of news announcements.
In the case of Iraqi WMDs, the only people claiming to have direct proof were the government officials, and they weren't sharing that proof; the news media was simply reporting on what they were told, not claiming to have direct evidence themselves.
the UN weapons inspection quarterly reports were saying it was likely that Iraq both still had WMDs and the capability to create them
The key word there being "likely", meaning the inspectors had no direct evidence of WMDs. By comparison, the certificate of live birth and the newspaper announcements were directly accessible to the media, who took pictures of them.
No, a conspiracy to hide the truth about Obama's birth isn't impossible at this stage (meaning prior to the release of the long form), but the reasonable belief is that Obama's status as a natural born citizen was demonstrated as well as could be expected in 2008.
the entire point of demanding an un-obfuscated document was to be able to independently verify it. It's just another piece of paper until it's verified. That will take some time as you pointed out earlier. Hell, it hasn't even been a month.
Actually, it's been three years that the serial number has been available--plenty of time to independently verify it.
But you do not even know what that stuff is and are blindly making those conclusions based on your faith.
Again with the straw man. I have heard the birther arguments, I have evaluated them and found them loony, and by comparison, Obama's certificate of live birth and the independent verification of it seems to demonstrate to me quite sufficiently that he was a natural born U.S. citizen. You, however, assume that I'm just an o-bot, when I've spent all this time without returning your sarcasm, presenting actual facts and verifiable evidence.
Slashdot Mirror
Not when the "anonymous reader" uses the word "effectively" to describe Schwartz's actions. That means he knows that the facts of the case are different than what he's describing, and he's trying to substitute his own interpretation for what's already asserted.
Well said.
Scientific Linux is specifically not binary compatible with RHEL. That's why they were able to move so fast getting 6.0 out.
How many database servers did you have backing your app?
My whole point was that the engineering concerns of a large-scale app are hugely different than those of a small app. I believe that you could port an app from MySQL to PostgreSQL in a couple weeks, assuming that it was as simple as going through the code and adjusting the queries. For Facebook there's a lot more than that to do, though. There's cluster performance, there's handling replication lag, there's integrating it back with memcached, etc. Cluster engineering is its own discipline.
A lot of people are saying "they'll be fine if they used an ORM layer", but it's highly doubtful they did because the first performance overhead to eliminate is the ORM layer by either writing directly to the database or customizing the ORM layer so much that it's equivalent to writing directly to the DB.
A minor difference that exists in 4,000 instances and who knows how many places in the code that's also distributed across multiple servers, isn't minor, especially when there are hundreds or even thousands of minor differences.
And no, the differences in SQL between Oracle and MySQL aren't minor. It's not just syntax, and it's not MySQL-can-Oracle-can't. It's the performance characteristics of various queries, the logic of how they're implemented, and the incredible investment in configuring a large cluster to work smoothly (which MySQL and Oracle do extremely differently. Large scale systems add a layer of complexity all their own that's a totally separate engineering challenge.
Short version: Switching from MySQL to anything else would be the equivalent to a ground-up rewrite, though this is largely true of any database system. MySQL hasn't somehow uniquely trapped them here.
Not mistaken, but the issue the NoSQL people face when trying to replicate something like a Facebook-sized cluster of relational databases is that they have to build the ACID features back in, which tends to negate the performance advantage.
Facebook isn't really using a relational database system either: they have a gigantic memcached layer on top of a gigantic MySQL layer. They have, effectively, a massive in-memory database that's continually being written back to MySQL for permanence. That's the only way they could get sufficient read performance.
Right, good point, because we all know that wars come with a flat price, so Obama's 3 are obviously much more expensive than Bush's 2. 3 > 2, it's simple math!
You're a freak. You probably try to eat a turkey by putting one end in your mouth and pushing really hard.
Seriously, good on you for picking it up the right way at 12, but you're extremely unique in terms of people, their optimal learning strategies, and what they're likely to keep working at without payoff.
Did you know that Friendster has 125 million active users still?
It's a little nuts because it's dead in North America, but it's going strong in southeast Asia, and is a very viable business still. Frequently we tend to equate Internet popularity with business success, and the reverse, but neither are very strong correlations.
