Software Patents
-
I've heard that software isn't a patentable item is better protected under copyright law. However, with all the publicity over software patents lately, I'm wondering how people are able to obtain a patent for software, and how enforceable is a software patent?
-
I've heard that software isn't a patentable item is better protected under copyright law. However, with all the publicity over software patents lately, I'm wondering how people are able to obtain a patent for software, and how enforceable is a software patent?
Bob X wrote:
I've heard that software isn't a patentable item is better protected under copyright law. However, with all the publicity over software patents lately, I'm wondering how people are able to obtain a patent for software, and how enforceable is a software patent?
Software is very patentable under current laws. There is debate on if the laws should be changed to prevent patenting software. Copyright is on a whole item, not on a part, a patent is on a part never a whole. Names of processes are trademarked. Without a patent, you could make a checksum:
unsigned long checksumA (char *string)
{
unsigned short sum1=0,sum2=0;
while (*string)
{
sum1+=*string;
sum2+=sum1;
string++;
}
return ((long)sum1)<<16|(long)sum2;
}the name checksumA could be trademarked to prevent someone from calling their algorithm the same, even if it used the same process. If you submitted it as a whole, you could copywrite it, though anymore it would fail because it is known now as a part. But even if you copywrited it the code, substituting 0xdead and 0xbeef for the variable initial state would cause the program to be something new and therefore not fall under the original copywrite. Enter patents. As terrible as they were, they were made to protect that which could not otherwise be protected, algorithms that required much R&D, the investment in time is then turned into a patentable process. Otherwise the time is wasted because you invest a year in developing a new process, and a million dollars in the process, and someone backward engineers your process changes a couple of variables and it is declared "something new". But patents have also been used in newer ways for software than they had been under physical item patent laws. It was unheard of for a company to give out parts for 5 years that they held a patent on, then suddenly declare everyone using all those parts had to pay them $1 five years later. This was one of the major abuses in the system. But only one. Because algorithms employ mathematics, mathematics was once declared public and exempt from patent and trademark laws. But as an algorithm, you can patent your math and anyone employing a computer procedure of your mathematical algorithm for say, generating random numbers must pay or look elsewhere. It is complex, and not pretty, and no matter how you look at it is controversial. Sure investments in R&D need to be protected, how, that is the crux of
-
Bob X wrote:
I've heard that software isn't a patentable item is better protected under copyright law. However, with all the publicity over software patents lately, I'm wondering how people are able to obtain a patent for software, and how enforceable is a software patent?
Software is very patentable under current laws. There is debate on if the laws should be changed to prevent patenting software. Copyright is on a whole item, not on a part, a patent is on a part never a whole. Names of processes are trademarked. Without a patent, you could make a checksum:
unsigned long checksumA (char *string)
{
unsigned short sum1=0,sum2=0;
while (*string)
{
sum1+=*string;
sum2+=sum1;
string++;
}
return ((long)sum1)<<16|(long)sum2;
}the name checksumA could be trademarked to prevent someone from calling their algorithm the same, even if it used the same process. If you submitted it as a whole, you could copywrite it, though anymore it would fail because it is known now as a part. But even if you copywrited it the code, substituting 0xdead and 0xbeef for the variable initial state would cause the program to be something new and therefore not fall under the original copywrite. Enter patents. As terrible as they were, they were made to protect that which could not otherwise be protected, algorithms that required much R&D, the investment in time is then turned into a patentable process. Otherwise the time is wasted because you invest a year in developing a new process, and a million dollars in the process, and someone backward engineers your process changes a couple of variables and it is declared "something new". But patents have also been used in newer ways for software than they had been under physical item patent laws. It was unheard of for a company to give out parts for 5 years that they held a patent on, then suddenly declare everyone using all those parts had to pay them $1 five years later. This was one of the major abuses in the system. But only one. Because algorithms employ mathematics, mathematics was once declared public and exempt from patent and trademark laws. But as an algorithm, you can patent your math and anyone employing a computer procedure of your mathematical algorithm for say, generating random numbers must pay or look elsewhere. It is complex, and not pretty, and no matter how you look at it is controversial. Sure investments in R&D need to be protected, how, that is the crux of
RSA (whatever they are called these days) had/have some of the more famous patents for algorithms (employing a lot of math as you no doubt know). They were probably some of the earliest "patenters" (is that a word?) of software algorithms (many of their patents have finally expired). Of course, I am not typing this for your benefit, Corazon, as I am sure you are well aware of all this. Just throwing in my 2 cents to the discussion. :)
WE ARE DYSLEXIC OF BORG. Refutance is systile. Your a$$ will be laminated.
-
RSA (whatever they are called these days) had/have some of the more famous patents for algorithms (employing a lot of math as you no doubt know). They were probably some of the earliest "patenters" (is that a word?) of software algorithms (many of their patents have finally expired). Of course, I am not typing this for your benefit, Corazon, as I am sure you are well aware of all this. Just throwing in my 2 cents to the discussion. :)
WE ARE DYSLEXIC OF BORG. Refutance is systile. Your a$$ will be laminated.
Tom Delany wrote:
Of course, I am not typing this for your benefit, Corazon, as I am sure you are well aware of all this. Just throwing in my 2 cents to the discussion.
