As far as I know the Intergraph patents did not relate to SMT/hyperthreading.
Re:MULTIthreading != Hyperthreading
on
AMD Quad Cores, Oh My
·
· Score: 2, Informative
Are you sure Intergraph invented it? As far as I know hyperthreading, otherwise known as simultaneous multithreading, was invented by DEC and a research group at the University of Washington.
The EV8 was indeed to have 4 such 'logical' processors, but was cancelled. The EV7 did not have SMT.
IBM was not happy at all with all the clones. The BIOS really did have to be reverse-engineered. Of course, without the clones the PC would not have taken off but that's something else. Actually, IBM published the BIOS code on purpose, with the aim of polluting the pool of outside programmers.
Also, I seem to remember that the reverse-engineering bit of the PC BIOS re-implementation process was done by analyzing the copyrighted BIOS code. You can't really expect a person that has seen, read and understood the code to provide new code with the same functionality and free of copyright issues.
Afaik, Tridgell did not look at the bitkeeper source or its derived object code, but instead analyzed its input and output to derive the protocols. As long as whatever he looked at is free from the Bitkeeper copyright, it should be fine (although this might vary between national laws).
As far as I can tell, they are different persons. The Dutch Advocaat-Generaal does not advise the government in legal matters (as the Attorney-General does in many countries, see wikipedia), but advises the Supreme Court (Hoge Raad).
In fact, it is wrong to talk about "the" AG, since there are several AGs advising the Hoge Raad and they are placed under "the" Procureur-Generaal. The PG and AGs of the Hoge Raad are primarily advisers, not prosecutors. I don't know if Supreme Courts in other countries have similar advisers, but I suppose that at least those countries whose judicial system is based on the French system have.
The European Court of Justice has similar advisers, and those are called Attorneys-General in English. I guess different court systems just make it impossible to come up with fully consistent naming;)
Indeed, the company can sue the employee. Their agreement only binds the two of them, so if the company doesn't get what the employer is supposed to give them, they can sue.
And about the GPL: no matter what the company does, the agreement only binds the two of them, so it cannot override the GPL that's already stuck to the code.
See here for a whole discussion from December 2001 and here for a suggestion for an IsNot operator dated January 2001. The patent application was filed in May 2003. Anyway this application is just funny and I hope it'll be granted just to watch the effect.
I would advise you to replace 8192 by 4095, or you will get funny effects from the VIC chip. Actually I found out the hard way about pointers and pokes and peeks after struggling to find out why my bouncing ball program destroyed itself (the ball bounced out of the screen area 1024-2023 into the program area starting at 2049).
Claim 1 does not depend on any other claim, and is therefore called an 'independent' claim. The independent claims define the scope of protection: something infringes the patent if and only if it has all the features of one of the independent claim (roughly speaking).
The dependent claims define smaller areas of protection by adding more features to those of an independent claim. The dependent claims are useful in case it turns out in court that an independent claim is not valid because of some prior art. The dependent claims then serve as a fall back.
It's only a patent application. Claim 1 seems to specify the idea of determining that two pointers are not aliases, which is of course well-known in compiler technology. This one is so broad that I expect even the USPTO will not grant it, at least not in the present form.
It's all caused by a press statement of a "Hogeschool" which is a level below academia. They don't have research-level mathematics teachers there to tell the student (and the school) that it's a nice try (and nothing more).
Exactly. And actually, you just proved that his method must contain a flaw: all the calculations are performed within the real numbers (even at the point where he needs to calculate the root of an equation of degree n-1, since obviously one should use recursion to get at that one). So he always finds a real-valued root. Maybe someone can try his method on x^2+1=0.
And another argument: even if the whole thing is algebraically correct, and the final result actually does converge, the method would still require the above-mentioned recursion to get an approximation of approximations of approximations (etc..) of the desired coefficients.
Yes, it claims to converge. However his reasoning in this part doesn't seem to follow much logic. The whole thing actually seems to come down to: first solve an infinite system of equations to get some coefficients, then see if it converges and if it doesn not, go back to the original equation and divide the coefficients by some large number. Of course, dividing the coefficients will change all the earlier calculations, so there is no reason why the result will then converge.
Still, RSA was not a trivial idea, so without misters R, S and A we might still have to do without it. In this case there is at least some benefit to society.
What are subclaims? If you mean the dependent claims ("a method according to claim 1 wherein..."), then remember those contain all the steps or parts of the claim they depend on.
No, you don't have to patent something to prevent some random guy from patenting that something. All you need to do is to disclose it, for example by offering it in your products. So Apple did what it had to. The patent should simply not have been granted.
Do you know the average cost of challenging a patent in court? This patent is worth exactly that amount. It is certain to fail in court, but it costs money. I don't know if you could get your legal costs reimbursed by the patent holder suing you, since after all, he did get that patent granted, so legally you can't really blame him for trying to use it (but I'm not an expert here...).
Slashdot Mirror
As far as I know the Intergraph patents did not relate to SMT/hyperthreading.
Are you sure Intergraph invented it? As far as I know hyperthreading, otherwise known as simultaneous multithreading, was invented by DEC and a research group at the University of Washington.
The EV8 was indeed to have 4 such 'logical' processors, but was cancelled. The EV7 did not have SMT.
Marc Andreessen it was.
