According to Wikipedia, Noonan vs. Staples is not a valid precedent. The validity of the Massachusetts law was untested, as both parties assumed it was valid and neither challenged it. Any relevant further case could challenge the Constitutional validity of the law.
(Could be argued being part of the the API and Oracle America, Inc. v. Google, Inc. [wikipedia.org] shows that using the same API isn't infringing.)
As I understand it, not being a lawyer, the ruling was that APIs can be copyrighted, since they are creative works in fixed form.
However, copyrights can't prevent you from doing something other than very specific actions, and so if you use an API to write a program or library you have to use that API, it isn't infringing.
The question after that was whether Google's use of the Java API was legitimate. Oracle was arguing that it wasn't to create a Java program or library, since Android programs were not designed to interoperate with standard Java programs, and therefore Google was using it only because it was a well-known API.
What the idiots backing Perens here don't realize is the GPL is about to get an important clause struck out.
I don't remember a severability clause in GPLv2, so the court has the choice between holding the license to be legally valid or legally invalid, nothing in between. If ruled as invalid, then GRSecurity had no license to distribute in the first place.
The GPL points out that you don't have to accept the GPL, but in that case you don't have a license, and may not copy, change, or redistribute the software.
Courts generally interpret contracts as narrowly as possible.
However, they tend not to declare the contracts invalid. They tend to interpret the clauses in a restrictive way. If the court absolutely disagreed with Perens' interpretation, the court could rule that "further restrictions" does not clearly apply to what GRsecurity is doing, and hence GRsecurity is in the clear on their actions.
Of course, GRsecurity would have to prove not only that Perens was wrong, but that he had no good reason to think he was right, to win their lawsuit.
You can agree not to do something you have a right to do, even under penalty. Look at any NDA which curtails your free speech rights.
Sure. However, the GPL appears to forbid such agreements relating to GPLed software. The contract isn't automatically illegal (like a contract giving an employer rights to stuff an employee does on his or her own time and resources in this state), but it appears to me to violate the GPL. After all, I can generally offer any legal contract, but I can be in an agreement where I can't offer a specific one.
However, I disagree with court opinions saying GPL is a contract; it's only a license since it doesn't require you to do anything and there is no consideration on the software author's side.
Your disagreement with the court is immaterial here, and doesn't actually change anything. If GRsecurity violated its license, then it doesn't have a license, and all the copies made and distributed are unlicensed, and that's good for a substantial amount of statutory damages (courtesy of the MAFIAA).
GRSecurity cannot grant a license to the Linux kernel if GRSecurity doesn't have a valid license. If they have violated the GPL, then they don't have a valid license. These licenses aren't free-floating; legally, they have to be granted. (The question of what you do when you violate the GPL and lose your license can get rather involved, and GPLv3 provided for automatic reinstatement of the license under certain conditions - however, Linux is GPLv2 only, and is not distributed under GPLv3.)
GRsecurity uses a separate Stable Patch Access Agreement [perens.com] with the supposed restriction, and that agreement is between GRsecurity and the individual customer, not the customer and any other recipient. That agreement also explicitly says that "The User has all rights and obligations granted by grsecurity's software license, version 2 of the GNU GPL," so the user would not be doing the same thing.
That sounds an awful lot like adding terms to the GPL, which is not permitted by the GPL.
Therefore, I'm claiming the following things about reality. GRsecurity may be violating GPLv2 (I'm not taking a definite position on that). If so, GRsecurity doesn't have a valid license for the Linux kernel, and is forbidden to change or further copy it. If GRsecurity doesn't have a valid license, GRsecurity can't grant a license, and therefore their customers are running unlicensed copies of software. We know from the MAFIAA and lawsuits that illegitimate copies of copyrighted works can cost a whole lot of money. The Linux kernel does not operate under a copyright-assignment principle, so there's a large number of people with copyrighted code in the kernel, and I believe any of them could sue.
Hence, it looks legally risky to me to rely on a kernel supplied from GRsecurity.
2. and 3. are arguably factual, but any evaluation of risk is open to subjective interpretation. Nevertheless if Grsecurity proves their product presents NO risk of contributory infringement or breach of contract, these statements would be unequivocally false.
However, if the statements are false, but Bruce has reason to believe they're true, it's still not libel. US law is stringent about what constitutes libel, and it can be hard to prove it.
