JP2002521751A - Modular language translation system - Google Patents

Abstract

(57) [Abstract] Any one user in various different user interfaces sends a translation request to any one of various different translation engines and receives a response from any of the above translation engines. A modular language translation system that allows you to. System users who are accustomed to using the user interface of the first type of translation system, such as the MT system for translating from Japanese to English, translate from the translation engine of the second type of translation system that translates Russian into English The user interface can be used to obtain In this case, there is no need to learn about the particularity of the second system and its interface. The user interface and the translation engine communicate via a distributed object protocol.

Description

DETAILED DESCRIPTION OF THE INVENTION

TECHNICAL FIELD The present invention relates to an automatic natural language translation system in which different user interface modules can interact with different translation engine modules. For example, one of the user interface modules resides on a computer and any of the various available translation engine modules reside on one or more other computers connected by a network. -Modules can be called.

BACKGROUND OF THE INVENTION [0002] Machine translation (hereinafter "MT") usually translates text into some natural (colloquial) language.
Computers that help or automate the translation of languages into other natural languages
Refers to the field of application. The idea of deciphering natural language using mathematical techniques became a reality during the Cold War period after World War II. Much of the early MT system and research has evolved because the intelligence community needed it. Early MT systems typically used verbatim translations (word-to-word translations) without using language rules.

[0003] As the use and increasing importance of computer systems connected by the Internet and other networks, the partial demand for MT systems remains high. Users around the world can exchange information at any time, with speed and convenience previously not possible. However, this information must often be translated into a natural language that is freely available to the user. In particular,
Business users often need to translate documents.

[0004] Typically, current MT systems have different user interfaces depending on which source language is translated into which object language. Therefore, it is difficult for the user to learn the interface or to quickly or automatically translate the document into one or more object languages.

For example, some MT systems can support translation options, such as “spell checking,” which automatically scans documents to check for spelling errors before translating. On the other hand, another MT system cannot support such an option. Thus, the user must make a trade-off between all requests to translate a document into a particular option and the requirements of the particular MT system in use.
You have to come to terms. Since each MT system can provide many different functions and options, the user must learn and remember which MT system can perform a particular option. Furthermore, different MT systems often refer to the same concept in different terms, making it difficult for the user to set up a "standard" translation request. As described above, since the standardization is not performed, it is difficult for the user to translate various documents.

SUMMARY OF THE INVENTION The present invention is directed to a method wherein a user of any of a variety of different user interfaces sends a translation request to any of a variety of different translation engines. Provide a translation system that allows to receive responses from any translation engine. That is, MT that translates from Japanese to English
Users of the present invention who are accustomed to using a user interface for an MT system of the first type, such as a system, will be able to translate Russian into English.
The user interface can be used to obtain translations from the T system translation engine. In this case, there is no need to learn about the specifics of the second system and its interface. Further, the present invention allows software developers to easily create applications that can communicate with many different MT systems. Furthermore, using the present invention, a particular MT system can handle translation requests from any of a number of other MT systems.

In certain embodiments of the disclosure, the present invention provides a component object model (COM) or common object request broker architecture (CORBA) for all communication between a client and a translation engine. Use a distributed object protocol such as The client and the translation engine each have an interface defined by an interface definition language (IDL). All translation requests sent by the client and all responses generated by the translation engine are sent according to the IDL that defines these interfaces. Therefore, the client interface and translation engine
Whatever machine the translation engine is using (as long as the machine is connected, such as over a network)
No matter what operating system the machine is using (as long as the protocol is supported) or in the computer language in which the client or translation engine is written (as long as the language system supports the distributed object protocol). Regardless, communication can be performed.

[0008] By using a distributed object protocol to standardize the way the user interface communicates with the translation engine, as proposed by the present invention, developers have developed a "modular" type MT system. Can be designed. This is convenient for MT system developers and users. A user who buys and wants to master an MT system for one type of language, such as English-Spanish, can use a completely different MT system, such as English-French, as needed, without having to relearn a new system. An interface from the system can be used to communicate with the system.

Generally, from a certain point of view, the present invention features a translation system having a client and a translation engine. The client sends a translation request containing the text to be translated from a first language to a second language using a distributed object protocol. The client receives a response to the request using a distributed object protocol. The translation engine receives the translation request using a distributed object protocol, creates a response corresponding to the translation of the text from the first language to the second language, and uses the distributed object protocol. And send a response to the client.

[0010] Such embodiments of the present invention may include the following functions. Translation requests and responses can be sent according to the IDL defining the interface at the client and the interface at the translation engine.
To communicate, the distributed object protocol used by clients and translation engines can support multiple user interfaces. The distributed object protocol uses the component object model (COM
) Standard or Common Object Request Broker Architecture (CO
BA) Operates according to standards. The system can have a notification mechanism that includes an interface defined by IDL by which the translation engine informs the client of errors that occur during translation. The system allows the translation engine to register itself for the client to find,
For components like user dictionaries that can be obtained with the translation engine,
It can include a registration mechanism with an interface defined by IDL that can inform clients. The client and the translation engine can be located on the same computer or on different computers interconnected. For example, a client and a translation engine can communicate over a computer network. The text to translate can be represented by a set of characters that can include characters used in many languages, such as UNICODE.

In general, from another perspective, the invention features a translation method whereby a translation request is sent from a client to a translation engine via a distributed object protocol. The translation request includes text translated from a first language to a second language. At the translation engine, the request is received by the distributed object protocol, and translation is performed according to the received request. The translation is included in the response to the request, the response being a distributed object
It is sent from the translation engine to the client according to the protocol.

[0012] Such embodiments of the invention may include the following functions. For example, the step of transmitting a translation request from a client includes a client interface,
And transmitting a translation request from the client to the translation engine in accordance with the IDL defining the interface of the translation engine. The translation method may include providing a registration mechanism by which the translation engine may register itself for the client to discover. The registration mechanism has an interface defined by IDL. The translation method may include scanning the translation request to check for errors in the text, such as spelling errors or undefined phrases, before sending them to the translation engine for translation. The translation method may include providing the client with information regarding the progress of the translation by the translation engine.