All pennies are welcome. :-D Yeah, the math issue actually came up with encryption in respect to national security. Many of the algorithms were required to be patented in order to protect the USA which until the last decade listed encryption right next to nuclear weapons in respect to priority related to national security. Transfer of an encryption algorithm to a foreign country was equivalent to transfer of nuclear technology to that same state. This meant that for its own protection the government deliberately set up the situation for patent protection being the easiest form of protection beyond tossing guns and arresting people. It also provided a papertrail that was international that they could follow. Why do all the work when you can get someone else to do half of it for you? Anyhow, the net result is that mathematics is ultimately patentable. The original intent was such that you could not patent say gravitation F=G*(M1*M2)/r2. But if you place said formula into a game engine and show that it has never been used in that way before and thus does not fall under common knowledge, you can get said formula patented. Theoretically, math is still not patentable, as long as you never attempt to "use" the math, any process that employs a patented mathematical process immediately falls under the patent, as you know with the RSA and others. You may discuss and even publish your math, as long as you do not attempt to program the math or even provide a step-through process of the math. Ironic as it is, suddenly the math is patented when reduced to process form either in software or in physical steps.
_________________________ Asu no koto o ieba, tenjo de nezumi ga warau. Talk about things of tomorrow and the mice in the ceiling laugh. (Japanese Proverb)
-
Tom Delany wrote:
Of course, I am not typing this for your benefit, Corazon, as I am sure you are well aware of all this. Just throwing in my 2 cents to the discussion.
All pennies are welcome. :-D Yeah, the math issue actually came up with encryption in respect to national security. Many of the algorithms were required to be patented in order to protect the USA which until the last decade listed encryption right next to nuclear weapons in respect to priority related to national security. Transfer of an encryption algorithm to a foreign country was equivalent to transfer of nuclear technology to that same state. This meant that for its own protection the government deliberately set up the situation for patent protection being the easiest form of protection beyond tossing guns and arresting people. It also provided a papertrail that was international that they could follow. Why do all the work when you can get someone else to do half of it for you? Anyhow, the net result is that mathematics is ultimately patentable. The original intent was such that you could not patent say gravitation F=G*(M1*M2)/r2. But if you place said formula into a game engine and show that it has never been used in that way before and thus does not fall under common knowledge, you can get said formula patented. Theoretically, math is still not patentable, as long as you never attempt to "use" the math, any process that employs a patented mathematical process immediately falls under the patent, as you know with the RSA and others. You may discuss and even publish your math, as long as you do not attempt to program the math or even provide a step-through process of the math. Ironic as it is, suddenly the math is patented when reduced to process form either in software or in physical steps.
_________________________ Asu no koto o ieba, tenjo de nezumi ga warau. Talk about things of tomorrow and the mice in the ceiling laugh. (Japanese Proverb)
El Corazon wrote:
Transfer of an encryption algorithm to a foreign country was equivalent to transfer of nuclear technology to that same state.
Yeah. I am well aware of that fact. I worked for a company that wrote a PC data security product in the pre-Windows, DOS days (Watchdog), that employed hard drive encryption. The product was listed on the US government's Munitions Control List. It was a real PITA to sell it in any foreign country (except Canada, as I recall). The company had to create an export version of the product that used a watered down encryption algorithm. They had to then apply for an export license from the US Dept. of State for every country where they wanted to sell it. The paperwork involved in preparing an export shipment was horrendous (I know because I worked in the shipping department, and had to fill it out). The export license had to be on file with US Customs at the appropriate port of departure, and had to be marked against every time we made a shipment. There were only a couple of carriers that would even be bothered in handling the shipments because of all the red tape and hoops that they had to jump through. It was all great fun (NOT!). Oh those were the days! :)
WE ARE DYSLEXIC OF BORG. Refutance is systile. Your a$$ will be laminated.
-
El Corazon wrote:
Transfer of an encryption algorithm to a foreign country was equivalent to transfer of nuclear technology to that same state.
Yeah. I am well aware of that fact. I worked for a company that wrote a PC data security product in the pre-Windows, DOS days (Watchdog), that employed hard drive encryption. The product was listed on the US government's Munitions Control List. It was a real PITA to sell it in any foreign country (except Canada, as I recall). The company had to create an export version of the product that used a watered down encryption algorithm. They had to then apply for an export license from the US Dept. of State for every country where they wanted to sell it. The paperwork involved in preparing an export shipment was horrendous (I know because I worked in the shipping department, and had to fill it out). The export license had to be on file with US Customs at the appropriate port of departure, and had to be marked against every time we made a shipment. There were only a couple of carriers that would even be bothered in handling the shipments because of all the red tape and hoops that they had to jump through. It was all great fun (NOT!). Oh those were the days! :)
WE ARE DYSLEXIC OF BORG. Refutance is systile. Your a$$ will be laminated.
Tom Delany wrote:
They had to then apply for an export license from the US Dept. of State for every country where they wanted to sell it.
And was only approved if and only if the watered down version was decryptable by NSA or equivalent agency based on the time period, AND that the algorithm didn't automatically evaluate to a stronger version with minor changes... and the guy signing it slept well that night. ;)
Tom Delany wrote:
Oh those were the days!
ahhh, the salad days of youth....
_________________________ Asu no koto o ieba, tenjo de nezumi ga warau. Talk about things of tomorrow and the mice in the ceiling laugh. (Japanese Proverb)
-
I've heard that software isn't a patentable item is better protected under copyright law. However, with all the publicity over software patents lately, I'm wondering how people are able to obtain a patent for software, and how enforceable is a software patent?
Software Patents aren't really Software Patents. They're Algorithm Patents, or Process Patents. Copyright protects an exact copy, or derived copies. Patents protect Ideas or Processes. For example, one could copyright an image of a patent attorney being sunk at the bottom of a lake, but not patent it. One could also patent the compression used in said picture, but not the image created by it.
-- Where are we going? And why am I in this handbasket?