If it needed a lowercase conversion, it was the retroactively designated Model I, not the Model II which was a large box with 8" disk drives. You may be thinking of Level 2 BASIC, the Microsoft stuff.
Radio Shack offered the 8-bit character expansion after some time, which involved soldering a memory chip piggy-back and replacing the character generator.
There was a special CP/M version for the TRS-80 Model III, but the Model 4 could run standard CP/M software and that's what I got.
Visicalc was offered for the TRS-80 (I had a copy), but it came out first for the Apple, and got identified as Apple software. Accountants were spending their own money on Apple IIs and bringing them into work, and although that had to happen with some of the other computer types I never heard about it. It gave Apple a boost, and some business cred, but wasn't decisive.
What happened was the IBM PC, which soon took over the industry. The old Apple II and TRS-80 computers stayed in production for some time, but couldn't sell in nearly as great numbers. The PET had long gone by the wayside.
The three of them tried making their mark in different ways. Radio Shack tried coming out with a line of IBM PC-like computers that were superior, but this was before IBM-compatible ROMs, so they died in the marketplace, and Radio Shack continued making PC clones for a while. Commodore came out with the 64 and later the Amiga, which were innovative, very good, and didn't last. (IIRC, Commodore wasn't as committed to the industry.) Apple fumbled with the Apple III, produced the Lisa that was too expensive for home users, and finally the Macintosh, which was successful as the only real competitor to the PC and PC clones in the marketplace.
Another thing about $A and $B was that there was that you could type the strings in and print them out, and do absolutely nothing additional with them. I thought the most interesting thing about the TRS-80 Level 1 BASIC were the error messages: WHAT? (syntax error), HOW? (runtime error) and SORRY (ran out of memory). There were (IIRC) 26 16-bit integer variables, the two strings (limited length, maybe 16 characters each), and one array A[] or A() that took up the rest of memory.
The TRS-80, initially, had no way of making sound other than the tape relay, which was not a good musical device had had the potential of wearing out the relay. (I no longer have my technical manual, so I can't tell you where the relay was or what it was actually supposed to do).
The really fun thing about the 300 baud acoustic modem was that, when the phone wasn't plugged in but it was on, I could whistle to it and it would whistle back.
The Model II was Radio Shack's attempt to get into business programming, with 64K RAM, a large screen, 8" disk drives, and an impressive price tag. The Model III was a considerable improvement on what retroactively became the Model I.
2400 baud networking? Kids these days. The original TRS-80 had a 300 baud serial port, if you paid enough to get one.
The PETs of the time had crappy keyboards (even by the microcomputer standards of the time) and were limited in the amount of memory they could have. One novel feature was the all-in-one case that had the chicklet keyboard, cassette drive, and display. For general use, the bus wasn't unusually good, and it lacked in other factors.
I did some comparison shopping in late 1977. I might have gotten an Apple if they were cheaper.
Entering BASIC programs on the TRS-80 was absolutely standard. There was nothing strange or confusing about it, if you were familiar with Microsoft BASIC or the original time-shared version. It actually had a somewhat superior Microsoft BASIC compared to many computers, since it had 12K rather than 8K ROMs.
The original TRS-80s were a hack job, because Radio Shack was taking a chance on them. The second production run was sized so that, if it proved a bust, there'd be a TRS-80 for each Radio Shack store, for use for inventory or something. The TI-99/4 was introduced a couple of years after that, and the TRS-80 Model III the year after that.
I don't remember the EX designation, but the 1000 series was designed to be better than contemporary IBM PCs. There were versions of some business software for it (like Visicalc and some word processing program), and the color display was far better than what IBM had at the time. This was before we had IBM-compatible ROMs, so any non-IBM computer was nonstandard to some extent.
If you bought the Level II (i.e., Microsoft) BASIC, you could type in all sorts of BASIC programs. I've still got some from the time period that were apparently designed for teletypes on time-sharing systems. Later on, there were plenty of programs for the system.
Most home computers had some sort of cassette storage system in those days. I had some friends come out with some air combat games that were cassette-only because anything on a disk drive could be copied.
CP/M required floppy disks, which were somewhat affordable by 1981. The first TRS-80s in that series (i.e., not the Model II, which was something different) that were able to run standard CP/M were the Model 4s, which came out in 1983. I never heard anyone refer to the ROM as an OS, but if you got a floppy disk you could run TRS-DOS.