In general, and in yet another respect, the invention features a translation system comprising a plurality of translation engines and a client, or a translation engine and a plurality of clients. In each case, one unit is
It can interoperate with any one of a number of units, and these units are distributed by a distributed object protocol such as COM or CORBA.
Can communicate. In the case of the first embodiment, one or more clients may be included, and in the case of the second embodiment, one or more translation engines may also be included. In certain embodiments, one or more of each client and one or more of each translation engine are
Has an interface defined by Therefore, all communication between the client and the translation engine is performed by IDL. In a specific example,
The translation engine can maintain a separate context for each translation request it receives, so that the translation engine translates each request according to the particular translation format that may be associated with that translation request. can do.

The above and other aspects, features and advantages of the present invention will become more apparent from the following detailed description when read in conjunction with the accompanying drawings.

Throughout the drawings, similar reference numbers designate the same parts. In addition, the drawings are not to scale, and generally focus on describing the principles of the invention.

DESCRIPTION OF THE INVENTION Referring to FIGS. 1 and 2, in one specific example, a language translation system 10 of the present invention includes one client 11 and one translation engine 16.
The client 11 uses a distributed object protocol to translate the translation engine 16
May be any type of system (e.g., a user interface) that can send translation requests to and receive responses from the translation engine 16 via a distributed object protocol. (The distributed object protocol will be described in more detail below.) For example, the client 11
World Wide Web (WWW) such as Gopher, File Transfer Protocol (ftp), Hypertext Transfer Protocol (http), Simple Network Management Protocol (SNMP) and Simple Mail Transfer Protocol (SMTP)
) May be a proxy server that operates according to a protocol that invokes the translation engine 16 in order to instantly translate content downloaded from a computer network. In another example, the client 11 may be a "web crawler" or "web robot" that searches the WWW for content and generally translates the content into a common language. Good. Client 11
The whole can be configured by a hard-wired logic circuit, but preferably,
It is preferably implemented with software running primarily on a general purpose computer such as a personal computer or workstation.

In one embodiment, the client is a UNICODE character set, an international character encoding standard that allows computers in one world language community to “converse” with computers in other language communities. Use character sets that include multiple characters used in multiple languages. For example, one example client is a word processing application running on a computer. Alternatively, other software applications, such as spreadsheets, database programs, web browsers, etc., can be clients 11. The client 11 has a client interface 94 specified by the distributed object protocol. Preferably, the client interface 94 is defined in an interface definition language (IDL), although those skilled in the art will be able to use other mechanisms to define the client interface 94. Will be able to understand. Software developers work together through this interface. Further, in another embodiment, the client may include a user that allows a user of the application to interact with the application.
It has an interface 24. The user interface 24 of the application or program can include a translation service that accepts translation requests. For example, in the case of a script-enabled application such as Microsoft Word 7.0, the word document may include a visual basic script that can send translation requests over a distributed object protocol. .

In one embodiment, the translation request includes text for translating from one natural language to another. Alternatively, the translation request may also include information that tells the translation engine how to translate, information about the particular text to be translated, information about the order of translation, and information about the particular dictionary to be used during translation. it can. For example, the information included in the translation request may include: part of speech settings (indicating a particular part of speech of one or more phrases in the text); translation hints; sentence boundary indicators;
Translation guidelines; indications of the type of phrase to translate in a particular way; or indicators of breaks in the text. The information is further provided by a suitable dictionary,
It may include information on the priority of the required translations, indicating different phrase settings, different translations, and different sentence translations, as well as indications for portions of the text to which the options apply. These examples of information do not limit the invention. One skilled in the art will appreciate that it may be desirable to include the priorities and options of many different types of translations in a translation request.
Information can also be stored in Hypertext Markup Language (“HTML”), and Standard Generalized Markup Language (“SGML”) markup; Rich Text Format (“RTF”) markup; Markup may also be included, such as run-off (“NROFF”) markup.

In one embodiment, a system other than a general-purpose computer is used to execute the client 11, including a palm-sized computer, a personal organizer, a mainframe computer, and a fax machine. Any system capable of communicating with a distributed object protocol is within the scope of the present invention. For certain embodiments, preferably, the distributed object protocol of the present invention preferably supports multiple user interfaces, computers, applications, and the like. Also, preferably, the IDL defines an interface of each system used to execute the client 11.

Referring to the specific example of FIGS. 1 to 3, the translation engine 16 includes a system capable of receiving a translation request by a distributed object protocol, and is preferably a general-purpose computer like the client 11. It is preferably performed primarily by software running on it. Also, like the client 11, the translation engine 16 of this embodiment includes a translation engine module interface 96 that is specified to follow a distributed object protocol. Preferably, translation engine interface 96 is defined by IDL. However, as one skilled in the art will appreciate, other mechanisms may be used to define the translation engine interface 96. However, the translation engine 16 and the client 11 may each communicate via a distributed object protocol, and any machine that may include an interface specified to conform to the distributed object protocol.
It should be understood that it can also be performed on a device or product. Translation engine 16
Translates the source text using the data in the storage device 18. Typically, the data in storage device 18 corresponds to all types of information that is used or useful in performing translations, including one or more dictionary domain keywords, grammar rules, other databases, and the like. . The translation engine 16 can be located on the same computer as the client 11 or on a separate computer in communication with the client 11.

After translation engine 16 receives the translation request from client 11, translation engine 16 translates the text from a first language to a second language. Thereafter, translation engine 16 uses the translated text to create a response to send back to client 11. The translation engine 16 sends the response to the client 11 using the distributed object protocol.