Others think that the votes reflect traditional antipathies between adjoining countries, meaning the Dutch won't vote for the Belgians.
No no no!!! The Belgians won't vote for the Dutch, the Dutch won't vote for the Germans.
Hmmm, then why did people claim Quake and/or Quake2 had been the perfect deathmatch FPS when Q3 came out?
:)
It just depends on what you grow up with.
Ehm, deathmatch Pong rulez
IBM was not happy at all with all the clones. The BIOS really did have to be reverse-engineered. Of course, without the clones the PC would not have taken off but that's something else. Actually, IBM published the BIOS code on purpose, with the aim of polluting the pool of outside programmers.
Also, I seem to remember that the reverse-engineering bit of the PC BIOS re-implementation process was done by analyzing the copyrighted BIOS code. You can't really expect a person that has seen, read and understood the code to provide new code with the same functionality and free of copyright issues.
Afaik, Tridgell did not look at the bitkeeper source or its derived object code, but instead analyzed its input and output to derive the protocols. As long as whatever he looked at is free from the Bitkeeper copyright, it should be fine (although this might vary between national laws).
See here for the BIOS reverse engineering story.
As far as I can tell, they are different persons. The Dutch Advocaat-Generaal does not advise the government in legal matters (as the Attorney-General does in many countries, see wikipedia), but advises the Supreme Court (Hoge Raad).
;)
In fact, it is wrong to talk about "the" AG, since there are several AGs advising the Hoge Raad and they are placed under "the" Procureur-Generaal. The PG and AGs of the Hoge Raad are primarily advisers, not prosecutors. I don't know if Supreme Courts in other countries have similar advisers, but I suppose that at least those countries whose judicial system is based on the French system have.
The European Court of Justice has similar advisers, and those are called Attorneys-General in English. I guess different court systems just make it impossible to come up with fully consistent naming
Indeed, the company can sue the employee. Their agreement only binds the two of them, so if the company doesn't get what the employer is supposed to give them, they can sue.
And about the GPL: no matter what the company does, the agreement only binds the two of them, so it cannot override the GPL that's already stuck to the code.
See here for a whole discussion from December 2001 and here for a suggestion for an IsNot operator dated January 2001. The patent application was filed in May 2003. Anyway this application is just funny and I hope it'll be granted just to watch the effect.
I would advise you to replace 8192 by 4095, or you will get funny effects from the VIC chip. Actually I found out the hard way about pointers and pokes and peeks after struggling to find out why my bouncing ball program destroyed itself (the ball bounced out of the screen area 1024-2023 into the program area starting at 2049).
Claim 1 does not depend on any other claim, and is therefore called an 'independent' claim. The independent claims define the scope of protection: something infringes the patent if and only if it has all the features of one of the independent claim (roughly speaking).
The dependent claims define smaller areas of protection by adding more features to those of an independent claim. The dependent claims are useful in case it turns out in court that an independent claim is not valid because of some prior art. The dependent claims then serve as a fall back.
It's only a patent application. Claim 1 seems to specify the idea of determining that two pointers are not aliases, which is of course well-known in compiler technology. This one is so broad that I expect even the USPTO will not grant it, at least not in the present form.
I'm not quite sure I can follow this, but see
this story by Maureen O'Gara. Hmmm, maybe they both live in the Netherlands.
As I one-clicked away my previous post I realized that Amazon's was the one-click-buy patent.
and not Microsoft.
It's all caused by a press statement of a "Hogeschool" which is a level below academia. They don't have research-level mathematics teachers there to tell the student (and the school) that it's a nice try (and nothing more).
Exactly. And actually, you just proved that his method must contain a flaw: all the calculations are performed within the real numbers (even at the point where he needs to calculate the root of an equation of degree n-1, since obviously one should use recursion to get at that one). So he always finds a real-valued root. Maybe someone can try his method on x^2+1=0.
And another argument: even if the whole thing is algebraically correct, and the final result actually does converge, the method would still require the above-mentioned recursion to get an approximation of approximations of approximations (etc..) of the desired coefficients.
Uhm, no this paper is not at all straightforward. I would be really surprised if it can be made to work for quadratic equations.
Yes, it claims to converge. However his reasoning in this part doesn't seem to follow much logic. The whole thing actually seems to come down to: first solve an infinite system of equations to get some coefficients, then see if it converges and if it doesn not, go back to the original equation and divide the coefficients by some large number. Of course, dividing the coefficients will change all the earlier calculations, so there is no reason why the result will then converge.
Still, RSA was not a trivial idea, so without misters R, S and A we might still have to do without it. In this case there is at least some benefit to society.
Since the Microsoft patent application was filed on May 20, 1999, any art from 2000 is not prior...
What are subclaims? If you mean the dependent claims ("a method according to claim 1 wherein ..."), then remember those contain all the steps or parts of the claim they depend on.
No, you don't have to patent something to prevent some random guy from patenting that something. All you need to do is to disclose it, for example by offering it in your products. So Apple did what it had to. The patent should simply not have been granted.
Do you know the average cost of challenging a patent in court? This patent is worth exactly that amount. It is certain to fail in court, but it costs money. I don't know if you could get your legal costs reimbursed by the patent holder suing you, since after all, he did get that patent granted, so legally you can't really blame him for trying to use it (but I'm not an expert here...).