Binary code is the same as octal or hex. It depends on how it's written, but it's 8-bit numbers in memory . The normal practice back then was to express 8-bit numbers in decimal and POKE them into memory locations.
The Z80 was not microcoded, IIRC, but rather executed the instructions in hardware.
Slashdot Mirror
According to Wikipedia, Noonan vs. Staples is not a valid precedent. The validity of the Massachusetts law was untested, as both parties assumed it was valid and neither challenged it. Any relevant further case could challenge the Constitutional validity of the law.
As I understand it, not being a lawyer, the ruling was that APIs can be copyrighted, since they are creative works in fixed form.
However, copyrights can't prevent you from doing something other than very specific actions, and so if you use an API to write a program or library you have to use that API, it isn't infringing.
The question after that was whether Google's use of the Java API was legitimate. Oracle was arguing that it wasn't to create a Java program or library, since Android programs were not designed to interoperate with standard Java programs, and therefore Google was using it only because it was a well-known API.
I don't remember a severability clause in GPLv2, so the court has the choice between holding the license to be legally valid or legally invalid, nothing in between. If ruled as invalid, then GRSecurity had no license to distribute in the first place.
The GPL points out that you don't have to accept the GPL, but in that case you don't have a license, and may not copy, change, or redistribute the software.
However, they tend not to declare the contracts invalid. They tend to interpret the clauses in a restrictive way. If the court absolutely disagreed with Perens' interpretation, the court could rule that "further restrictions" does not clearly apply to what GRsecurity is doing, and hence GRsecurity is in the clear on their actions.
Of course, GRsecurity would have to prove not only that Perens was wrong, but that he had no good reason to think he was right, to win their lawsuit.
Sure. However, the GPL appears to forbid such agreements relating to GPLed software. The contract isn't automatically illegal (like a contract giving an employer rights to stuff an employee does on his or her own time and resources in this state), but it appears to me to violate the GPL. After all, I can generally offer any legal contract, but I can be in an agreement where I can't offer a specific one.
Your disagreement with the court is immaterial here, and doesn't actually change anything. If GRsecurity violated its license, then it doesn't have a license, and all the copies made and distributed are unlicensed, and that's good for a substantial amount of statutory damages (courtesy of the MAFIAA).
Sure. However, offering your services to the general public except for "those people" is potentially infringing a right.
GRSecurity cannot grant a license to the Linux kernel if GRSecurity doesn't have a valid license. If they have violated the GPL, then they don't have a valid license. These licenses aren't free-floating; legally, they have to be granted. (The question of what you do when you violate the GPL and lose your license can get rather involved, and GPLv3 provided for automatic reinstatement of the license under certain conditions - however, Linux is GPLv2 only, and is not distributed under GPLv3.)
That sounds an awful lot like adding terms to the GPL, which is not permitted by the GPL.
Therefore, I'm claiming the following things about reality. GRsecurity may be violating GPLv2 (I'm not taking a definite position on that). If so, GRsecurity doesn't have a valid license for the Linux kernel, and is forbidden to change or further copy it. If GRsecurity doesn't have a valid license, GRsecurity can't grant a license, and therefore their customers are running unlicensed copies of software. We know from the MAFIAA and lawsuits that illegitimate copies of copyrighted works can cost a whole lot of money. The Linux kernel does not operate under a copyright-assignment principle, so there's a large number of people with copyrighted code in the kernel, and I believe any of them could sue.
Hence, it looks legally risky to me to rely on a kernel supplied from GRsecurity.
However, if the statements are false, but Bruce has reason to believe they're true, it's still not libel. US law is stringent about what constitutes libel, and it can be hard to prove it.
If there were 8K TRS-80s (I'm not quite sure) they were pretty darn rare. The sizes they came up with normally were 4K and 16K.
If you were into assembly language programming, the CoCo had a sweet, sweet 6809 from the start. Loved that chip.
If it needed a lowercase conversion, it was the retroactively designated Model I, not the Model II which was a large box with 8" disk drives. You may be thinking of Level 2 BASIC, the Microsoft stuff.
Radio Shack offered the 8-bit character expansion after some time, which involved soldering a memory chip piggy-back and replacing the character generator.
There was a special CP/M version for the TRS-80 Model III, but the Model 4 could run standard CP/M software and that's what I got.
