Each AMD module has two cores. Each core has a control unit (scheduler) 4 execution units (2ALU 2AGU) and an I/O unit to get data in and out. Each core is therefore capable of eight independent single-precision operation or four independent double-precision operations per cycle. With two each of ALU and AGU, arguably that's TWO cores and AMD should call it a 16-core CPU rather than an 8-core.
Several years ago, Intel's CPU cores were what most would call 75% of a core, sharing more parts between "cores" than AMDs' did, so Intel tried to redefine "core" to mean just those parts. Now the situation has reversed a bit - AMD cores can either have separate, dedicated DP FPUs, or two can share it as a DP FPU. Intel's are separate FPUs, so now Intel changed their mind and wants to redefine "core" again.
What can be objectively said is that at the moment, Intel's cores share fewer resources. It used to be the other way around, and it probably will be again.
There are two different legal theories used to justify this. There is the third-party approach, and the business records approach. Around half of Slashdot commenters actively support the business records excuse, whether or not they realize it.
Government passes law saying that carriers must _____ (provide good service in the sticks, hire enough black people, whatever), liberals cheer. Goverment inspects carrier's call/personnel records to ensure that the _______ (carrier has good service in BFE/hired enough black people, whatever). Carrier gets tired of monthly audits, goes to court. Court rules that govt can inspect the carrier's records any time they want to, liberals cheer. Govt inspects YOUR call records and YOUR personnel file. Liberals get mad at carrier, demand law that carriers must _____. Rinse and repeat.
It's actually a tough issue because in some ways you DO want the government to be able to look at Sprint's business records; on the other hand Sprint's records are records about you, their customer.
A well known quote called Sturgeon's Revelation or Sturgeon's law is "90% of everything is crap". It's certainly true of the software code I've seen, and of the (small) sample of accounting work I've had reason to examine.
In addition, two other factors are probably are work.
If I intend to use an app for something I do often, I frequently click to download the top two or three, trying out each one as the next one downloads. If I'm going to use it often, I may as well select the one I like best. This is more true on Android than iOS, because iOS has fewer free apps. I'm unlikely to BUY three apps in order to compare them.
On the other hand, if I download an app for something I do NOT do frequently, I may well use the app for the task at hand and be done with it. It's not that I didn't use it again because it sucked. Maybe I only used it once because I only need to build one set of stairs, or fix one ipad, or whatever. It may have worked perfectly well, so the job is done and I don't need it anymore.
> So you either support terrorism posts on Twitter and Facebook (ISIL et al.) or you don't.
I very strongly support terrorists posting on Twitter.
A long time ago, I was a licensed private investigator and I did a bit of bounty hunting - finding people who asked a bail bondsman to out their bail, then skipped out leaving the bondsman holding the bag. Once in a while, some idiot fugitive would post a picture or location on Twitter or Facebook; "Hangin with my homies at Kirk's Bar". An hour later, the idiot would be in jail. I would very much like for terrorists to be as stupid.
That's interesting that your software takes 100 times as long as the Linux kernel does. My first thought was "I bet that could be reduced by 90%", but you said you already did that, reducing it from 10 hours on 20 build servers. I'm curious why it takes so much longer, what the difference is. Does this project have a lot more tightly coupled dependencies than the kernel does?
AMD has x86 processors with 16 cores. As I recall you can have up to 4 CPUs per motherboard, so 64 cores total. Whether that's appropriate for your "desktop" is your decision. Their APUs have 4 CPU and 8 GPU cores.
Fry's and Microcenter are reasonable choices for brick-and-mortar retailers.
256 is th minimum allowed value to be POSIX compliant. Think real hard I bet you can guess how that value ended up there. I don't think there's ever been a *nix below 1024.
It's great that Windows is getting some of the capabilities that *nix had in the 1970s and 1980s. Let's see, they now have full support for storage over 2TB with Windows 7 service pack 2; only 20 years behind Linux. Before that, they were only 20 years behind on 2GB storage. And now they're only 30-40 years behind on long paths.
Based on this schedule, in another 10 years, they'll move beyond simplistic "any program run by Administrator can do anything, to anything" and get grown-up security, mandatory access control like SELinux and grsecurity have provided for a decade on Linux.
For every problem, there is a solution that's simple, easy, and wrong.
> Traffic can be managed at the end points.
Netflix IS an endpoint of Comcast's network. There is a router at which Comcast's network ends. Netlfix wants to plug in to Comcast's network and dump billions of dollars of traffic. Comcast wants to do as you suggest and manage that endpoint.
