Integrated authoring and translation system

A computer-based system for monolingual document development, comprising:

• a text editor adapted to accept interactively from a author information elements written in a source language;
• a language editor, which is an extension of said text editor, which interactively enforces first lexical and then grammatical constraints on a natural language subset used by said author to create unambiguous constrained information elements, wherein said author interactively aids in enforcing first said lexical constraints and then said grammatical constraints on said input text so as to produce said unambiguous constrained information elements;
• memory means for storing said unambiguous constrained information elements for subsequent use;
• a machine translation system, responsive to said language editor, that is configured to translate said unambiguous constrained information elements into a foreign language without pre-editing and without postediting;
• and
• a domain model, which communicates with said language editor, wherein said domain model provides pre-determined domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, so as to assist said language editor in said enforcement of said lexical and grammatical constraints wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and said machine translation system, wherein said lexical information includes all lexical items within said natural language subset along with associated semantic concepts, parts of speech, and morphological information
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said natural language subset, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation.

A computer-based method for monolingual document development, comprising the steps of:

• (1) entering input text in a source language into a text editor;
• (2) checking, via a language editor said input text against a pre-determined set of constraints stored in a domain model that provides pre-determined domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, said pre-determined set of constraints includes a set of source sublanguage rules concerning vocabulary and grammar, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items that satisfy said pre-determined set of constraints along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes synonyms for items that do not satisfy said pre-determined set of constraints, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (3) providing to an author interactive feedback relating to said input text, said interactive feedback indicating if said predetermined set of constraints is met, said interactive feedback is performed subsequent to referring to said domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, and grammar of a subset of a natural language;
• and
• (4) producing, after completion of step (3), unambiguous constrained text.

A computer-based method for monolingual document development, comprising the steps of:

• (1) entering input text in a source language into a text editor;
• (2) checking, via a language editor, said input text against a constrained source language;
• (3) providing to an author interactive feedback relating to said source input text if non-constrained source language is present in said source input text until said author modifies said source input text into a constrained source text, said interactive feedback is performed after consulting a domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said constrained source language, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (4) checking for syntactic grammatical errors and semantic ambiguities in said constrained source text by consulting said domain model;
• and
• (5) providing to said author interactive feedback to remove said syntactic grammatical errors and said semantic ambiguities in said constrained source text to produce unambiguous constrained text.

A computer-based method for monolingual document development, comprising the steps of:

• (1) entering into a text editor at least one information element created in a source language;
• (2) checking, via a language editor, said at least one information element against a constrained source language;
• (3) providing to an author interactive feedback relating to said at least one information element if non-constrained source language is present in said at least one information element until said at least one information element has been modified into a constrained source text, said interactive feedback is performed after referring to a domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and their combinations, wherein said domain model is a tripartite domain model, said tripartite domain model comprising:
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said constrained source language, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (4) checking for syntactic grammatical errors and semantic ambiguities in said constrained source text by consulting said domain model;
• (5) providing interactive feedback to said author to remove said syntactic grammatical errors and said semantic ambiguities in said constrained source text to produce at least one unambiguous constrained information element;
• and
• (6) saving said at least one unambiguous constrained information element for later use.

A computer-based system for translating source language input text to a foreign language without pre-editing and without postediting, comprising:

• a text editor adapted to accept interactively from an author the input text written in a source language;
• a language editor, which is an extension of said text editor, which interacts with said author to produce from said input text an unambiguous constrained source text by interactively enforcing first vocabulary constraints and then grammatical constraints against a constrained source language;
• a machine translation system, responsive to said language editor, which is configured to translate said unambiguous constrained source text into the foreign language without pre-editing and without postediting;
• and
• a domain model, which communicates with said language editor and said machine translation system, and which provides predetermined domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, so as to aid in producing said unambiguous constrained source text and in said translation to the foreign language, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and said machine translation system, wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes synonyms for items not within said constrained source language, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation.

A computer-based system for monolingual document development and multilingual translation, comprising:

• a text editor adapted for accepting interactively from an author information elements written in a source language;
• a language editor, which is an extension of said text editor, which interactively enforces lexical and grammatical constraints on a natural language subset used by said author to create said input text, wherein said author is interactively aided in enforcing said lexical and grammatical constraints on said information elements to produce said unambiguous constrained information elements;
• machine translation system, responsive to said language editor, which translates said unambiguous constrained information elements into a foreign language, wherein the translated text requires no postediting;
• and
• a domain model, which communicates with said language editor and said machine translation means, wherein said domain model provides pre-determined domain knowledge and linguistic semantic knowledge about lexical units and their combinations, so as to aid in producing said unambiguous constrained source text and in said translation to said foreign language, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and said machine translation system, wherein said lexical information includes all lexical items within said natural language subset along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said natural language subset dictionary definitions of said lexical items, and examples of using said lexical item, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation.