Referring to FIG. 2, which illustrates a particular embodiment of the present invention, translation system 10 also includes
Preferably, it includes a registry 84 by which the client 11 can find the location of the translation engine 16 that can be used. For example, client 11
If he wants to send a translation request, the client 11 can check the registration 84 to find a translation engine that he can access. In one embodiment, this check includes a search for a translation engine 16 located at the client 16. In another embodiment, the registry 84 is used to register other translation engines 16 operating according to the distributed object protocol.
Search a computer network, such as WWW, to find. In yet another embodiment, the registry 84 provides information regarding options and / or components that can be used on each translation engine 16 found.
Typically, these components include one or more of the following: That is, a user dictionary, a syntax dictionary, a user dictionary editor, a dictionary browser, a second translation engine, another phrase finding device, another sentence translating device, a sentence boundary finding device, an asynchronous (background) translating device, and a synchronous Translator, fuzzy translation alignment, part-of-speech settings, selective translation, spell checker, and translation memory archive.

The translation engine 16 converts it into one or more translation options 86,
Can be related. Translation options 86 correspond to “global” options that apply to any translation requests sent to translation engine 16. For example, translation option 86 may be an operation that scans for undefined phrases. Therefore, for any translation request that the client 11 sends to the translation engine 16 including this option, the translation engine scans the text of the translation request for undefined phrases. Further, a registration 84 that includes a list of translation engines 16 having this option may indicate as part of registration 84 that the translation option is available at this translation engine 16.

The translation request sent by client 11 to translation engine 16 typically includes at least one type of text that translates from a first natural language to a second natural language. For example, the translation request may include some English text that translates into Japanese. It should be understood that the invention is not limited to any particular set of languages. As described above, the translation request may include information other than the text to be translated, such as information instructing the translation engine 16 to perform the translation in a particular manner or in a particular format. For example, the translation request is sent to the translation engine 16,
A translation priority may be included that instructs a particular user dictionary 38 to translate. It may be desirable to use a particular user dictionary 38 if the text to translate relates to a particular area (such as technology or finance). The use of technical or financial term dictionaries helps to ensure that translations with specific meanings in these areas are accurate.

The above information that directs the translation engine 16 to translate in a particular manner, commonly referred to as translation priority, can be processed by the translation engine 16 in several ways. In one embodiment, the translation engine uses one or more translation priority objects 88 that respond to information in the translation request relating to the format required for translation. In another embodiment, translation system 10 may further include a priority editor 26 by which a client may provide translation engine 16 as a translation priority object.

A translation request can be created in a number of well-known ways. FIG. 3 can provide text to be translated, input interface 12, display 20,
1 illustrates one embodiment of the present invention for use with a system that includes a client 11 having an output interface 14 and a user input device 22. In certain embodiments, user input interface 12 may be a keyboard,
It may be a mouse, a touch screen, or a light pen. It should be understood that any user input device that can be used by client 11 can be used as user input interface 12.
For example, in one embodiment, input interface 12 may be connected to a source capable of providing text to be translated, such as a modem, serial line, web page, or other system communicating with client 11. Including the connection. In certain embodiments, display 20 includes a computer display or printer. Any device or system that can send information to the client 11 can be used as the display. Output interface 14 sends the final translation of the source text in a second language. In this case, the translation is from the response sent by translation engine 16. Output interface 14 may include an electronic interface such as a printer, fax, voice interface, modem or serial line, or may include other means for transmitting text to an end user.

The client 11 and the translation engine 16 can be interconnected in any number of well-known ways so that the client 11 can send a translation request to the translation engine 16. For example, in one embodiment, client 11 and translation engine 16 are located on the same computer system. In another embodiment, client 11 and translation engine 16 are located on different computers that can communicate with each other. In another embodiment, the communication occurs over a computer network, such as the Internet. In yet another embodiment, the communication is a coaxial cable, such as a cable TV line,
It is performed by a wireless communication system through a high-speed communication link or a telephone line.

Further, in another embodiment, the client 11 and the translation engine 16 share a thin (narrowly defined) interface. In this case, the thin interface is flexible enough for client 11 and translation engine 16 to communicate, but supports only a limited number of application program interfaces (APIs).

Accordingly, the distributed object protocol in which the client 11 and the translation engine 16 communicate regulates the setting of communication between the object and the distributed object (an object that communicates with an object on another network node). It is a standard to do. Distributed object computing is the development and maintenance of systems by generating reusable code (usually called objects) and by allowing objects located on different computers to communicate over a network. Make it easy. A software object is a collection of related functions (or artificial brains) and related states of functions (or artificial brains). A set of distributed objects, such as a set of all objects in FIG. 2 (client 11, translation engine 16, dictionary browser 46, etc.) is a collection of independent operating nodes. In one embodiment, each node includes an independent process. In another embodiment, each node includes independent storage. However, it should be understood that a node need not include an independent process that can be used with the present invention. For certain embodiments, each node may be located on a separate and independent computer, may be located on the same computer as all other nodes, or It may be distributed between two or more computers.

A number of different distributed object protocols are currently available and can be used with the present invention. These distributed object protocols include, for example, Microsoft's Component Object Model (COM), Microsoft's Distributed Component Object Model (DCOM), Visual Work's Distributed Small Talk, JAVA® RML and For example, the Common Object Request Broker Architecture (CORB A), which includes the Internet Inter Orb Protocol (IIOP) of the Object Management Group. For the disclosed embodiment, COM and CORBA are the preferred distributed object protocols. COM is currently the de facto standard for building component architectures.

Under the distributed object protocol, objects on one node know the names of objects on other nodes. Handles to distributed objects (also called interface pointers or object references) can be moved between machines, between programs, and between processes. Any entity that has access to the handle to the distributed object is
You can call the method on it. (The term invocation method is synonymous with sending a message. Therefore, an object on one node can send a message directly to an object on another node, and an object on one node cannot One way to ensure that objects operating according to the distributed purpose protocol communicate with each other is to specify them on the interface on each object using the Interface Definition Language (IDL). That is, if the interface is specified by the IDL, the object itself can be written by different vendors in different programming languages or as different versions of the same program. Can communicate with each other.