> definition "To circulate, distribute, or print information for the public at large."
Let's try it and see. This directive says that if it's published, it must be made available to the public. Let's try substituting in that definition of "publish":
Any report circulated, distributed, or printed for the public at large must be made available to the public.
I'm not sure that makes sense, seems redundant. If that's what they mean, they should say that. It's certainly not clearly obvious that they mean "anything made available to public at large must be made available to the public at large". If so, I suggest they add "anything printed on paper must be printed on paper".
> Plnaes get moved around from route to route routinely (sorry)
You don't move a plane or crew certified for 75 minutes between airports (ETOPS-75) to a route with 180 minutes between airports (ETOPS-180). The systems in place do a pretty damn good job of preventing that.
This is where someone brings up the August, 2015 LA-Honolulu flight by American Airlines. Yes, once a plane was wrongly assigned. The Airbus was a model that was certified for the flight, but without the extra oxygen and fire extinguishers provided by the ETOPS-180 option package. It it made national news. Once, a plane flew a route it wasn't supposed to and it was national news because that occurs so rarely. American Airlines immediately reported it to the FAA as required.
(It didn't crash or anything, just flew a route that required more fire extinguishers becuase there are no diversion airports between LA and Honolulu).
I see your point about "scan". I'm not sure how a judge and jury might interpret the old wording, but adding the definition for "scan" does seem move the needle in Facebook's favor.
Speaking of defining words, now I shall try to remember a new word I learned, a discussion such as this is called "dialectic". https://en.m.wikipedia.org/wik...
> I feel a bit conflicted on this one.... But on the other hand Windows 10 has some stuff in it I really like.
Suppose for a moment that Windows 10 was awesome, as good as sex. And Microsoft is forcing it upon people who don't want it. How do you feel about forcing sex on someone who doesn't it? Still conflicted?
In my case, I have expensive hardware which is controlled by a Windows application, an application which doesn't run in Windows 10. Without Windows 7 or earlier, I have to throw out several thousand dollars worth of equipment.
You have a point, there is a new bit. The additional wording stating that not only are photographs not included under the statute, but neither are analyses of photographs, is new. Not necessarily a huge change - the old wording could have been interpreted the same way, but it is a change. Good point.
Your position that the categorization of "photographs" was changed is based on the premise that last week "photograph" meant on analog photos printed on paper? That's an interesting theory. Perhaps you're thinking that since this law was passed in 2008, the word "photograph" should be read with the meaning it would have had in 2008? In 2008, over 98% of cameras sold were digital. Kodak stopped manufacturing consumer film at all a few months later, in 2009. So certainly when the law was originally passed in 2008 nearly all photographs were digital photographs.
Sometimes it's temping to pay more attention to what we WISH the law to be than what the law actually says. This law says, and always said, that photos are not included in its scope.
To cite one example from my own experience, I'm not sure how giving away an open source mapping script which works with whatever image-processing library happens to be installed is "cheating the GPL", but okay.
> But the biggest one is: how will your users get that library? If give them both to your customer
Probably the most obvious example is pretty much every C program ever written, which is maybe half of all software ever written. Every C program uses the standard library. Approximately none of them include libc in the installer, or any script to get it.
Another example would be the approximately 60,000 packages available from any major Linux repository. When you create an RPM or a.deb package, you simply declare which other packages it depends on.
> because you are still using all the code from the library
COPYright essentially controls the right to make copies, not the right to use. If I buy a book, I can't legally make and distribute copies of it; I CAN use it to prop up a wobbly chair, swat a bug, etc. Generally I can USE it without worrying about copyright. It's the making of copies that invokes copyright. Static linking makes a copies.
Further, my distribution of a dynamically linked executable doesn't actually even USE the library. The library is used by the recipient when they RUN the program. Presumably they have the right to use the library.
Again, I'm not saying that no court will ever, under any circumstances, hold any kind of dynamic linking to invoke copyrights. Maybe they will. Maybe. I'm saying that making copies by static linking DEFINITELY invokes copyrights. Dynamic, since it doesn't create a copy, is much less likely to involve COPYrights.
Thanks for your thoughtful post. After reading it, I double-checked the amendment to see if you saw something I missed. Central to your point, I think, is your statement:
> excluding photographs of any kind from the definition of "biometric information"
I don't see where the amendment makes ANY change to the definition of "biometric information". The lines are helpfully numbered, perhaps you can point out where you see that?
