A court sided with Google on the "fair use" question mostly because Google's scanning process (a) was transformative and (b) did not substantially affect the market for the original work. Google provided a way to search within books -- which was not a capability offered before -- and when Google shows the context from the original work, it does not show all the pages of the book. Instead, it cuts chunks out so that readers have a reason to get the book through an authorized channel. The decision did not depend on whether Google is a for-profit or non-profit enterprise, because copyright law does not inquire about that.
In this case, the Internet Archive doesn't have either of those copyright-relevant factors on its side.
The AC who submitted the story also distorts what TFA said "leaves critical legal issues unresolved": It is not the fact that SFWA is raising a hue and cry 10 years after the Internet Archive launched this effort, but rather the fact that much of what the Internet Archive does goes below the radar of content creators in general.
Fortunately for you, the FTC isn't the only party that can prosecute violations of anti-trust or fair-competition laws. State and local prosecutors and even private parties can, too.
That is conclusory and unconvincing. If net neutrality's best argument is "nuh uh, net neutrality means you have the right to be spoken to, not to speak", no wonder the FCC canned it. Are you so new to the Internet that you do not remember the way people accessed the Internet before corporations tried to lock people into their walled gardens? People who access the Internet have a right to speak, but the net neutrality crowd doesn't like that for some reason.
"[W]hether the FCC can govern intrastate commerce" is not a very accurate description of the question before the Supreme Court, or that court's decision. That case was an extremely narrow ruling on whether two particular sections of the federal law establishing the FCC gave the FCC authority to preempt state rules on depreciation schedules for equipment where both the FCC and the state had jurisdiction over setting telecom rates.
Contrast that to the rulings in Wickard and is progeny, through Gonzales v. Raich (2005), where federal law can govern even intrastate activities as long as the local effects are part of an overarching scheme of national regulation.
Hosting companies were not the only ones to block the racist morons: registrars also did, which is what OP was complaining about. Exactly where does net neutrality say that ISPs can't discriminate, but hosting companies and domain registrars are free to discriminate, and why? Because "net neutrality" is largely being pushed by hosting companies and other people who want to force ISPs to carry their content?
"Company X blocks or throttles network traffic between X's users and X's competitor Y" is a straightforward violation of existing, well-founded laws about fair competition. The FCC's net neutrality rules did not, and cannot, change those laws or make the inapplicable.
No, it's really not clear. Maybe that is the implication that Microsoft wants people to draw, but they don't directly say who provided the documentation in question.
It is bizarre how often "non-linear" is used in this thread, and for how many reasons. Is your model based on fantasies? It doesn't matter, the model is non-linear, so inaccurate corrections don't matter! Pump more energy into the system, and non-linearity makes it behave more chaotically! (Hint: Non-linear and chaotic are different concepts.) Non-linearity makes people get confused about short-term high-frequency responses versus long-term low-frequency responses, and then picks their pockets!
An awful lot of Puerto Rico's hurricane costs are due to "deferred" maintenance of public facilities. The territorial government was too busy diverting funds as political favors to keep the infrastructure up-to-date. They apparently even forgot how to keep things in good repair. As a result, things broke badly when they finally got a strong storm.
ARM is a RISC architecture, and plenty of RISC architectures suffer from Spectre. Meltdown is an Intel-only bug -- AMD doesn't have it because they implemented an obvious security rule, and presumably Cyrix and other x86 implementations didn't either.
Of course it correlates with the other white noise. All signals in the same domain correlate with each other. The real question is how strongly they correlate with each other, and over what duration.
The DMCA does not require this. It requires services like YouTube to implement a takedown process with particular criteria. Google's demonetization and reassignment of ad revenue are its own creations, unmoored from the law's requirements.
You just don't see innovations like the "COME FROM" statement in modern languages. Now there's a language feature with a noteworthy impact on defect density.
ISO 7185 ("Standard Pascal") defines conformant arrays to be an optional feature, which gives us another -- entirely redundant -- reason to dislike K. S. Kyosuke, because K. S. Kyosuke either lies or is willing to make definitive, but wrong, claims out of ignorance.
Well, they are using data from GitHub, so even though some of the projects go back 18 years, the projects should all be fairly well-maintained. However, the study's authors give no good reason to think that the kind of linear regression they use is appropriate. Even assuming a linear model is appropriate, their choice of variables seems arbitrary. For example, why is "log of project age" the right parameter to use, rather than some quantized version of age, or the square root of the age? Did they choose those by exploratory analysis (p-value fishing)?
