You are currently viewing our boards as a guest, which gives you limited access to view most discussions and access our other features. By joining our free community, you will have access to post topics, communicate privately with other members (PM), respond to polls, upload content, and access many other special features. Registration is fast, simple, and absolutely free, so please, <a href="/profile.php?mode=register">join our community today</a>!
''We believe that Nvidia has hired former Transmeta staff extensively, and that instruction code "morphing" requirements have declined as more x86 instructions have come off of patent coverage,'' he said.
I guess I hadn't really thought about it in detail, but I assumed the rights Intel had over x86 were more than just patents that would expire, they were more along the lines of copyright or similar that does not expire. Is that not the case? If they were merely protected by patents then all the pre-SSE instructions would long ago have become public. Once the patents expire, is anyone free to make a processor that can execute those instructions natively? Does each separately introduced extension to the x86 instruction set have its own 14 or 20 year patent?
''We believe that Nvidia has hired former Transmeta staff extensively, and that instruction code "morphing" requirements have declined as more x86 instructions have come off of patent coverage,'' he said.
I guess I hadn't really thought about it in detail, but I assumed the rights Intel had over x86 were more than just patents that would expire, they were more along the lines of copyright or similar that does not expire. Is that not the case? If they were merely protected by patents then all the pre-SSE instructions would long ago have become public. Once the patents expire, is anyone free to make a processor that can execute those instructions natively? Does each separately introduced extension to the x86 instruction set have its own 14 or 20 year patent?
Patents don't cover instructions, they cover efficient ways of implementing instructions. The 386 patents are set to expire soon, however all the patents for the 486, Pentium, P6, MMX, SSE1-4, virtualization, etc. are all still in force.
You cannot copyright hardware... I guess that is how confused people get from software being both copyrightable and patentable (which makes no bloody sense to begin with).
Anyways, they can buy VIA's CPU business and get the necessary patents that way. You should be able to license X86-64 from AMD, but you cannot sell a processor in the market today without SSE extensions. Lots of software won't even work without them.
software being both copyrightable and patentable (which makes no bloody sense to begin with).
An algorythm may be patentable, not a software or a software implementation of it, a software implementation may be copyrightable.
For example, the gif format, the compression algorythm is patented (note, on some countries algorythms aren't patentable), so nobody may implement such algorythm (and thus, compress or decompress gifs) without a license, but this patent is going to expire soon, after that everybody may implement it, but if I implement all the source code will be copyrighted to me and only those with a license may use it, you may make another piece of code to make the same thing (another implementation) and so, note copyrigths take too long to expire.
In the case of instructions sets, I can't see how an instruction may be patented, the algorythm to implement it may, but as most instructions do trivial operations (add, for example) patents are likely to be in public domain for a long time, if patentable at all.
For example, the gif format, the compression algorythm is patented (note, on some countries algorythms aren't patentable), so nobody may implement such algorythm (and thus, compress or decompress gifs) without a license, but this patent is going to expire soon, after that everybody may implement it
FYI the patent on LZW (compression algorithm for GIF) expired a few years ago already
In the case of instructions sets, I can't see how an instruction may be patented, the algorythm to implement it may, but as most instructions do trivial operations (add, for example) patents are likely to be in public domain for a long time, if patentable at all.
If a file format is patentable, why a instruction format can't be patentable too? Have x86,x86-64 etc been patented?
A file format isn't patentable, algorythms used to generated its content mnay be patentable.
yfe wrote:
why a instruction format can't be patentable too?
Because file formats can't to start with...
yfe wrote:
Have x86,x86-64 etc been patented?
I never saw any patent covering an ISA and can't imagine how they could be patentable, also, Intel lost every law suit they filed against other companies about instruction sets.
I never saw any patent covering an ISA and can't imagine how they could be patentable,
One of the LISP machine vendors put the complete low level micro code into a patent. I have never read it myself, but seen references to it. That's not a ISA, but large parts of it.
Here's a guy who tried to OCR it and run it again on Franz Lisp:
software being both copyrightable and patentable (which makes no bloody sense to begin with).
An algorythm may be patentable, not a software or a software implementation of it, a software implementation may be copyrightable.
For example, the gif format, the compression algorythm is patented (note, on some countries algorythms aren't patentable), so nobody may implement such algorythm (and thus, compress or decompress gifs) without a license, but this patent is going to expire soon, after that everybody may implement it, but if I implement all the source code will be copyrighted to me and only those with a license may use it, you may make another piece of code to make the same thing (another implementation) and so, note copyrigths take too long to expire.
In the case of instructions sets, I can't see how an instruction may be patented, the algorythm to implement it may, but as most instructions do trivial operations (add, for example) patents are likely to be in public domain for a long time, if patentable at all.
It's still stupid. A cooking recipe is an algorithm. Business methods are an algorithm. Patents were made because "copying" machines is a different process than copying literature and whatever. Yeah copyrights last too long as well. I guess we can thank Hollywood for that.
what makes you think GIF use arithmetic coding ?, it was plain impractical (way too slow) in the pre-Web Compuserve era
Well... The algorythm allows it... But ok, I really wasn't sure about it, if expired then fine, not that I care very much about image formats.
andikleen wrote:
One of the LISP machine vendors put the complete low level micro code into a patent. I have never read it myself, but seen references to it. That's not a ISA, but large parts of it.
Here's a guy who tried to OCR it and run it again on Franz Lisp:
Well... It's not the ISA itself and for lines of code they could got a longer protection with a simpler and cheaper process by copyright laws.
cheesybagel wrote:
It's still stupid. A cooking recipe is an algorithm.
Kind of, once I saw a patent of a method to convert a G3 into a bullpop rifle, wich is kind of an algorythm too.
Since I'm not very close to cooking activities I have no ideia if anyone have tried to patent something like that, but don't give the idea to others, they may suceed.
yfe wrote:
You sure? What about MP3 files? Makers have to pay for support of MP3 files in their players even if they use their own decoding algorithm. Isn't it?
If you change the decoding algorythm you won't be able to reproduce that, do not confuse algorythm and implementation.
Well... The algorythm allows it... But ok, I really wasn't sure about it, if expired then fine, not that I care very much about image formats.
not that I care much myself about GIF, but data compression in general is a very rich subject when discussing software patents since it's such a minefield now honestly the distinction between algorithm and file format isn't that clear cut / practically useful IMHO. If I have to pay a license to implement myself a codec to support a given file format because the algorithm is patented, for all practical purposes it's the same thing than saying the file format itself is patented
Users browsing this forum: No registered users and 1 guest
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot post attachments in this forum