Visicalc was offered for the TRS-80 (I had a copy), but it came out first for the Apple, and got identified as Apple software. Accountants were spending their own money on Apple IIs and bringing them into work, and although that had to happen with some of the other computer types I never heard about it. It gave Apple a boost, and some business cred, but wasn't decisive.
What happened was the IBM PC, which soon took over the industry. The old Apple II and TRS-80 computers stayed in production for some time, but couldn't sell in nearly as great numbers. The PET had long gone by the wayside.
The three of them tried making their mark in different ways. Radio Shack tried coming out with a line of IBM PC-like computers that were superior, but this was before IBM-compatible ROMs, so they died in the marketplace, and Radio Shack continued making PC clones for a while. Commodore came out with the 64 and later the Amiga, which were innovative, very good, and didn't last. (IIRC, Commodore wasn't as committed to the industry.) Apple fumbled with the Apple III, produced the Lisa that was too expensive for home users, and finally the Macintosh, which was successful as the only real competitor to the PC and PC clones in the marketplace.
Another thing about $A and $B was that there was that you could type the strings in and print them out, and do absolutely nothing additional with them. I thought the most interesting thing about the TRS-80 Level 1 BASIC were the error messages: WHAT? (syntax error), HOW? (runtime error) and SORRY (ran out of memory). There were (IIRC) 26 16-bit integer variables, the two strings (limited length, maybe 16 characters each), and one array A[] or A() that took up the rest of memory.
The TRS-80, initially, had no way of making sound other than the tape relay, which was not a good musical device had had the potential of wearing out the relay. (I no longer have my technical manual, so I can't tell you where the relay was or what it was actually supposed to do).
The really fun thing about the 300 baud acoustic modem was that, when the phone wasn't plugged in but it was on, I could whistle to it and it would whistle back.
The Model II was Radio Shack's attempt to get into business programming, with 64K RAM, a large screen, 8" disk drives, and an impressive price tag. The Model III was a considerable improvement on what retroactively became the Model I.
2400 baud networking? Kids these days. The original TRS-80 had a 300 baud serial port, if you paid enough to get one.
The CoCo was a sweet machine, particularly the 2 which had a halfway decent keyboard.
The PETs of the time had crappy keyboards (even by the microcomputer standards of the time) and were limited in the amount of memory they could have. One novel feature was the all-in-one case that had the chicklet keyboard, cassette drive, and display. For general use, the bus wasn't unusually good, and it lacked in other factors.
I did some comparison shopping in late 1977. I might have gotten an Apple if they were cheaper.
Note that inflation has been around a factor of four since then, so the lowest-cost one would cost about $40K in today's dollars.
Entering BASIC programs on the TRS-80 was absolutely standard. There was nothing strange or confusing about it, if you were familiar with Microsoft BASIC or the original time-shared version. It actually had a somewhat superior Microsoft BASIC compared to many computers, since it had 12K rather than 8K ROMs.
The original TRS-80s were a hack job, because Radio Shack was taking a chance on them. The second production run was sized so that, if it proved a bust, there'd be a TRS-80 for each Radio Shack store, for use for inventory or something. The TI-99/4 was introduced a couple of years after that, and the TRS-80 Model III the year after that.
I don't remember the EX designation, but the 1000 series was designed to be better than contemporary IBM PCs. There were versions of some business software for it (like Visicalc and some word processing program), and the color display was far better than what IBM had at the time. This was before we had IBM-compatible ROMs, so any non-IBM computer was nonstandard to some extent.
If you bought the Level II (i.e., Microsoft) BASIC, you could type in all sorts of BASIC programs. I've still got some from the time period that were apparently designed for teletypes on time-sharing systems. Later on, there were plenty of programs for the system.
Most home computers had some sort of cassette storage system in those days. I had some friends come out with some air combat games that were cassette-only because anything on a disk drive could be copied.
CP/M required floppy disks, which were somewhat affordable by 1981. The first TRS-80s in that series (i.e., not the Model II, which was something different) that were able to run standard CP/M were the Model 4s, which came out in 1983. I never heard anyone refer to the ROM as an OS, but if you got a floppy disk you could run TRS-DOS.
Binary code is the same as octal or hex. It depends on how it's written, but it's 8-bit numbers in memory . The normal practice back then was to express 8-bit numbers in decimal and POKE them into memory locations.
The Z80 was not microcoded, IIRC, but rather executed the instructions in hardware.