Therefore, one advantage of the system of the present invention, as defined in IDL, using a distributed object protocol, is that it is interoperable. Clients and translation engines developed according to the present invention should be treated as interoperable software modules, unlike incompatible, dedicated systems, such as those used in conventional language translation systems. Can be. Further, the client and translation engine developed according to the present invention are neutral in terms of language, merchant, and version. Thus, by using the present invention, software developers can work with multiple providers of translation services in a transparent manner, ie, professionally as to which provider or implementation is currently in use. You can design without any knowledge. By using the present invention, translation service providers will be able to
New, enhanced, or unique translation components can be exposed to potential consumers in a standard way, without having to worry about compatibility with the consumer's system.

A software developer who wants to be able to develop a user interface that can use multiple translation engines can use any one of a number of applications to any of a number of printers. It's somewhat like a PC user who wants to be able to put a document in one. The above user
The application maker and / or printer maker does not want to know the various names and options for the print options. To make this task easier, developers of desktop applications, such as word processors and spreadsheets, have tasks that fit the needs of a particular printer.
Established a standard "printer driver" that is transparent to the user. The user
The user simply needs to have knowledge of the interface of the particular printer of the application being used. The user simply needs to display the standard "Print" menu from the application interface, select from a set of standard print options (such as "Duplex"), and select the required printer. The user can confidently configure the default options and / or the default printer with the knowledge that the application and printer will achieve the required functionality or will be notified if the functionality is not available. You can choose. Thus, users of multiple applications can interoperate with multiple printers. This is because printers and applications function as objects defined by standard protocols. By requiring that the translation engine and the client be defined in a standard such as a distributed object protocol, the present invention allows the client and the translation engine to work together.

Referring to FIG. 2, this illustrates some of the options and features that certain embodiments of the translation system 10 of the present invention can have. The translation system 10 includes a client 11 having a user interface 24, a translation priority server 28, a synchronous translation device 30, an asynchronous translation device 32, another phrase server 34, a translation server 36, and a user dictionary editor 44. , Dictionary browser 46, priority editor 26, user dictionary 38, syntax dictionary 40, basic dictionary 42
, Part of speech registration 90, registration 84, another sentence server 92, translation option 86, and translation priority object 88.

The user interface 24 has a translation service for accepting a translation request. In one embodiment, the user interface 24 includes a standard set of fields in which the client 11 provides text to translate and / or translation options. The user interface 24 need not be located on the same computer as the client. The user interface 24 may be a site, such as a site accessible by a web browser,
It can be located at a remote location where the client 11 can be accessed.

In the embodiment of the present invention shown in FIG. 2, the user interface 24 is the same regardless of what translation request the client 11 wants to request.
That is, whatever the application or system, the client is provided with the same user interface 24. The user interface 24 allows the client to specify options that affect the translation performed by the translation engine 16. Further, the client 11 using the user interface 24 may use the user interface 24 to inform the user whether the translation engine 16 supports a particular translation priority, such as spell checking. it can. The client 11 sends a message to the priority editor 26, which checks with the translation priority server 28 to check whether the selected translation priority is available at the translation engine 16. connect. Preferably, priority editor 2
6 preferably has an interface defined by IDL. Thereafter, the priority editor 26 informs the client 11 whether the translation priority can be used, so that the client 11 can formulate a translation request with this information, , This information can be sent to the user interface 24. If the translation engine 16 receives the translation request sent by the client 11 according to the translation priority, the translation server 36 receives the priority information and ensures that the translation engine 16 translates accordingly. I do.

As an example of the use of priorities, consider an example of a request for a particular user dictionary. When the client 11 interacts with the priority editor 26, the priority editor 26 displays on the client 11 only a user dictionary 38 that can be used by the translation engine 26. If the client 11 selects a particular user dictionary 38, the priority editor 26 will ensure that the client 11
Sends a translation request to the translation engine so that the translation priority object 88 reflects the preferred dictionary selected by the client 11. In one embodiment, the preferred dictionary is used until the client 11 resets the priority. Therefore, translation priorities can work the same for priorities used in other applications, such as web browsers.

As an example of using options, consider the example of a spell checker. If the translator 16 cannot use spell checking and the user interface 24 is displayed to the user of the system, the user cannot select it as an option. Spell check options are displayed in such a way as to indicate that For example, the selectable box for spell checking on the user interface 24 changes to "gray" to indicate to the user that the user cannot select the option. However, if spell checking can be performed at the translation engine 16, the user interface 24 allows the user to select it as an option. In this way, the client 1
1 and the user interface 24 can interact with the translation engine 16 having different types of options. This type of function is also registered 84
Can be used with the items shown in

The embodiment of FIG. 2 also provides other features that are useful to users seeking translation of the text. The synchronous translation device 30 is used when translation needs are simple.
Preferably, the synchronous translator is implemented such that it has an interface defined by IDL. In this case, the translation is performed immediately.
When the client 11 sends a translation request to use the synchronous translation device 30, the client waits for the translation engine 16 to send a response containing a translation of the text.

For example, an end-user application, such as a word processing application, can include a script that requests an immediate translation of a portion of the text. Such a script (usually a visual basic script) calls the synchronous translator 30. Those skilled in the art will appreciate that synchronous translator 30 can be used to write many other programs in IDL.

In contrast, if the client 11 sends a translation request to use the asynchronous translation device 32, the translation engine 16 performs the translation in the background, and as a result, the client 11 Other work can be performed while 16 is performing the translation. While the client 11 is doing other work, the translation engine 16 provides information about its progress to
Keeps you informed of the progress of the translation. The translation engine 16 may send this information directly to the client 11, or the client 11 may poll the translation engine 16 for this information.

