Image via WikipediaFundamental problem of current
patent law, too many things are patented and people have been offering
mutually exclusive patents to multiple people for the same device or concept. The suggestion of the book is that some sort of automagic search system compares patents to see who has what. Sounds like a good idea and steps in this direction are being taken so perhaps all is well. There is a problem,
natural language processing. In order for patents to be acessible to the common man and indeed comprehensible at all patents in regular everyday lanuage must eb considered allowable. However, language is fluid and can describe a phenomenon or device in hundred's of ways. A Binary Metallic
Decision Making Device is just a complicated way of saying, flip a coin. The taxonomy of ideas that must be involved for a truly comprehensive and comprehensible patent system may well be impossible. Not to say it isn't worth trying but any time parsing a truly stupendous amount (terabytes if you include pictures, text etc.) of data in next to no time is the easy part of your task you have quite the difficult task. I am not certain the optimism of the book is warranted nor am I sure that technology is the appropriate solution. Perhaps a more precise measure of language in encoding the patents in the first place would be worthwhile. Natural language patents could be accepted but then a summary or abstract version would be encoded by a patent worker and then entered into the system. Unfortunately this solution would be wickedly expensive and time consuming but it is hard to imagine a system which isn't.
Image via WikipediaFundamental problem of current patent law, too many things are patented and people have been offering mutually exclusive patents to multiple people for the same device or concept. The suggestion of the book is that some sort of automagic search system compares patents to see who has what. Sounds like a good idea and steps in this direction are being taken so perhaps all is well. There is a problem, natural language processing. In order for patents to be acessible to the common man and indeed comprehensible at all patents in regular everyday lanuage must eb considered allowable. However, language is fluid and can describe a phenomenon or device in hundred's of ways. A Binary Metallic Decision Making Device is just a complicated way of saying, flip a coin. The taxonomy of ideas that must be involved for a truly comprehensive and comprehensible patent system may well be impossible. Not to say it isn't worth trying but any time parsing a truly stupendous amount (terabytes if you include pictures, text etc.) of data in next to no time is the easy part of your task you have quite the difficult task. I am not certain the optimism of the book is warranted nor am I sure that technology is the appropriate solution. Perhaps a more precise measure of language in encoding the patents in the first place would be worthwhile. Natural language patents could be accepted but then a summary or abstract version would be encoded by a patent worker and then entered into the system. Unfortunately this solution would be wickedly expensive and time consuming but it is hard to imagine a system which isn't.
Posts
No comments:
Post a Comment