Quote: The software is protected -- by copyright law.

The difference between software and hardware can sometimes run thin. What runs in software some day might be run using dedicated silicon circuitry the following day. Cases in mind: alpha blending, Z-buffer, texture compression. Copyright laws won't help you here.

Quote: The biggest problem with software patents is that software is already protected by copyright.

The legal protection between the two is not the same, though.

Maybe I should make myself clearer here: by software, I mean algorithms.

-cb

[Edited by - cbenoi1 on May 1, 2006 7:09:19 PM]

Quote: Original post by cbenoi1
Maybe I should make myself clearer here: by software, I mean algorithms.

Ah... they should not be covered by anything. You can protect an implementation of an algorithm (say, in hardware and patent it, or in software and copyright it).

Quote: Original post by chollida1
You could just write something original:) But seriously I believe you need to strike a balance between protecting R&D IP that was done by a company and lowering barriers to entry for other companies.

It's hard to be entirely original when so many things are patented. It's not really a "balance" when the existing companies are not held to the same standard (as they simply licence each others' IP). And the R&D IP is already protected by copyright law.

Quote: Original post by Sander

Quote: Original post by cbenoi1
Maybe I should make myself clearer here: by software, I mean algorithms.

Ah... they should not be covered by anything. You can protect an implementation of an algorithm (say, in hardware and patent it, or in software and copyright it).

I'd have to argue that point. :) You're assuming that the algorithms involved were relatively easy (hence cheap) to develop and that the implementation of said algorithms is where the real money is spent. While this is true most of the time, there could be cases where the cost of developing the algorithms themselves far exceeds the cost of implementation.

As a theoretical example, suppose a company developed algorithms detailing the interaction between a complex organic organism (like the human body) and a foreign compound (think drug research). Now pursuing such development would undoubtedly take many decades and cost hundreds of billions of dollars in research. So let's say that a company goes ahead and invests the required capital, and in 50 years has finally developed a working model. The company is then going to want to recoup their massive investment, so they'll undoubtedly sell an implementation of their previous effort, whether software or hardware based, to other companies for large sums of money, plus royalties. Now human nature being what it is, it wouldn't be very long before their product is reverse-engineered and the algorithms extracted from their implementation. So without patents that protect the algorithms themselves, what's to stop a competitor from stealing them, writing their own implementation, and then undercutting the original company that invested all of those billions right out of the market? Admittedly this is an extreme example, but I believe that in cases such as this software patents would serve a valuable purpose.

Now having said that, I believe that software patents as they stand today should be eliminated, with a new system put in place that only protects algorithms that are developed at a significant financial cost. The current system is simply too easy to abuse, but that doesn't mean that there aren't any cases where protecting an "idea" is necessary in order for that idea to take fruit. [smile]

1) It's impossible to extract the software model/alogorithm used for creating the drug you speak of simply by looking at the drug itself.
2) The drug itself can be patented, so the competitors chemically reversed engineered product that undercuts you is infringing.

So from a software parent point of view the example doesn't hold water. I'm not going to try arguing for/against drug patents. That's a whole mess on their own which I'll gladly leave alone for now.

Quote: Original post by Sander
1) It's impossible to extract the software model/alogorithm used for creating the drug you speak of simply by looking at the drug itself.
2) The drug itself can be patented, so the competitors chemically reversed engineered product that undercuts you is infringing.

So from a software parent point of view the example doesn't hold water. I'm not going to try arguing for/against drug patents. That's a whole mess on their own which I'll gladly leave alone for now.

Sorry, I guess I should have been more clear. :) I didn't mean that a company who developed such software would sell drugs researched with it, I meant that the company (quite possibly a biotech firm and not a pharmaceutical) would sell the software system itself. As the system would be built upon a complex series of algorithms developed at great expense, the company would run the risk that someone else would reverse-engineer it and avoid copyright law by writing their own implementation using the stolen algorithms. The second company could then claim that they independently developed the algorithms and avoid any litigation.

