4 ms·
If you don't want to read through all the explanation of what patents are and why the they are too strict here is the two paragraphs of news: "America’s Suprem
by csmeder 17y ago
If you don't want to read through all the explanation of what patents are and why the they are too strict here is the two paragraphs of news:
"America’s Supreme Court is about to issue a ruling which, by all accounts, will make it difficult, if not impossible, to get a patent for a business process. And because most business processes are, at bottom, computer algorithms, the Supreme Court’s judgment could also bar all sorts of software patents in the process. As a result, a lot of patents for online shopping, medical-diagnostic tests and procedures for executing trades on Wall Street could be invalidated.
The roots of the dispute go back more than a decade to two inventors, Bernard Bilski and Rand Warsaw, who sought to patent a method of hedging weather-related risks in energy prices. The USPTO concluded that the process was too abstract and denied the application. On appeal, the Federal Circuit actually upheld the patent office’s decision—and said, crucially, that the process would be eligible for a patent only if it was “tied to a particular machine or apparatus” or if it “transformed a particular article into a different state or thing”. Failing on both counts, the inventors took their case to the Supreme Court, which has been pondering for the past few months whether the “machine-or-transformation” test is an appropriate standard. It is hard to conclude it is not. "
- Sukotto 17y agoAbout frikken time (imho)
- cduan 17y agoNot to be nitpicky with the article, but it's a little misleading when it says (right above this quoted section): "What is clear is that the “non-obviousness” part of the test for patentability has not been applied anywhere near rigorously enough to internet and business-process patents. Because they lack a history of “prior art” to refer to, examiners and judges have granted a lot of shoddy patents for software and business processes." Bilski has nothing to do with obviousness. Bilski is about the subject matter requirement, which is that patents can only be granted on certain types of inventions. For example, physical machines can be patented, but electronic signals cannot. This is another reason (in addition to Grellas's below) that I doubt Bilski will have a big effect on software patents. Regardless of the outcome of the case, it will be most likely easy to work around, simply by tying the software to some sort of machine, be it a computer, mobile phone, embedded circuit, or anything else.