I do see where three words were inserted into the definition of "biometric identifier". It originally said "photographs" were excluded, the update says "analog or digital photographs". So it ALREADY excluded photographs, that was NOT changed.
The original author of the law says that when he originally wrote iris scans, fingerprints... excluding photographs, he meant that digital photographs were excluded along with analog ones . Importantly, he put language in his amendment saying that it was clarifying, not changing the current law. I suppose that if you can prove that he originallly meant only analog, printed photos, you can call him a liar. I believe him when he says that by "photographs" he meant digital photographs too.
He got in because the password was left as "password". In what programming language is "password" a secure password?
Having said that, ten years or fifteen ago PHP had serious security issues, given that it is designed to be used on web, where the application will be attacked daily. It was literally impossible to write a secure program in PHP; literally "hello world" had a security vulnerability. Much has changed. PHP was originally a CMS, written in Perl with a bit of C. It's now an actual programming language, one used by clueless little companies like Facebook. Seriously, it has improved a lot. The world's largest web sites wouldn't be running on PHP if it were junk.
Having said THAT, it's still an "easy" language to start learning. You can start writing little PHP scripts without being trained and educated as a programmer. If you do that in any language and put your scripts on the web, you'll get hacked. While PHP as a language is pretty decent now, PHP "scripters" who don't know any programming language other than PHP are still mostly people who don't know much. But the same is true of.Net or many other languages. If you learned a bit of a language but never learned programming and especially security issues of web programming, you probably shouldn't expose your software to internet hackers.
Well a statically compiled binary, which contains a copy of the library, is obviously a derived work, no question about it.
Judge Alsup ruled that using an API dynamically (and even re-implementing it) isn't, because the library's API can't be copyright protected anyway. The appeals court disagreed in this case. So it's debatable so using an API via dynamic linking. It may come down to the specific facts of the case, or the mood of the judge or jury.
On to fair use. A key consideration in fair use is how much of the original work is used. If you copy the whole thing verbatim, it's generally unlikely to be fair use. If you use 1% of the original, it's much more likely to be fair. Static linking copies the entire function verbatim. Dynamic copies only the header line. Not much different than copying the full text of a book versus copying only the chapter titles.
With dynamic linking, there's an argument to be made either way. Different courts may decide different cases of dynamic linking differently. Static linking, verbatim copying of the entire implementation - that's not even arguable, not at all.
That's a very thoughtful post. Some clarification might be helpful regarding this sentence:
> In Illinois they decided the state has the right to restrict the freedom to use pictures
Here's the old text of the law:
Biometric identifiers do not include writing samples, written signatures, photographs
Here's the new text:
Biometric identifiers do not include writing samples, written signatures, physical or digital photographs
So they explicitly did NOT restrict the use of photographs. The law, from the time it was initially passed, regulated the storage and use of biometric identifiers, which it defined as "retina scans, fingerprints", etc. It never did cover photographs. The new text clarifies that digitial photos are photos, and therefore not restricted by this law.
I think by "worse" GP means "more clearly brings the result under the license used by the library". Static linking creates a strong case that the resulting executable must abide by the license of the library.
The reason why is because copy right is essentially about the right to create copies. Static linking EMBEDS a -copy- of the library into the executable. Therefore you're clearly making and distributing copies, and must follow the GPL is whichever license applies.
With dynamic linking, you're not copying the full compiled code of the library before the program runs. You're "only" copying the external interface of the library, much like Google copied the external interface of Java. Therefore it MAY be fair use and you MAY not have to follow the license of the library.
I capitalize MAY because this case did not rule that all use of any and all APIs, in any way, is always fair use. The jury found that in this particular case, the way that Android used the Java APIs was fair, after considering statutory criteria such as how it affected the market for Oracle's Java products.
The summary is complete bull. Here is the primary change bring made to the law. It used to say:
Biometric identifiers do not include writing samples, written signatures, photographs
They are trying to update it to say:
Biometric identifiers do not include writing samples, written signatures, physical or digital photographs
In other words, they are clarifying that yes, a digital photo is a photo. The amendment also has wording stating that this is clarifying, not changing the law - that digital photographs were photographs last week too.
Here's the full text of the amendment. Underlined words are the words being added, words crossed out are being deleted.
Slashdot Mirror
Each AMD module has two cores. Each core has a control unit (scheduler) 4 execution units (2ALU 2AGU) and an I/O unit to get data in and out. Each core is therefore capable of eight independent single-precision operation or four independent double-precision operations per cycle. With two each of ALU and AGU, arguably that's TWO cores and AMD should call it a 16-core CPU rather than an 8-core.