The point I was trying to make is that algorithms are not always intuitive, easy things to come up with, and thus should sometimes enjoy the legal protections that encourage necessary capital investment. In my "Holy Grail of drug testing" example, serious amounts of research would need to done in order to develop the appropriate scientific/mathematical understanding of the human body. Now if all of this research is funded by governments and universities then there's no need to protect it, but if one expects private enterprise to foot the bill then they're going to want some assurances that they'll be able to protect the fruit of their research from being stolen by their competitors. [smile]

Quote: You can protect an implementation of an algorithm (say, in hardware and patent it, or in software and copyright it).

Copyrights do not protect the inherent algorithms (or methods herein), which are the actual valuable pieces of the whole software application. An astute programmer can extract those critical algorithms from the application and integrate them into other software applications. In that case, copyrights will only protect litteral copying (or "derivative works") of the entire application but not those valuable pieces of know-how.

-cb

Quote: Original post by Ramius
Sorry, I guess I should have been more clear. :) I didn't mean that a company who developed such software would sell drugs researched with it, I meant that the company (quite possibly a biotech firm and not a pharmaceutical) would sell the software system itself.

1) The software is protected by copyright so copying is out. As for reverse engineering it, yeah that's possible. But if company A had to spend millions in order to research the algorithm itself (before even writing the software) then you can bet company B needs to spend vast ammount of money to be able to reverse engineer it. Think millions as well. So company A is quite safe. That is, even if it can be done at all. I think that the complexity of the algorith to make the scenario possible would be nearly impossible to reverse engineer.

2) Now suppose company B does the research itself as well. It also spends the millions and also makes a software model to sell. Should company A be able to put them out of business? Ofcourse not.

So, should the whole of the software industry suffer from the drawbacks of software patents just so a very, very rare and quite theoretical scenario as the above can be covered? I think not.

Would the algotithm still have been researched in a software patent free world? Most likely. Maybe by a university rather than company A, but company A has ample oportinity to make money from the algorithm even without software patents.

Honestly I'm starting to wonder how it is there's a serious debate on this topic. I've spent a fair amount of time looking around on Google and Wikipedia, and the whole matter seems pretty one-sided against software patents. For instance, if you look at the external links section of Wikipedia's Software Patent Debate article, there's six sites listed supporting software patents, and many times that many in opposition. Of course I understand that could represent a bias on the author's part, but ultimately I've noticed about the same in looking around on Google. I can appreciate the comments of those of you who support software patents at least to some extent, but I'm wondering if you could point out any particular articles or organizations that make a case for software patents, whether you agree with them or not.

And of course, any comments you have on these perspectives are welcome as well.

A few of the meager examples I've been able to locate so far:
Cisco Systems: Defending Innovation
Shoring up Rights to Technology Inventions

While I'm too lazy to compile an actual list of software patent advocates, I'd recommend checking out organizations that have a lot to lose if such patents didn't exist. :) This should apply to most cutting edge, time consuming research endeavors such as advanced image/voice recognition software, sophisticated compression techniques, voice synthesis software, and pretty much anything biotech. Basically any company that has to devote considerable resources to developing new levels of understanding required for algorithm construction is likely going to want at least limited software patent protection. Corporations that could care less would probably include any who rely primarily on common, well understood, and most importantly public domain scientific knowledge and research. If a given algorithm takes a couple of averagely skilled guys a day to come up with then there's really not much value in protecting it, is there? Whereas if it's the result of a year of research by a large team then there's likely to be a lot more interest expressed in protecting it. [smile]

As I stated earlier, I'm actually against the software patent system that we have in the United States currently. It seems that it's far too easy for a corporation/individual to abuse, so I'm all for some legal restructuring. However, I believe that the assumption that algorithms are never in need of any kind of legal protection is just as faulty as the current software patent system itself. Just my opinion though. :)