When the client 11 sends a translation request to the asynchronous translation device 32, the client 11 similarly supplies the text to be translated and a place to write the result to the asynchronous translation device 32. The client 11 provides two additional interfaces for this purpose: a translation source interface 96 and a translation sink interface 100. Each of these interfaces is specified by IDL. The translation source interface 96 supplies the text, and the translation sink interface 100 consumes the translation (the results are written there). In other embodiments, the translation source interface 96 may also specify the use designation portion of part of speech settings, annotations such as HTML markup, and the like. In yet another embodiment, translation sink interface 100 not only retrieves translated text, but also provides part of speech information and dictionary source information from translation engine 16.

Further, when the asynchronous translation device 32 is used, the client 11
Similarly, you may want to know how to get progress information, or you may want to know when the method is performed. In one embodiment, the client can provide an associated translation progress object 102 to the asynchronous translation device 32 and receive a progress notification. The translation progress object 102 does not have an interface defined by IDL. In another embodiment, the client 11 can instead poll the asynchronous translator 32 to obtain progress information.

Other functions of the specific example of FIG. 2 include the following. That is, the end-of-sentence detection device 80 having the interface defined by IDL tries to find a break between sentences in the text to be translated without performing translation. The error processing device 82 notifies the client 11 whether or not an error has occurred during translation, during reception of a translation request, and during reception of priority information and the like. Preferably, error handling unit 82 is implemented using a callback interface defined by IDL to specify errors, provide severity information, and provide numeric error identification. In another embodiment, the error callback object is the asynchronous translator 32
, Client 11 or other objects operating according to the distributed object protocol, such as the user interface 24.

In certain embodiments, the translation engine 16 responds to the client 11 with information about the format required for translation, and in some of these embodiments, Notification that the required format is not available. In certain embodiments, translation system 10 further locates at least one alternative phrase selection for a given translation request and causes client 11 and translation engine 16 to select this alternative phrase selection. Is provided with another phrase server 34. In certain embodiments, the system translates Japanese into English or English into Japanese. In certain embodiments, the translation engine includes a translation server to receive information related to the format required for translation. As one of ordinary skill in the art will appreciate, many other types of custom dictionaries, spell checkers, background (asynchronous) translators, immediate (synchronous) translators, error checkers, etc. A translation system function may also be included.

Referring to FIG. 3, in another specific example, the translation engine further comprises:
A storage device 18 that can include one or more disk areas (eg, a hard disk, a floppy disk, an optical disk, etc.), and / or a memory (eg, RAM) storage device. Etc. are provided. In certain embodiments, storage device 18 stores input text information in the source natural language, output text information in the target natural language, and one or more dictionaries, domain keywords, grammar rules, and other databases. Used when performing a translation, including or used to store all kinds of useful information.

FIGS. 4 and 5 illustrate two embodiments of the present invention that take advantage of the interoperability provided by the distributed object protocol used in the present invention. 4 and 5
, It is to be understood that all clients 11 and translation engine 16 operate according to the distributed object protocol.

FIG. 4 is an embodiment of the present invention having a client 11 and a translation engine 16. Each translation engine 16 can receive a different translation request from the client according to the distributed object protocol, and can create a response corresponding to the translation of the text of the translation request. The response is sent to the client by a distributed object protocol.

Therefore, for example, in the case of FIG. 4, the user of the client 11 has a predetermined portion of the English text and intends to translate it into Japanese. Users also want to be able to communicate the accuracy of translation. The user inputs the text, and the client 11 formulates a translation request to be transmitted to one of the translation engines 16. In this case, the translation request is sent to a translation engine 16 that performs English-Japanese (English-Japanese) translation. Because the system can access multiple translation engines 16 simultaneously, the system can translate the response created by the English-Japanese translation engine into other translation requests sent to the Japanese-English ("Japanese-English") translation engine. Can be sent as Thereafter, both the English-Japanese translation engine and the Japanese-English translation engine return a response to the client 11. The user at the client 11 compares the response provided by the Japanese-English translation engine (the translated text (Japanese) is translated into English again) with the original English text to confirm the accuracy of the English-Japanese translation device. can do. Such a technique is useful when the user detects a small translation error that occurs when the translation engine cannot recognize a phrase in a particular language. Regardless of the nature of a particular translation engine, being able to access multiple translations simultaneously is very useful, as these translation mistakes can result in translations that have a different meaning than the original.

Whatever the source, the ability of this embodiment of the present invention to remotely host multiple translation engines also has other advantages. For example, in one embodiment, the client 11 may select the entire user translation engine (often tens or more megabytes or more) without having to download it at a high cost. Can be used to test a remote translation engine 16.

In another example, as shown in FIG. 4, the user of client 11 has a predetermined portion of the English text. The user can copy this text, for example,
They want to translate into multiple languages, such as Japanese, Spanish, and German. Client 11 only wants to formulate a translation request. The client can then send this same translation request to any one (or more) of translation engine 16. In this case, there is no need to customize the translation request for each translation engine 16. Translation engine 1 that receives the above translation request
6 sends a response corresponding to the translation of the text to the client 11. Thus, the client 11 can receive, by one translation request, translations of the predefined portions of English text from Japanese, Spanish and German from different translation engines 16.

Such an embodiment of the present invention includes a user interface having a translation service for accepting translation requests, and a distributed object in which the client and the plurality of translation engines communicate to support the plurality of translation engines. -Can include clients with protocols. In another embodiment, the distributed object protocol used by multiple translation engines and clients to communicate operates according to the COM standard. In another embodiment, the distributed object protocol used by multiple translation engines and clients to communicate operates according to the DCOM standard.

FIG. 5 is an embodiment of the present invention having a plurality of clients 11 and one translation engine 16. In this embodiment, the translation engine 16 is located far away from one or more of the clients 11. Any one or more of the clients 11 can send a translation request to the translation engine 16 and receive a response from the translation engine corresponding to the text of the translated message.