Several years ago, Intel's CPU cores were what most would call 75% of a core, sharing more parts between "cores" than AMDs' did, so Intel tried to redefine "core" to mean just those parts. Now the situation has reversed a bit - AMD cores can either have separate, dedicated DP FPUs, or two can share it as a DP FPU. Intel's are separate FPUs, so now Intel changed their mind and wants to redefine "core" again.
What can be objectively said is that at the moment, Intel's cores share fewer resources. It used to be the other way around, and it probably will be again.
There are two different legal theories used to justify this. There is the third-party approach, and the business records approach. Around half of Slashdot commenters actively support the business records excuse, whether or not they realize it.
Government passes law saying that carriers must _____ (provide good service in the sticks, hire enough black people, whatever), liberals cheer.
Goverment inspects carrier's call/personnel records to ensure that the _______ (carrier has good service in BFE/hired enough black people, whatever).
Carrier gets tired of monthly audits, goes to court.
Court rules that govt can inspect the carrier's records any time they want to, liberals cheer.
Govt inspects YOUR call records and YOUR personnel file.
Liberals get mad at carrier, demand law that carriers must _____.
Rinse and repeat.
It's actually a tough issue because in some ways you DO want the government to be able to look at Sprint's business records; on the other hand Sprint's records are records about you, their customer.
A well known quote called Sturgeon's Revelation or Sturgeon's law is "90% of everything is crap". It's certainly true of the software code I've seen, and of the (small) sample of accounting work I've had reason to examine.
In addition, two other factors are probably are work.
If I intend to use an app for something I do often, I frequently click to download the top two or three, trying out each one as the next one downloads. If I'm going to use it often, I may as well select the one I like best. This is more true on Android than iOS, because iOS has fewer free apps. I'm unlikely to BUY three apps in order to compare them.
On the other hand, if I download an app for something I do NOT do frequently, I may well use the app for the task at hand and be done with it. It's not that I didn't use it again because it sucked. Maybe I only used it once because I only need to build one set of stairs, or fix one ipad, or whatever. It may have worked perfectly well, so the job is done and I don't need it anymore.
> So you either support terrorism posts on Twitter and Facebook (ISIL et al.) or you don't.
I very strongly support terrorists posting on Twitter.
A long time ago, I was a licensed private investigator and I did a bit of bounty hunting - finding people who asked a bail bondsman to out their bail, then skipped out leaving the bondsman holding the bag. Once in a while, some idiot fugitive would post a picture or location on Twitter or Facebook; "Hangin with my homies at Kirk's Bar". An hour later, the idiot would be in jail. I would very much like for terrorists to be as stupid.
That's interesting that your software takes 100 times as long as the Linux kernel does. My first thought was "I bet that could be reduced by 90%", but you said you already did that, reducing it from 10 hours on 20 build servers. I'm curious why it takes so much longer, what the difference is. Does this project have a lot more tightly coupled dependencies than the kernel does?
AMD has x86 processors with 16 cores. As I recall you can have up to 4 CPUs per motherboard, so 64 cores total. Whether that's appropriate for your "desktop" is your decision. Their APUs have 4 CPU and 8 GPU cores.
Fry's and Microcenter are reasonable choices for brick-and-mortar retailers.
PATH_MAX is the length of path gauranteed to work on this OS (and can be different for each OS).
POSIX_ PATH_MAX is the lowest value of PATH_MAX allowed on other POSIX systems, which could matter for network communication etc. It's 256.
256 is th minimum allowed value to be POSIX compliant. Think real hard I bet you can guess how that value ended up there. I don't think there's ever been a *nix below 1024.
It's great that Windows is getting some of the capabilities that *nix had in the 1970s and 1980s. Let's see, they now have full support for storage over 2TB with Windows 7 service pack 2; only 20 years behind Linux. Before that, they were only 20 years behind on 2GB storage. And now they're only 30-40 years behind on long paths.
Based on this schedule, in another 10 years, they'll move beyond simplistic "any program run by Administrator can do anything, to anything" and get grown-up security, mandatory access control like SELinux and grsecurity have provided for a decade on Linux.
For every problem, there is a solution that's simple, easy, and wrong.
> Traffic can be managed at the end points.
Netflix IS an endpoint of Comcast's network. There is a router at which Comcast's network ends. Netlfix wants to plug in to Comcast's network and dump billions of dollars of traffic. Comcast wants to do as you suggest and manage that endpoint.
> definition "To circulate, distribute, or print information for the public at large."
