Does this analyzer server meet the “full, clear, concise” terms of 35 USC 112?

Question

Could you explain me how the 35 USC 112 rule is satisfied by the patent US5946647:

The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.

The US5946647 patent specification defines an analyzer server. The analyzing server consists of a parser and a grammar file. The parser retrieves a grammar from the grammar file and parses text using the retrieved grammar.

[00030] Referring now to FIG. 3, a block diagram illustrating an analyzer server 220 is shown. In this figure, analyzer server 220 is described as having a parser 310 and a grammar file 320, although alternatively or additionally a fast string search function or other function can be used. Parser 310 retrieves a grammar from grammar file 320 and parses text using the retrieved grammar. Upon identification of a structure in the text, parser 310 links the actions associated with the grammar to the identified structure. More particularly, parser 310 retrieves from grammar file 320 pointers attached to the grammar and attaches the same pointers to the identified structure. These pointers direct the system to the associated actions contained in associated actions file 330. Thus, upon selection of the identified structure, user interface 240 can locate the linked actions.

[00031] FIG. 4 illustrates an example of an analyzer server 220, which includes grammars 410 and a string library 420 such as a dictionary, each with associated actions. One of the grammars 410 is a telephone number grammar with associated actions for dialing a number identified by the telephone number grammar or placing the number in an electronic telephone book. Analyzer server 220 also includes grammars for post-office addresses, e-mail addresses and dates, and a string library 420 containing important names. When analyzer server 220 identifies an address using the "e-mail address" grammar, actions for sending e-mail to the identified address and putting the identified address in an e-mail address book are linked to the address.

I am a normal software engineer. And, after reading the patent, I do not know how to make the parser, because there are many possible implementations of the parser, and all of them have drawbacks. For example, to extract an e-mail from the text I can use one from many of regular expressions. I will have to strongly consider how to implement the parser and how to define the grammar file. Also, I do not know, which is the best mode contemplated by the inventor of carrying out the parser.

Please, help me understand how the patent law works?

The US5946647 is used by Apple against HTC and Samsung.

Answer 1

Applying the criteria that you quote from section 112, the question is not whether you can build the same parser element but whether you can build the invention - perhaps using any one of several parsers to satisfy the parser element of the invention. Not every element of a claimed invention has to be defined in excruciating detail. The question is whether one of ordinary skill in the art can build the invention based on the information set out in the patent - Figures, Specification and Claims all taken together.

It is not at all uncommon that some claim elements can be satisfied in multiple ways. If one of ordinary skill in the art can choose among several options - each of which permits assembling the claimed invention - then the inventor has given a full description of that claim element.

Answer 2

I am a normal software engineer. And, after reading the patent, I do not know how to make the parser, because there are many possible implementations of the parser, and all of them have drawbacks. ... I will have to strongly consider how to implement the parser and how to define the grammar file. Also, I do not know, which is the best mode contemplated by the inventor of carrying out the parser.

As user96 said, you've already stated you do, in fact, know not just one but several possible ways to implement the parser, you just don't know which of the several implementations would be best for your purposes. The general rule is that the invention has to be enabled such that a person having ordinary skill in the art (you, it sounds like) could create the claimed invention using your basic knowledge and skills combined with the teachings of the specification and without resorting to "undue experimentation."

Without more information, it's hard to give a solid answer, but it doesn't sound like what you're talking about, e.g. having to "strongly consider" how to implement the parser or define the grammar file, would rise to the level of "undue experimentation."

Also, if the inventor has a "best mode" in mind when they file, they are supposed to disclose it, but they are not obligated to have actually contemplated a best mode prior to filing.

Answer 3

To hopefully further expound on what user96 said, this patent sounds a whole lot like the Interpreter Pattern, but in patent-speak. One or more regular expressions would certainly constitute an appropriate set of grammars for certain actions (e.g., call a phone number), and for more complex grammatical rules, you could use a tool (ANTLR, Bison, Yacc, etc) to generate some code.

I note that the filing date is 1996; the Interpreter Pattern was included in the Gang of Four's Design Patterns book in 1995, and must have been in general use before that date to have been included in the book. The diagrams also seem to be of ... well, computers. The fourth diagram of Bob and Alice IM-ing each other is nifty, but it's still probably just running the input through an implementation of the Interpreter Pattern. Am I overlooking something about this patent and/or these claims that was truly novel in 1996?