Slashdot Mirror
Microsoft [to patent] Verb Conjugation
streepje writes "Here [to be] the latest egregious patent application. Microsoft [to be] [to apply] for a patent for [to conjugate] verbs. Future postings [to look] like this."
← Back to Stories (view on slashdot.org)
It seems that slashdot routinely posts headlines claiming "Microsoft patents X!" Where X is something obviously nonpatentable. However, in almost every instance what Microsoft has actually done is patented a specific method or system of performing X. This is no exception. Microsoft has not patented conjugating verbs. They are applying for a patent for a specific type of system which helps users identify verb forms from verbs and vice versa. Again: patenting a method or system for performing X != patenting X. Can we get an end to all these misleading "Microsoft patents smiley faces!" type of headlines?
Yes - but they are effectively patenting all methods of doing this. And that is the big problem. Amazon didn't patent one particular method of providing one-click shopping, they pretty much patented them all. As such, Microsoft will have a lock on anyone doing verb conjugation on a computer.
Nowhere in this patent do they describe the method in anything but the broadest generality - they are not patenting a specific implementation (which is what covers programs under copyright law).
As you imply - it's not unusual but it's still a bad idea to allow method patents like this.
You're kidding right? Their policy is to automatically grant every patent application, and let the courts figure out validity later. Basically, in order to show that they've reduced their budget, they fired all their patent analysts and let them work as consultants to civil courts at one hundred times the overall cost, once you factor in all the legal costs associated with resolving patent disputes the hard way. In a reasonable enlightened nation, this would get the government officials responsible for this decision horsewhipped in a public square before being exiled. In America, the people responsible were instead paid hundreds of thousands of dollars for their efforts and will live some of the cushiest lives in the entire world, while the tax payer grapples the massive extra costs introduced by this monstrous decision (as well as paying for the officials' pensions, rather than for a few bullwhips and an exile-barge at a fraction the price). Nice, huh?
That specific method here is "on a computer." This is exactly the type of patent that slashdot people get up in arms about. The patent application requests that they be the only ones allowed to conjugate verbs on a computer.
Though, I for one [to welcome] our new language [to own] overlords. (btw, way to go article submitter. you've made something dull into something interesting.)
//TODO: signature
Yes, that is true, but that doesn't make it any less straightforward and simple.
>"which is something I, for one, haven't seen before and think could be pretty useful..."
I, for one, have created a simple Perl-module which conjugates a given Latin verb in all tenses and forms. Let me tell you: conjugating a verb "on the fly" is trivial. Exceptions to every rule do, however, mess things up a little, but the exceptions themselves build up very simple and trivial rules.
Prior art? Hell, yeah!
Non-obvious? Hell, no!
Lemon curry???
You very likely don't work in natural language processing. People have been generating whole paradigms for a long time. For a set of published examples, check out the Xerox Finite State Morphology software and textbook. The software provides ways of describing the morphology and lexicon of a language and compiling it into an efficient finite state transducer. Once you've got the transducer, you can run it in either direction, that is, you can parse, or you can generate. A common test, and exercise in courses on doing this, is to generate the entire paradigm of a particular word or set of words.
Actually, there is a big change.
I have made a couple of inventions, which I did NOT want patents for. I want the general public to benefit from them (besides, filing for a patent is too expensive for my meagre budget).
Now Microsoft (or another evil big company) reads about my research, and files for a patent. The consequence is that they will get a patent for my work, which I did not allow them to get. And the main reason is that there is NO WAY to apply for NOT getting a patent. The only thing I can do is to publish my invention, and hope that it takes Microsoft more than one year to discover that publication. One year after first publication a patent cannot be applied for anymore, so that would make my invention safe.
It happens quite often that I present research at conferences, and someone in the audience gets up and asks with a gleam in his eye, "Did you apply for a patent yet?" I know what that guy is thinking.
It doesn't matter whether other systems have had on-the-fly verb conjugation. It only matters if they used the same implementation as described in this patent. If the MS implementation is new, then it's arguably patentable. Most here seem to intentionally misunderstand that.
-- "I never gave these stories much credence." - HAL 9000