3 ms·
The patent system of today depends on a single core concept left from the 19th century: Government knows everything. It's a 200 year old administrative artifact
by belorn 13y ago
The patent system of today depends on a single core concept left from the 19th century: Government knows everything. It's a 200 year old administrative artifact, removed from every aspect of law, except patents.
The assumption of the 19th century patent system was that everything that should be patented already was patented already; therefore, when you would bring a invention to the office, the correct way to determine whether it is novel is to ask whether it is novel in light of the art already in the office. When asking whether it is obvious, you are essentially asking whether the very semi-skilled form of art knowledge possessed by the patent examiner would have regarded this as a natural extension of the existing art deposited in the office.
In other words, determining a patent comes down to searching the patents offices' own archive. If the old patent texts would be so lucky as to say: "To unlock the cellphone, the user’s finger moves across the touch screen", it won't match when Apple adds the word "continuous". The patent examiner is forced into a position where the least resistance is simply to grant Apple the patent.
This system is not working. The patent office archives do not have every iteration and written description about everything. By certainty, they won't have all the implied descriptions. As such, one should not assume that it is efficient to do an binary determination - 0 or 1, novel or obvious or not - on the basis of the consultation of the patent offices' own administrative records.
The simple change suggested by the article would only remove the ability for companies to brute force a claim phrase so it does not match exactly something in the patent offices' archives. It would not fix the patent system, but it might patch it up so exploiters can't run a brute force method like some script kids attacking web forms and see what string returns a SQL injection vulnerability.
- pbhjpbhj 13y agoThe UK Patent Office certainly searches far, far more art than just the global patent literature. They make use of many specialist databases, extensive internet searches (carefully tailored not to disclose novel material), journals relevant to the art. The search and substantive examiners in the UK are one and the same (or were a few years ago) and the majority I knew of had doctorates. Your example of mismatch according to a word/phrase is laughable. Novelty means novelty of the invention [defined by the claims], no matter how it's claims are worded, not novelty of the claims. If indeed USPTO examiners are searching only on specific word forms then I'd be amazed; that certainly isn't the case in Europe. You give an example of adding "continuous" - the problem here lies in that either that is anticipated or the objection is that it's obvious based on prior art. If an obviousness objection is made the applicant can claim that "if it's obvious why hasn't anyone done it already" - that then forces their hand to leave the analysis to the courts. The classic UK example (quite old now) of the breadth of prior art is citing a children's comic against a "doggy doorbell". The breadth of related arts used in searches has vastly expanded since then. In addition to that search and examination is divided up based on the IPC/ECLA classification systems (and additional subdivisions cast over those that are local) - which allows for specialisation.
- belorn 13y ago> They make use of many specialist databases, extensive internet searches (carefully tailored not to disclose novel material), journals relevant to the art. How is that contrary to the core concept of the assumption that "Government knows everything"? If they do use specialist databases (source?), and subscribe to journals, it is still simply just more files collected in their computer system, which patent examiners use when searching key words/phrases located in the patent application. A few months, there was a article here on HN about how patent writers carefully choose words in patent application just so key word searching Google and patent databases won't result in prior art. The industry that writes the large mass of patents knows how to exploit the mismatch according to a word/phrase system. Europe patent offices might have a current better time dealing with this, since the brute force method illustrated in the article do not exist here. Like with SQL injections, its harder to do an exploit without the possibility to try until the claim string matches.
- pbhjpbhj 13y ago[to reiterate I can't address the specifics of the USPTO, but can address how patent search and examination work from personal experience] >which patent examiners use when searching key words/phrases located in the patent application. // You said examiners only search patent documents in their computer systems. That's just wrong. There are many commercial databases of technology - IBM's Technical Disclosure Bulletins (TDB) that they used for defensive publication for example or compendiums of papers from various scientific journals. All areas of technology have trade magazines that give details of current tech, enough to provide a suggestion of where to search in more detail. Of course the internet is a vast resource too >how patent writers carefully choose words in patent application just so key word searching Google and patent databases won't result in prior art // Can you link that. It sounds like nonsense. Claims use language such as "fixing means" to allow breadth to the legal monopoly defined encompassing all relevant/possible forms of fixation - glue, screws, bolts, magnetics, welding, flexible retention. Searches are made for the idea, not the language used to express it. Again, it's not about string matching. If it were then it would be a simple computerised process. Example of sources of non-patent disclosures - http://www.delphion.com/search-prior_art#tdb http://www.delphion.com/search-prior_art#tdb, http://www.questel.com/index.php/en/product-and-services/prior-art-and-alert http://www.questel.com/index.php/en/product-and-services/pri... [questel orbit], http://www.sciencedirect.com/ http://www.sciencedirect.com/. It used to be that the UK PO had console based access (eg telnet) to create search queries against these (and other) databases. Now of course the web is [often] used as the interface medium.