This embodiment is particularly useful when the client 11 is running on a user computer that does not have enough memory or processing power to install the translation engine 16 thereon. This embodiment is also useful for multiple "on-the-fly" translations requested by multiple remote clients 11. For example, several employees of a company may each carry a palm-sized computer, personal digital assistant (PDA), or the like, on a business trip. Users of such systems can communicate with remote computers via satellite and cellular connections. According to this embodiment of the present invention, each user's portable system can be a client 11 that communicates with the same remotely located translation engine 16 and sends the translation engine 16 a plurality of translation requests. Each client 11 communicates according to the distributed object protocol, and the translation engine 16 sends a response according to the distributed object protocol, so that each client 11 (eg, PDA, laptop,
Whatever the type of each client 11 (such as a palm-sized computer), a translation can be obtained from a translation engine.

Those skilled in the art will appreciate that all embodiments that can be applied to various types of the present invention, including one client and one translation engine, are also equivalent to multiple clients and / or multiple translation engines. It can be understood that the present invention can be applied to embodiments of the present invention. As those skilled in the art will appreciate, many other types of translation system features, such as custom dictionaries, translation options, spell checkers, etc., may also be included.

FIG. 6 illustrates steps performed in a specific embodiment of the method disclosure of the present invention. The above steps will be described using an example of an English-Japanese translator. It should be understood that the invention is not limited to any particular language. Referring to step 48, the client receives input text in a first language. This text may be from a word processing document, voice input, or a string entered by a user operating keyboard. For example, the following statement can be entered into the client:

"Do you speak English?" Next, in steps 50 and 52, the client creates a translation request from the input text and sends it to the translation engine via a distributed object protocol. In one embodiment, the translation method includes providing the translation engine with information about the format required for translation, and in one embodiment, the translation method includes: Informing, according to the information, whether a translation cannot be performed. Therefore, step 52 may further include the steps shown in FIG. 7, if desired. In this example, as shown in step 64, the above text string is first scanned to check for mistakes (such as typos). The source text may also include annotations scanned by the translation engine, HTML markup, sentence punctuation markup, part of speech settings, and the like. Turning now to step 66, the user can select the required translation format. In this case, the required translation format is to select a Japanese translation that will be the English alphabet (Roman) instead of the Kanji. Turning now to steps 68 and 70, the translation priority object is sent to the translation engine to inform the translation engine of the configured attributes of the translation. In one embodiment, the translation priority object is sent as part of a translation request. In another embodiment, the translation priority object can be sent independently of the translation request.

Referring again to FIG. 6, in step 54, the translation engine receives a translation request according to the distributed object protocol. Referring to step 56, the translation engine then translates the translation request text into a second language according to any selected available translation options and according to any selected translation priority. In certain embodiments, the method of the present invention further comprises:
The method may include notifying the client if an error occurred during any of the steps. Those skilled in the art will be able to use additional standard translation steps, such as translation with a particular dictionary, translation while performing other tasks of the client, translation with a particular format, etc. with this feature of the invention. Will be able to understand. So, for example, if you want to,
6 can include the steps shown in FIG. Referring to step 76 of FIG. 8, the client may receive a notification as to whether an error has occurred during translation. This can occur, for example, if the translation engine encounters a character for which it does not have a translation. In the case of the above message, if the translation engine did not have a suitable translation for the question mark at the end of the text string, the translation engine would notify the client of this error.
In certain embodiments, the notification is made immediately. In another embodiment, the notification is part of a response sent to the client.

Another function of the notification mechanism of the present invention is that the notification mechanism of the present invention has an interface defined by IDL. Therefore, the reporting of error messages between the client and the translation engine is standardized.

For example, in certain embodiments, as would be understood by one of ordinary skill in the art, specifying an error, providing severity information, and providing a numeric identifier for the error Using a callback interface for In one embodiment, the error callback object is associated with the object that caused the error (such as an object associated with asynchronous translation). In other embodiments, an error callback object may be provided as an argument to other steps. In one embodiment, the client user interface implements this interface with the conversation box. However, it should be understood that any user communication that can report an error can be used to inform the user of the error.

In another embodiment implementing this notification mechanism, the numerical error identifier is
Associated with a custom user interface language to provide error messages to the user interface application.

In step 78 of FIG. 8, the translation engine in one specific example can notify the client of the progress of the translation. Such a notification may be sent when the text to be translated is very long and / or when the client translates, such as an asynchronous translation (where the client is free to perform other tasks and translates in the background). Useful if you select an option.

Referring again to FIG. 6, after the translation step 56 is completed, the translation engine creates a response (step 58) and sends the response to the client via the distributed object protocol (step 60).

Using the previous example, in response to the translation request below, the “Do you speak English?” Client receives a response containing text in a second language. For English-Japanese translation engines, the option to use English alphabetic characters
The received text is as follows.

“Eigo o hanashimasu ka?” Any embodiment of the modular language translation system of the present invention, including all of the above features, may be implemented on a general-purpose computer (eg, Apple Macintosh, IBM PC).
Or computer software on a computer readable medium, such as a diskette or optical compact disc (CD) for execution on a compatible machine, Sun workstation, etc.).

Various alterations, modifications and other implementations of the present invention will occur to those skilled in the art without departing from the spirit and scope of the invention. Therefore,
The invention is not defined by the foregoing illustrative description, but rather by the spirit and scope of the following claims.

[Brief description of the drawings]

FIG. 1 is a block diagram of a modular language translation system of the present invention.

FIG. 2 is a block diagram of a particular embodiment of the modular language translation system of FIG. 1 of the present invention.

FIG. 3 is another block diagram of the modular language translation system of the present invention.

FIG. 4 is a block diagram of a modular language translation system having multiple translation engines according to one embodiment of the present invention.

FIG. 5 is a block diagram of a modular language translation system having a plurality of clients according to another embodiment of the present invention.