Let's try it and see. This directive says that if it's published, it must be made available to the public. Let's try substituting in that definition of "publish":
Any report circulated, distributed, or printed for the public at large must be made available to the public.
I'm not sure that makes sense, seems redundant. If that's what they mean, they should say that. It's certainly not clearly obvious that they mean "anything made available to public at large must be made available to the public at large". If so, I suggest they add "anything printed on paper must be printed on paper".
> Plnaes get moved around from route to route routinely (sorry)
You don't move a plane or crew certified for 75 minutes between airports (ETOPS-75) to a route with 180 minutes between airports (ETOPS-180). The systems in place do a pretty damn good job of preventing that.
This is where someone brings up the August, 2015 LA-Honolulu flight by American Airlines. Yes, once a plane was wrongly assigned. The Airbus was a model that was certified for the flight, but without the extra oxygen and fire extinguishers provided by the ETOPS-180 option package. It it made national news. Once, a plane flew a route it wasn't supposed to and it was national news because that occurs so rarely. American Airlines immediately reported it to the FAA as required.
(It didn't crash or anything, just flew a route that required more fire extinguishers becuase there are no diversion airports between LA and Honolulu).
I see your point about "scan". I'm not sure how a judge and jury might interpret the old wording, but adding the definition for "scan" does seem move the needle in Facebook's favor.
Speaking of defining words, now I shall try to remember a new word I learned, a discussion such as this is called "dialectic".
https://en.m.wikipedia.org/wik...
> I feel a bit conflicted on this one. ... But on the other hand Windows 10 has some stuff in it I really like.
Suppose for a moment that Windows 10 was awesome, as good as sex. And Microsoft is forcing it upon people who don't want it. How do you feel about forcing sex on someone who doesn't it? Still conflicted?
In my case, I have expensive hardware which is controlled by a Windows application, an application which doesn't run in Windows 10. Without Windows 7 or earlier, I have to throw out several thousand dollars worth of equipment.
You have a point, there is a new bit. The additional wording stating that not only are photographs not included under the statute, but neither are analyses of photographs, is new. Not necessarily a huge change - the old wording could have been interpreted the same way, but it is a change. Good point.
Your position that the categorization of "photographs" was changed is based on the premise that last week "photograph" meant on analog photos printed on paper? That's an interesting theory. Perhaps you're thinking that since this law was passed in 2008, the word "photograph" should be read with the meaning it would have had in 2008? In 2008, over 98% of cameras sold were digital. Kodak stopped manufacturing consumer film at all a few months later, in 2009. So certainly when the law was originally passed in 2008 nearly all photographs were digital photographs.
Sometimes it's temping to pay more attention to what we WISH the law to be than what the law actually says. This law says, and always said, that photos are not included in its scope.
To cite one example from my own experience, I'm not sure how giving away an open source mapping script which works with whatever image-processing library happens to be installed is "cheating the GPL", but okay.
> But the biggest one is: how will your users get that library? If give them both to your customer
Probably the most obvious example is pretty much every C program ever written, which is maybe half of all software ever written. Every C program uses the standard library. Approximately none of them include libc in the installer, or any script to get it.
Another example would be the approximately 60,000 packages available from any major Linux repository. When you create an RPM or a .deb package, you simply declare which other packages it depends on.
I'm guessing you used an older version. Moodle too has improved dramatically in the last four years. It has really grown up.
>> Ten or fifteen years ago PHP sucked
> I'm maintaining/refactoring a large legacy PHP
I feel your pain. I've done the same with a million-line PHP project called Moodle.
Since you are refactoring, I hope you study modern PHP and apply it where it makes sense.
> because you are still using all the code from the library
COPYright essentially controls the right to make copies, not the right to use. If I buy a book, I can't legally make and distribute copies of it; I CAN use it to prop up a wobbly chair, swat a bug, etc. Generally I can USE it without worrying about copyright. It's the making of copies that invokes copyright. Static linking makes a copies.
Further, my distribution of a dynamically linked executable doesn't actually even USE the library. The library is used by the recipient when they RUN the program. Presumably they have the right to use the library.
Again, I'm not saying that no court will ever, under any circumstances, hold any kind of dynamic linking to invoke copyrights. Maybe they will. Maybe. I'm saying that making copies by static linking DEFINITELY invokes copyrights. Dynamic, since it doesn't create a copy, is much less likely to involve COPYrights.
