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."

Re:Oh please

[zeruch]

And I for one, have seen things that are certainly similar. At best what you are creating is a series of like values (I live (Engliah) = Eu vivo (Portuguese) = Iskun (Arabic), etc), and that is if you are doing translation (where such things have already been around). If it is for one language, then it is basically taking a "501 X Verbs" Book and making it searchable electronically, and adding it to the grammar/cpell check of a writing application. Unless there is some that extends beyond the simple idea of large tables of word/phrase data and maybe some kind of expert system with grammar rules that accounts for some of the varied iregular verbs of somelanguages, what you have is a rather bogus patent application.

More prior art

[Virtual_Raider]

In the spanish speaking world, unlike in english, there is an official academy of the language which monitors its development throughout all the spanish-speaking countries and updates the official Dictionary of the Academy accordingly. In their website they have a tool that does exactly the same as this patent describes. Would that count as prior-art or the fact that its in a different language might count as sufficient difference even though the process is about the same (if not more complex given that there are a lot more perks to spanish conjugation)?

Verbix

[Anonymous Coward]

According to Archive.org [archive.org], Verbix [verbix.com] has been around since at least March of 2000.

Re:Misleading headline....

[jorghis]

The abstract is general yes. But that is the abstract, the specifics are in the pages that follow. I think that this is where all the confusion on slashdot comes from. People read the abstract and assume that anything which is remotely similar to the abstract is what they are trying to patent. When in reality it really is just an abstract. You need to look at the entire application and realize that they are patenting a specific method of doing xyz, not just "a method for doing xyz" as is usually claimed in the abstract.

Re:Oh please

[Anonymous Coward]

Most dictionaries don't conjugate all of the verbs. They simply give the infinitive form and redirects for some oddballs. For example, I just checked my dictionary and it doesn't list the conjugations of the standard verb 'to utter.' It does list the nonstandard conjugations of 'to eat' (such as 'ate'). How about a Spanish dictionary? Hell no! None of mine even have tables in the back for conjugations. But they do have non-standard conjugations just like my English dictionary. If you want to learn the conjugations you buy a language book or you buy one of the 555 red books, not a dictionary. They are different books used for different purposes.

But this isn't really that surprising, a book on STL functions doesn't tell you all of the possible 'conjugations' of a function call either. You will find something like:

template<class InputIterator, class EqualityComparable> InputIterator find(InputIterator first, InputIterator last, const EqualityComparable& value);

The conjugation to "p = find(a.begin(), a.end(), N);" is up to you.

Re:Oh please

[martin-boundary]

Looks like you're willfully misunderstanding the point as well. There is nothing difficult about listing all the possible conjugations of a verb: It's trivial to do it by applying the algorithms expressed in a good grammatical reference.

It's trivial to do it for a fixed language, and it's trivial to iterate over any set of candidate languages with a well defined grammar, doing it for each.

The fact that a book doesn't list all possible forms for each possible verb in an explicit table is irrelevant. The book is enough to generate those forms on demand, which is all an algorithm is required to do.

Now, there are certainly optimal (smallest number of operations, or maybe smallest RAM requirements, etc) algorithms out there which perform equivalently to any given published grammar book, but finding those is at best a cause for buying the programmers a case of beer, it's not worthy of a patent. After all, it doesn't significantly advance the state of the art.

Re:Misleading headline....

[a_n_d_e_r_s]

Of course Microsoft has bullied programmers from releasing their code because it contains patents that Microsoft claims it owns. Yes, against small time people who cant afford the tens or houndred of thousands of money to get the patent revoked.

One highly publized example is VirtualDub which no longer support the .asf file format since Microsoft sent them a threat to stop VirtualDyb from using .asf files.

http://en.wikipedia.org/wiki/VirtualDub [wikipedia.org]

So yes Microsoft has no qualms about using their patents to stop open software being developed.

Re:Oh please

[caveymon]

Have you taken a look at http://www.verbix.com/ [verbix.com] ? Pretty nice program, with loads of languages. Input verb, output any possible conjugation form. Heck, I use the online conjugator all the time when I'm trying my best at the Finnish language.

Link to the actual implementation

[Anonymous Coward]

http://encarta.msn.com/encnet/homework/languagever bs.aspx [msn.com]

Re:Oh please

[Rock-n-Rolf]

While I was studying computational linguistics at http://www.linguistik.uni-erlangen.de/en_contents/ index.php [uni-erlangen.de] me and three other students implemented such a system in C on HP-UX within 6 weeks for German-English and English-German wordform translation during an internship. That was about 13 years ago, if I remember correctly. So really nothing new and extraordinary.

If somebody needs a reference for prior art, feel free to contact me.

Re:Yay, whatever

[DaphneDiane]

Jim Breen's WWWJDIC Japanese-English Dictionary Server [monash.edu.au] offers Japanese verb conjugation support. This sites been around for a while: main page from 1999 [archive.org]. I found mentions of conjugation support back as far as 2003-02-11 [archive.org].

A note on software patents

[Yumi Saotome]

The USPTO is obligated to publish patent applications 18 months after filing. Given their backlog of software patents, they usually don't address the merits of one until after 33 months or so after filing. That is why you might see some strange looking applications.

Re:Already been invented.

[Phisbut]

IANAL, thank God, but it seems to me that that would only be prior art if you had publicized it somehow. Prior art has to be public, for obvious reasons.

If you didn't publicize it, your prior invention only gives you the personal right to use your version of the technology without paying Microsoft. Until they sue you of course, then you'll either pay them or lawyers.

This piece of software [druide.com] has been for sale since 1996 (for French), and it does much more than what the patent covers (conjugate verbs), it's also a dictionnary with definitions (partly in the patent application for verbs), a thesaurus, a grammar, a spell and grammar checker (way better than what's embedded in MS-Word... it's a totally different league), and much much more. It's a must-have if you're even only remotely interrested in the French language.

Re:Already been invented.

[back_pages]

It has everything to do with this. This is an obvious patent to strike down unmercifully. The patent itself says that verb conjugation has been around online and that the only major difference is that this method *might* detect spelling errors (gee, like *that* hasn't been done) and would allow a person to imput their native verb to get conjugations in another language (not novel in the least). If the USPTO can't even pass basic common sense tests, why should they be allowed to issue anything that could lead to million dollar court battles?

This is not insightful.

The article is NOT in reference to a PATENT. There is no "patent to strike down unmercifully". The article describes a patent application. A PATENT APPLICATION PUBLICATION HAS NOT BEEN EXAMINED. All of your comments about the USPTO are literally, and according to the dictionary definition, BASELESS in this instance.

SLASHDOT IS THE FOX NEWS OF PATENTS.

Carry on.