So you are saying that the years since BWK wrote that article have given us even more reasons to dislike Pascal, such as the fact that the only versions that are useful are either dead for 20+ years (Turbo Pascal) or need vendor-proprietary extensions (Delphi)?
That was the joke. If you were less of a humorless scold, maybe you would have realized that the bit about co-worker complaints was a flag to indicate where the costs of that tradeoff come in.
One occasionally does have use for write-only source code, where the source code is small enough that changing the purpose of the program would inherently require rewriting such a large proportion of the code that it is just as efficient to start from a blank file, but that is seldom going to be the case for even a 500-line program.
No, that is a shitty tradeoff. If easily spotting a bug when someone tells you where to look is the most common situation, you are writing too much obviously bad code, and you should look at every line you write as a possible bug and write more (or better) unit tests.
In the group I work with, if you look at code that gets committed to shared repositories, the most common Python errors involve runtime detection of errors on code paths that only run in uncommon situations: in particular, use of undefined function or variable names, and putting a string-typed value where a number-typed value is expected (or vice versa). Statically typed languages avoid both of these problems, and loosely typed languages mostly avoid the second problem. Type confusion errors are often hard to trace their origin because the wrong-typed value can be passed through a long chain of function calls and storage.
Particularly annoying is a functional-style programming habit where my coworkers use tuples to hold strongly structured data, apparently mostly to avoid defining a class. They seem to think it is overall cheaper to write foo(1) than to write foo.job_id.
No one else has pointed out the old saw that there are three kinds of lies (lies, damned lies, and statistics)? Or that one could -- and someone actually did -- literally write a book on "How to Lie with Statistics"?
That is very much like saying that because Microsoft Visual Studio lets you write in C++.NET, there has been a lot of unfair badmouthing of C++. Pointing to one vendor's proprietary extensions (which essentially make a new language) doesn't make the general complaint wrong.
My god, you are a genius! The solution to America's woes is to turn every city into San Jose, California! Why hasn't anyone thought of this before? </sarcasm>
Slashdot Mirror
A court sided with Google on the "fair use" question mostly because Google's scanning process (a) was transformative and (b) did not substantially affect the market for the original work. Google provided a way to search within books -- which was not a capability offered before -- and when Google shows the context from the original work, it does not show all the pages of the book. Instead, it cuts chunks out so that readers have a reason to get the book through an authorized channel. The decision did not depend on whether Google is a for-profit or non-profit enterprise, because copyright law does not inquire about that.
In this case, the Internet Archive doesn't have either of those copyright-relevant factors on its side.
The AC who submitted the story also distorts what TFA said "leaves critical legal issues unresolved": It is not the fact that SFWA is raising a hue and cry 10 years after the Internet Archive launched this effort, but rather the fact that much of what the Internet Archive does goes below the radar of content creators in general.
Fortunately for you, the FTC isn't the only party that can prosecute violations of anti-trust or fair-competition laws. State and local prosecutors and even private parties can, too.
That is conclusory and unconvincing. If net neutrality's best argument is "nuh uh, net neutrality means you have the right to be spoken to, not to speak", no wonder the FCC canned it. Are you so new to the Internet that you do not remember the way people accessed the Internet before corporations tried to lock people into their walled gardens? People who access the Internet have a right to speak, but the net neutrality crowd doesn't like that for some reason.
"[W]hether the FCC can govern intrastate commerce" is not a very accurate description of the question before the Supreme Court, or that court's decision. That case was an extremely narrow ruling on whether two particular sections of the federal law establishing the FCC gave the FCC authority to preempt state rules on depreciation schedules for equipment where both the FCC and the state had jurisdiction over setting telecom rates.
Contrast that to the rulings in Wickard and is progeny, through Gonzales v. Raich (2005), where federal law can govern even intrastate activities as long as the local effects are part of an overarching scheme of national regulation.
Hosting companies were not the only ones to block the racist morons: registrars also did, which is what OP was complaining about. Exactly where does net neutrality say that ISPs can't discriminate, but hosting companies and domain registrars are free to discriminate, and why? Because "net neutrality" is largely being pushed by hosting companies and other people who want to force ISPs to carry their content?
"Company X blocks or throttles network traffic between X's users and X's competitor Y" is a straightforward violation of existing, well-founded laws about fair competition. The FCC's net neutrality rules did not, and cannot, change those laws or make the inapplicable.
Sex is a trait of people, gender is a trait of nouns or pronouns.
No, wait, those are the 20th century definitions.
Gender is what you feel you are, sex is what other people feel when they feel your body.
No, wait, that's offensive and patriarchic.
I give up. What is the difference between the two, as recently redefined by identity activists?