FIG. 6 is a flowchart showing an operation of the system of FIG. 2;

FIG. 7 is a more detailed flowchart of the step of transmitting a translation request in FIG. 6;

FIG. 8 is a more detailed flowchart of the translation transmission step of FIG. 6;

Claims (62)

[Claims] 1. A translation system, comprising: (a) transmitting a translation request including text to be translated by a distributed object protocol, and corresponding to a translation from a first language to a second language; A client for receiving a response to the request by the distributed object protocol; and (b) receiving the translation response by the distributed object protocol and transmitting the response to the client by the distributed object protocol. And a translation engine for performing the translation. 2. The translation system according to claim 1, wherein said translation request and said response are in an interface definition language (IDL) defining an interface at said client and an interface at said translation engine. The translation system sent. 3. The translation system according to claim 1, wherein the distributed object protocol used by the client and the translation engine operates according to a Component Object Model (COM) standard. 4. The translation system according to claim 1, wherein the distributed object protocol used by the client and the translation engine operates according to a Common Object Request Broker Architecture (CORBA) standard. . 5. The translation system according to claim 1, wherein the client and the translation engine are located on a same computer. 6. The translation system according to claim 1, wherein the client and the translation engine are located on different computers communicating with each other.
Translation system. 7. The translation system according to claim 6, wherein said client and said translation engine communicate over a wide area network. 8. The translation system according to claim 6, wherein said client and said translation engine communicate over a local area network. 9. The translation system according to claim 6, wherein said client and said translation engine communicate over the World Wide Web. 10. The translation system according to claim 1, wherein the text to be translated is represented by a set of characters including characters used in a plurality of languages. 11. The translation system according to claim 1, wherein the text to be translated is represented by UNICODE. 12. The translation system according to claim 1, wherein the translation request further includes: a part-of-speech setting, an annotation, an HTML markup,
At least one of GML markup, RTF markup, NROFF markup, translation options, different phrase settings, different translations, end-of-sentence annotations, translation guidelines, translation hints, and indicators of punctuation in the text. Translation system, including one. 13. The translation system according to claim 1, further comprising a notification mechanism whereby the translation engine can notify the client about at least one error that occurs during translation. system. 14. The translation system of claim 13, wherein the notification mechanism comprises an IDL-defined callback interface having at least one callback object, and wherein the translation engine communicates with the client. A translation system for notifying an error that occurs during translation by associating the error with the error callback object. 15. The translation system according to claim 1, wherein the client further comprises a user interface having a translation service for accepting a translation request. 16. The translation system of claim 15, wherein the distributed object protocol used by the client and the translation engine supports the user interface and a plurality of other user interfaces. . 17. The translation system according to claim 1, wherein the distributed object protocol used by the client and the translation engine comprises:
A translation system that supports said translation engine and a plurality of other translation engines. 18. The translation system according to claim 1, further comprising an IDL-defined registration mechanism whereby the translation engine can register itself to be discovered by the client. , Translation system. 19. The translation system according to claim 18, wherein the registration mechanism includes information on at least one item that can be obtained by the translation engine. 20. The translation system according to claim 18, wherein the item is:
User dictionary, syntax dictionary, user dictionary editor, dictionary browser, second translation engine, another phrase discovery device, another sentence, end-of-sentence discovery device, (background) translation, fuzzy translation matching, part of speech setting, selective translation, And a translation system selected from the group consisting of translation memory archives. 21. The translation system according to claim 1, wherein the translation request further includes information on a format required for translation, and controls a method of processing the translation request by the translation engine from the client. Using said translation priority object having an interface defined by IDL, said translation engine responding to information for said translation request using said distributed object protocol in a format requiring translation. A translation system that creates responses. 22. The translation system according to claim 21, further comprising: a client that can input a format required for translation into the priority editor by a distributed object protocol. , I
A translation system, comprising: a priority editor having an interface defined by a DL, the priority editor providing the translation engine with the information on a format requiring translation as a translation priority object by a distributed object protocol. . 23. The translation system according to claim 1, wherein the translation engine further comprises:
A translation system comprising a synchronous translation device defined by IDL for providing an immediate response to said translation request. 24. The translation system according to claim 1, wherein said client further comprises a script-enabled application for calling said synchronous translator, wherein said client invokes said synchronous translator via a distributed object protocol. A translation system capable of translating text in the application by supplying a script in the application. 25. The translation system according to claim 1, wherein the client further comprises a source that supplies the text to be translated and a location for storing the response sent by the translation engine. The engine further provides the client with a message while the client is performing other tasks.
An asynchronous translator defined by IDL for providing a response to the translation request according to a distributed object protocol, wherein the source translates to the asynchronous translator according to the distributed object protocol A translation system for providing the text, wherein the location receives the response sent by the asynchronous translation device according to a distributed object protocol. 26. The translation system according to claim 25, wherein the client polls the asynchronous translation device according to the distributed object protocol for information on the progress of translation. 27. The translation system according to claim 25, wherein the asynchronous translation device supplies a client with information on the progress of translation while translation is being performed by the distributed object protocol. Translation system. 28. The translation system of claim 1, wherein the client is further configured to receive, from the translation engine, information regarding a progress of creating a response by the translation engine via the distributed object protocol.
A translation system including a progress object having an interface defined by IDL. 29. The translation system of claim 1, further comprising: the distributed object protocol for locating at least one other phrase in response to a translation request from the client. A translation system, comprising another phrase server having an interface defined by IDL, said another phrase server informing the client of the selection of the another phrase by a distributed object protocol. 30. The translation system according to claim 1, further comprising: an IDL for locating a selection of at least one other sentence in response to a request from the client by the distributed object protocol. A translation system comprising a server for another sentence having a defined interface, wherein the server for another sentence communicates with the selection of the other sentence via a distributed object protocol. 31. The translation system according to claim 1, further comprising a distributed object protocol in response to a request from the client,