The difference being that in Debt of Honor, a virus stopped recording transactions at noon while trading continued until 5:00 p.m., so no one knew who owned what after the afternoon's trading. They didn't roll back so much as agree to ignore all the non-recorded trading.
What MtGox is proposing is to revert a bunch of known transactions, and a lot of bitcoin traders are rightly pointing out that they profited by the volatility and shouldn't lose their profits, especially since whoever scammed the exchange likely didn't leave their bitcoins in MtGox, so reverting does nothing to address the underlying "crime".
The only way to lower housing prices is to increase the amount of housing on the market. The foreign investors drive prices up, but it's the self-satisfied yuppies and seniors sitting in million dollar shacks who keep prices high by refusing to allow the city to build anything higher than three stories outside of downtown.
They're not housing projects. It's a certain amount of any development that is set aside for low-income or rent-controlled housing. You don't get granite countertops, but you do get the same neighbourhood and development complex as middle and upper class people. You mix with people with jobs and a decent life, and because the well-off live right next door, the police pay attention to calls and crime rates.
Nothing cures homelessness and poverty like access to housing, and apportioned social housing avoids the problems for which housing projects are famous.
I don't dispute the problems with poverty in Vancouver. I see it every day. My wife teaches at a high school in Whalley, Surrey. I go through Chinatown a couple days a week.
What I'm disputing is that the riot was caused by social unrest. The pictures tell an obvious tale: half the crowd is wearing expensive Canucks jerseys and have nice haircuts. The ones who've been identified from the photos are rich kids from the burbs. If you see sunglasses, they're expensive designer sunglasses. This was hooliganism, not the poor rising up.
Foreign investment in real estate is a mixed blessing. It drives up prices, preventing the poor and the young from owning property, but it also gluts the rental market driving rental prices down. A common complaint among investors is that they're unable to finance the purchase solely by renting it out. The only unambiguously good thing is that the demand for more condos allows the city to require an apportionment of new construction to go for social housing.
Vancouver housing prices weren't the cause of the riot. Neither were disaffected youth angry about the cost of living.
While the housing market is grossly overinflated, the rental market is sane. Young people simply rent instead of buying, and rent quite nice places too because the main driver of inflating housing costs are foreign investors buying up all the condo stock. Metro Vancouver's unemployment rate (7.6%) is lower than Canada's overall, and has been pretty constant for the last decade. There's no large, pent up reservoir of anger.
The cause of the riots was 1) corralling 100,000 fans downtown to watch the game on outdoor screens, and 2) a large portion of those fans being drunk suburban kids looking to get their riot on. Blame lackluster police presence if you want. It was hooliganism pure and simple. Look at the photos. Look at their expensive shoes. Those Canucks jerseys they're all wearing aren't cheap. They're young, middle-class drunks having fun.
Logic tells us that there's no way that the U.S. government would announce bin Laden's death without being 100% sure that he was dead, because it would be too easy for bin Laden to release a video reading the headlines from today's newspaper.
Maybe they know he died a year ago from kidney failure and the U.S. government wants to take credit, but you can bet that bin Laden is dead right now.
You get the reliability you select. All you need to do to avoid this issue with brtfs is 1) use ext4 instead, or 2) use a hard disk that doesn't lie about whether or not it's buffering.
Fedora is not production ready even in the sense of a daily use desktop with data that can't be recovered from backup, even without this problem. It's explicitly a testbed distro, and always has been.
Because it's Fedora, which is explicitly a testbed distro for newer technologies, and the problem is limited to disks that don't handle flush requests correctly. If you care about a 100% reliable system, you shouldn't be using Fedora in the first place. If you're willing to accept a very small risk of losing everything and having to restore from backup, you get to play with the shiny new toys.
We try... Lord knows, we try... but when you can't tell parody from reality, that's its own special kind of gaffe.
Interesting stats. For smoking I was sure it was the other way.
Also, claiming 28% of ALL health care dollars goes to administrative costs makes me really skeptical. Unless you're counting stuff like doctor salaries as administrative
I drew my numbers from memory, but they're backed up here: http://www.ncbi.nlm.nih.gov/pubmed/12930930, and slightly worse: 31% to 16%. If I read this correctly, it does NOT include salaries for doctors, only for employees engaged in administrative roles:
You're correct that closing that gap wouldn't close the larger gap in overall health care spending. There are many areas of financial inefficiency in American health care. However, even if I grant that the difference in average life span is due entirely to non-medical factors, and assume that equalizing the non-medical factors would yield comparable life expectancies, you still have the fact that basically equivalent overall health care is vastly cheaper under a government run system.