No, it's really not clear. Maybe that is the implication that Microsoft wants people to draw, but they don't directly say who provided the documentation in question.
It is bizarre how often "non-linear" is used in this thread, and for how many reasons. Is your model based on fantasies? It doesn't matter, the model is non-linear, so inaccurate corrections don't matter! Pump more energy into the system, and non-linearity makes it behave more chaotically! (Hint: Non-linear and chaotic are different concepts.) Non-linearity makes people get confused about short-term high-frequency responses versus long-term low-frequency responses, and then picks their pockets!
An awful lot of Puerto Rico's hurricane costs are due to "deferred" maintenance of public facilities. The territorial government was too busy diverting funds as political favors to keep the infrastructure up-to-date. They apparently even forgot how to keep things in good repair. As a result, things broke badly when they finally got a strong storm.
There is a very simple rule about this: If it's hot, that is climate. If it's cold, that is just weather.
ARM is a RISC architecture, and plenty of RISC architectures suffer from Spectre. Meltdown is an Intel-only bug -- AMD doesn't have it because they implemented an obvious security rule, and presumably Cyrix and other x86 implementations didn't either.
Of course it correlates with the other white noise. All signals in the same domain correlate with each other. The real question is how strongly they correlate with each other, and over what duration.
The DMCA does not require this. It requires services like YouTube to implement a takedown process with particular criteria. Google's demonetization and reassignment of ad revenue are its own creations, unmoored from the law's requirements.
So there is a hodgepodge of mutually incompatible dialects of Pascal that are available, some of which are useful?
This is not helping Pascal's case, and makes it look more and more like my C++.NET analogy was accurate.
You just don't see innovations like the "COME FROM" statement in modern languages. Now there's a language feature with a noteworthy impact on defect density.
ISO 7185 ("Standard Pascal") defines conformant arrays to be an optional feature, which gives us another -- entirely redundant -- reason to dislike K. S. Kyosuke, because K. S. Kyosuke either lies or is willing to make definitive, but wrong, claims out of ignorance.
Well, they are using data from GitHub, so even though some of the projects go back 18 years, the projects should all be fairly well-maintained. However, the study's authors give no good reason to think that the kind of linear regression they use is appropriate. Even assuming a linear model is appropriate, their choice of variables seems arbitrary. For example, why is "log of project age" the right parameter to use, rather than some quantized version of age, or the square root of the age? Did they choose those by exploratory analysis (p-value fishing)?
So you are saying that the years since BWK wrote that article have given us even more reasons to dislike Pascal, such as the fact that the only versions that are useful are either dead for 20+ years (Turbo Pascal) or need vendor-proprietary extensions (Delphi)?
That was the joke. If you were less of a humorless scold, maybe you would have realized that the bit about co-worker complaints was a flag to indicate where the costs of that tradeoff come in.
One occasionally does have use for write-only source code, where the source code is small enough that changing the purpose of the program would inherently require rewriting such a large proportion of the code that it is just as efficient to start from a blank file, but that is seldom going to be the case for even a 500-line program.
No, that is a shitty tradeoff. If easily spotting a bug when someone tells you where to look is the most common situation, you are writing too much obviously bad code, and you should look at every line you write as a possible bug and write more (or better) unit tests.
In the group I work with, if you look at code that gets committed to shared repositories, the most common Python errors involve runtime detection of errors on code paths that only run in uncommon situations: in particular, use of undefined function or variable names, and putting a string-typed value where a number-typed value is expected (or vice versa). Statically typed languages avoid both of these problems, and loosely typed languages mostly avoid the second problem. Type confusion errors are often hard to trace their origin because the wrong-typed value can be passed through a long chain of function calls and storage.
Particularly annoying is a functional-style programming habit where my coworkers use tuples to hold strongly structured data, apparently mostly to avoid defining a class. They seem to think it is overall cheaper to write foo(1) than to write foo.job_id.
No one else has pointed out the old saw that there are three kinds of lies (lies, damned lies, and statistics)? Or that one could -- and someone actually did -- literally write a book on "How to Lie with Statistics"?
And my 50 line Perl version is much more concise than both of those, but my co-workers keep complaining that it's actually modem line noise!
That is very much like saying that because Microsoft Visual Studio lets you write in C++.NET, there has been a lot of unfair badmouthing of C++. Pointing to one vendor's proprietary extensions (which essentially make a new language) doesn't make the general complaint wrong.
My god, you are a genius! The solution to America's woes is to turn every city into San Jose, California! Why hasn't anyone thought of this before? </sarcasm>