A translation system comprising: an end-of-sentence server having an interface defined by IDL for determining the location of a sentence portion, wherein the end-of-sentence server communicates this information to the client via a distributed object protocol. 32. The translation system according to claim 1, further comprising: a part-of-speech registration mechanism having an interface defined by IDL indicating a part-of-speech that can be used at the translation engine, A translation system that communicates this information to the client via a distributed object protocol. 33. A translation method, comprising: (a) transmitting a translation request including text to be translated from a first language to a second language from a client to a translation engine by a distributed object protocol; b) receiving, at the translation engine, a request according to the distributed object protocol; and (c) at the translation engine, responding to the received translation request by translating the text from a first language to a second language. Performing a translation to create a response corresponding to the translation of the text from the first language to the second language; and (d) translating the response from the translation engine by the distributed object protocol from the translation engine. Transmitting to the client. 34. The method of claim 33, wherein the step of sending a translation request from a client further comprises: an IDL defining an interface at the client and an interface at the translation engine. A method comprising sending a translation request from a client to a translation engine. 35. The method according to claim 33, wherein the translation engine comprises a registration mechanism including an interface defined by IDL that can register itself so that it can be found by the client. ,
Method. 36. The method of claim 33, wherein transmitting the translation request further comprises transmitting a translation request including information about a format that requires translation, and wherein translating the text. The method wherein the steps further comprise the step of translating the text in a format requiring translation. 37. The method of claim 33, further comprising:
A method, comprising: a notification mechanism that includes an interface defined by IDL, wherein the translation engine can inform the client if a translation cannot be performed in a format that requires translation. 38. The method of claim 33, further comprising, before performing the translation, undefined words, undefined columns, undefined characters, spelling errors, and punctuation errors. Scanning the translation request for at least one error in the text selected from the group. 39. The method of claim 33, further comprising providing the client with information regarding the progress of the creation of the response at the translation engine according to the distributed object protocol. . 40. The method of claim 33, wherein the distributed object protocol allows the translation engine to inform the client about at least one error that occurs during translation.
Providing an annotation mechanism having an interface defined by IDL. 41. A translation system, comprising: (a) each receiving a translation request including a text to be translated from a first language to a second language by a distributed object protocol; A plurality of translation engines for creating and transferring a response to the translation request including translation of the text from the first language to the second language in response to the translation request; A translation system, comprising: a client for transmitting the translation request to any of the plurality of translation engines according to a protocol, and receiving the response from the translation engine according to the distributed object protocol. 42. The translation system according to claim 41, wherein the translation request and the response are transmitted by an IDL defining an interface at each of the plurality of translation engines and the client. . 43. The translation system according to claim 41, further defined by IDL, wherein each of said plurality of translation engines can register itself as discovered by said client. A translation system with a registration mechanism with an interface. 44. The translation system according to claim 41, wherein the client further has a translation service for accepting translation requests and the client and the plurality of translation engines use to communicate. A translation system where the type object protocol supports multiple translation engines. 45. The translation system of claim 41, wherein the plurality of translation engines and the distributed object protocol used by the client to communicate operate according to the COM standard. 46. The translation system according to claim 41, wherein the distributed object protocol used by the plurality of clients and the translation engine to communicate operates according to CORBA standards. 47. The translation system according to claim 41, wherein the translation request further includes information regarding a format required for translation, and as a result, the translation engine receiving the translation request includes the distributed object. -A translation system that can create responses using protocols. 48. The translation system according to claim 41, further comprising: providing the client with the distributed object protocol for availability of the format required for translation at each of the plurality of translation engines. A translation system comprising a notification mechanism having an interface defined by IDL for notification. 49. The translation system according to claim 41, wherein the plurality of translation engines respond to the translation of the text from the first language to a plurality of second languages by the distributed object protocol. Create multiple responses,
Translation system. 50. The translation system according to claim 4, wherein at least one of the client and the translation engine is located on the same computer. 51. The translation system according to claim 41, wherein at least one of the client and the plurality of translation engines are located on different computers communicating with each other. 52. The translation system according to claim 41, wherein at least one of the client and the translation engine communicate with a proxy server. 53. A translation system, comprising: (a) text to translate from a first language to a second language to at least one translation engine by a distributed object protocol; A plurality of clients for sending a response corresponding to the translation of the text from the second language to a second language and receiving a response to the translation request by the distributed object protocol; and (b) by a distributed object protocol. A translation engine for receiving at least one translation request from any one of a plurality of clients, creating a response, and sending the response to the request to the client according to the distributed object protocol. A transmission system comprising: 54. The translation system according to claim 53, wherein each of the plurality of translation requests and the response are transmitted by an IDL defining an interface at each of the client and the translation engine. system. 55. The translation system of claim 53, wherein the distributed object protocol used by the plurality of clients and the translation engine to communicate operates according to the COM standard. 56. The translation system according to claim 53, wherein said distributed object protocol used by said plurality of clients and said translation engine to communicate operates according to CORBA standards. 57. The translation system according to claim 53, wherein each of said plurality of clients further has a translation service for accepting translation requests.
A translation system comprising a user interface defined by a DL and wherein said distributed object protocol used by said plurality of clients and said translation engine to communicate comprises supporting a plurality of user interfaces. 58. The translation system according to claim 53, wherein at least one of the translation engine and the plurality of clients are located on a same computer. 59. The translation system of claim 53, wherein at least one of the translation engine and the plurality of clients are located on different computers communicating with each other. 60. The translation system of claim 53, wherein at least one of the translation engine and the plurality of clients communicate over a World Wide Web. 61. The translation system of claim 53, wherein the translation engine further comprises an IDL defined interface for maintaining a separate context for each translation request received by the translation engine. A translation system that includes multiple translation objects with. 62. The translation system according to claim 61, wherein at least one of the plurality of translation objects further responds to a translation request including information on a format required for translation, and Responding to the information to control how the translation engine processes the translation request.