Really, this is what perplexes me about American health care. In Medicare/Medicaid and the Veterans Administration, you effectively have UHC for certain segments of the population. While health care costs are growing for Medicare, they're growing faster in the private health care sector.
You have your own working models that demonstrate that UHC style programs are at least as effective as your overall system, and they're much cheaper.
Canada has a life expectancy 2.5 years greater than the U.S. We have more smokers per capita, five times the number of donut shops per capita, and a province representing a quarter of our population who thinks a good meal is a plate french fries covered in cheese curds and gravy. I don't think you can say that Canadian lifestyles are significantly different from American ones in a way that accounts for that big a gap.
And there's more to health care costs than the efficiency of health care.
Yes: There's the fact that as a portion of health care dollars spent, the U.S. spends almost twice as much as Canada on administrative costs (28% vs. 16%, last time I checked).
So explain why the first world countries with universal health care run by the government in some fashion all have life expectancies two years greater than Americans, and do it at 55% the cost.
You have to turn off your non-wifi ebook reader because when the flight attendants are getting things going, "turn off all portable electronic devices" is a lot easier and faster than learning to tell which are and which aren't, and checking that those passengers who are still poking a screen are using something non-wifi enabled.
And really, it's not that big a hardship to turn off your device for ten minutes during takeoff.
You're correct that there are ways to handle this situation, but having to do so rather voids the point about it being cross-platform, doesn't it?
No, you can't just use MS's .NET. If you write C# for linux to run on Mono, you have to avoid the proprietary bits of the .NET platform and use swap-ins. This makes your code non-portable to MS's .NET, which lacks the swap-ins but includes the proprietary bits.
In other words, thanks to the proprietary stuff, a Mono platform simply cannot be a complete implementation of .NET. When Novell had cut a deal with MS, it could be because it could include the proprietary parts, but since Novell shed Mono, you're back to a separate and incompatible library for the virtual machine.
I cited a government official verifying that the certificate of live birth was real, was given by them, and accurately represented underlying paperwork. You said "you think we should believe everything the government or people in the government says at face value and never question it". I offered a data point; you claimed I made a categorical statement about the trustworthiness of government officials. That's a straw man.
No, the difference is that one has multiple, independently verified pieces of evidence directly supporting it; the other had a lot of self-interested government officials saying "trust me".
The certificate of live birth was released generally in obscured form at one point, but members of the media were invited to see and photograph the document. Several did, including staffers from factcheck.org; Here's their photograph of it, unobscured, so go crazy verifying the serial number that was available back in 2008.
More than that, members of the media were able to visit the archives of the newspapers that published the birth announcement and view the archived newspapers with the birth announcement in them, proving pretty conclusively that the announcement (which was placed by the hospital as a matter of procedure) was made at the correct time, in the correct place, providing independent corroboration of the certificate of live birth.
That the serial number was available in 2008 from independent sources in the media, who found (i.e., saw and touched) corroborating evidence of his birth in the form of news announcements.
In the case of Iraqi WMDs, the only people claiming to have direct proof were the government officials, and they weren't sharing that proof; the news media was simply reporting on what they were told, not claiming to have direct evidence themselves.
The key word there being "likely", meaning the inspectors had no direct evidence of WMDs. By comparison, the certificate of live birth and the newspaper announcements were directly accessible to the media, who took pictures of them.
No, a conspiracy to hide the truth about Obama's birth isn't impossible at this stage (meaning prior to the release of the long form), but the reasonable belief is that Obama's status as a natural born citizen was demonstrated as well as could be expected in 2008.
Actually, it's been three years that the serial number has been available--plenty of time to independently verify it.
Again with the straw man. I have heard the birther arguments, I have evaluated them and found them loony, and by comparison, Obama's certificate of live birth and the independent verification of it seems to demonstrate to me quite sufficiently that he was a natural born U.S. citizen. You, however, assume that I'm just an o-bot, when I've spent all this time without returning your sarcasm, presenting actual facts and verifiable evidence.