Thanks for your thoughtful post. After reading it, I double-checked the amendment to see if you saw something I missed. Central to your point, I think, is your statement:
> excluding photographs of any kind from the definition of "biometric information"
I don't see where the amendment makes ANY change to the definition of "biometric information". The lines are helpfully numbered, perhaps you can point out where you see that?
I do see where three words were inserted into the definition of "biometric identifier". It originally said "photographs" were excluded, the update says "analog or digital photographs". So it ALREADY excluded photographs, that was NOT changed.
The original author of the law says that when he originally wrote iris scans, fingerprints ... excluding photographs, he meant that digital photographs were excluded along with analog ones . Importantly, he put language in his amendment saying that it was clarifying, not changing the current law. I suppose that if you can prove that he originallly meant only analog, printed photos, you can call him a liar. I believe him when he says that by "photographs" he meant digital photographs too.
He got in because the password was left as "password". In what programming language is "password" a secure password?
Having said that, ten years or fifteen ago PHP had serious security issues, given that it is designed to be used on web, where the application will be attacked daily. It was literally impossible to write a secure program in PHP; literally "hello world" had a security vulnerability. Much has changed. PHP was originally a CMS, written in Perl with a bit of C. It's now an actual programming language, one used by clueless little companies like Facebook. Seriously, it has improved a lot. The world's largest web sites wouldn't be running on PHP if it were junk.
Having said THAT, it's still an "easy" language to start learning. You can start writing little PHP scripts without being trained and educated as a programmer. If you do that in any language and put your scripts on the web, you'll get hacked. While PHP as a language is pretty decent now, PHP "scripters" who don't know any programming language other than PHP are still mostly people who don't know much. But the same is true of .Net or many other languages. If you learned a bit of a language but never learned programming and especially security issues of web programming, you probably shouldn't expose your software to internet hackers.
Well a statically compiled binary, which contains a copy of the library, is obviously a derived work, no question about it.
Judge Alsup ruled that using an API dynamically (and even re-implementing it) isn't, because the library's API can't be copyright protected anyway. The appeals court disagreed in this case. So it's debatable so using an API via dynamic linking. It may come down to the specific facts of the case, or the mood of the judge or jury.
On to fair use. A key consideration in fair use is how much of the original work is used. If you copy the whole thing verbatim, it's generally unlikely to be fair use. If you use 1% of the original, it's much more likely to be fair. Static linking copies the entire function verbatim. Dynamic copies only the header line. Not much different than copying the full text of a book versus copying only the chapter titles.
With dynamic linking, there's an argument to be made either way. Different courts may decide different cases of dynamic linking differently. Static linking, verbatim copying of the entire implementation - that's not even arguable, not at all.
That's a very thoughtful post. Some clarification might be helpful regarding this sentence:
> In Illinois they decided the state has the right to restrict the freedom to use pictures
Here's the old text of the law:
Biometric identifiers do not include writing samples, written signatures, photographs
Here's the new text:
Biometric identifiers do not include writing samples, written signatures, physical or digital photographs
So they explicitly did NOT restrict the use of photographs. The law, from the time it was initially passed, regulated the storage and use of biometric identifiers, which it defined as "retina scans, fingerprints", etc. It never did cover photographs. The new text clarifies that digitial photos are photos, and therefore not restricted by this law.
I think by "worse" GP means "more clearly brings the result under the license used by the library". Static linking creates a strong case that the resulting executable must abide by the license of the library.
The reason why is because copy right is essentially about the right to create copies. Static linking EMBEDS a -copy- of the library into the executable. Therefore you're clearly making and distributing copies, and must follow the GPL is whichever license applies.
With dynamic linking, you're not copying the full compiled code of the library before the program runs. You're "only" copying the external interface of the library, much like Google copied the external interface of Java. Therefore it MAY be fair use and you MAY not have to follow the license of the library.
I capitalize MAY because this case did not rule that all use of any and all APIs, in any way, is always fair use. The jury found that in this particular case, the way that Android used the Java APIs was fair, after considering statutory criteria such as how it affected the market for Oracle's Java products.
The summary is complete bull. Here is the primary change bring made to the law. It used to say:
Biometric identifiers do not include writing samples, written signatures, photographs
They are trying to update it to say:
Biometric identifiers do not include writing samples, written signatures, physical or digital photographs
In other words, they are clarifying that yes, a digital photo is a photo. The amendment also has wording stating that this is clarifying, not changing the law - that digital photographs were photographs last week too.
Here's the full text of the amendment. Underlined words are the words being added, words crossed out are being deleted.
http://www.ilga.gov/legislatio...