A computer-based system for monolingual document development and multilingual translation, comprising:

• (A) a text editor adapted to accept interactively from an author input text written in a source language;
• (B) a language editor, which is an extension of said text editor, which interactively enforces lexical and grammatical constraints on a natural language subset used by said author to create said input text, said language editor comprising,
• (i) a vocabulary checker which identifies occurrences of words that do not conform to said lexical constraints and which interactively aids said author in finding valid lexical replacements for said words that do not conform, and
• (ii) a grammar checker which provides interactive feedback to said author concerning syntactic and semantic ambiguity, said interactive feedback producing unambiguous constrained text;
• and
• (C) a domain model which communicates with said language editor, wherein said domain model provides pre-determined domain knowledge and linguistic semantic knowledge about lexical units and their combinations;
• and
• (D) a machine translation system, responsive to said language editor, which is configured to translate said unambiguous constrained text into a foreign language, wherein the translated text requires no postediting;
• wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and said machine translation system, wherein said lexical information includes all lexical items within said natural language subset along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said natural language subset, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation.

A computer-based method for translating source language text to a foreign language without pre- or postediting, comprising the steps of:

• (1) entering input text in a source language into a text editor;
• (2) checking, via a language editor, said input text against a constrained source language;
• (3) providing to an author interactive feedback relating to said source input text if non-constrained source language is present in said source input text until said author modifies said source input text into a constrained source text, said interactive feedback includes allowing said author to select, from a list of at least one synonym, a word or phrase to replace said non-constrained source language;
• (4) checking for syntactic grammatical errors and semantic ambiguities in said constrained source text;
• (5) providing interactive feedback to said author to remove said syntactic grammatical errors and said semantic ambiguities in said constrained source text to produce unambiguous constrained source text;
• and
• (6) translating, via a machine translation system, said unambiguous constrained source text into a target language;
• wherein steps (2) and (4) further includes the step of communicating with a tripartite domain model (DM), wherein said tripartite DM provides pre-determined domain knowledge and linguistic semantic knowledge about lexical units and their combinations, said tripartite domain model including,
• a kernel which contains all lexical information that is required by said language editor and said machine translation system wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes synonyms for items not within said constrained source language, dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts for unambiguous mapping and semantic verification in translation.

A computer-based method for monolingual document development and multilingual translation, comprising the steps of:

• (1) entering input text in a source language into a text editor;
• (2) checking, via a language editor, said input text against a pre-determined set of constraints stored in a domain model, wherein said pre-determined set of constraints includes a set of source sublanguage rules concerning vocabulary and grammar, wherein said interactive feedback is performed in order to make said input text conform with said set of source sublanguage rules and to eliminate ambiguities, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items that satisfy said pre-determined set of constraints along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes synonyms for items that do not satisfy said pre-determined set of constraints, dictionary definitions of said lexical items and examples of using said lexical items, and
• a machine translation domain model, which contains information which is required by only said machine translation systems, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (3) providing to an author interactive feedback relating to said input text if said pre-determined set of criteria is not met, said interactive feedback is performed subsequent to consulting said domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and their combinations, whereto said author produces, through said interactive feedback, unambiguous constrained source text;
• (4) translating said unambiguous constrained source text into a target language.

A computer-based method for monolingual document development and multilingual translation, the computer-based method comprising the steps of:

• (1) entering input text in a source language into a text editor;
• (2) checking, via a language editor, said input text against a constrained source language;
• (3) providing to an author interactive feedback relating to said source input text if non-constrained source language is present in said source input text until said source input text has been modified into a constrained source text, said interactive feedback being done subsequent to consulting a domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and their combinations, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor, wherein said information includes natural language subset synonyms for items not within said constrained source language dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (4) checking for syntactic grammatical errors and semantic ambiguities in said constrained source text by consulting said domain model;
• (5) providing interactive feedback to said author to remove said syntactic grammatical errors and said semantic ambiguities in said constrained source text to produce a unambiguous constrained source text;
• and
• (6) translating with said machine translation system said unambiguous constrained source text into a foreign language with the aid of said domain model.

A computer-based method for monolingual document development and multilingual translation, comprising the steps of:

• (1) entering into a text editor at least one information element created in a source language;
• (2) checking, via a language editor, said at least one information element against a constrained source language;
• (3) providing to an author interactive feedback relating to said at least one information element if non-constrained source language is present in said at least one information element until said at least one information element has been modified into a constrained source text, said interactive feedback is performed after consulting a domain model which provides the necessary domain knowledge and linguistic semantic knowledge about lexical units and of their combinations, wherein said domain model is a tripartite domain model, said tripartite domain model comprising,
• a kernel which contains all lexical information that is required by said language editor and a machine translation system, wherein said lexical information includes all lexical items within said constrained source language along with associated semantic concepts, parts of speech, and morphological information,
• a language editor domain model which contains information that is required only by said language editor wherein said information includes natural language subset synonyms for items not within said constrained source language dictionary definitions of said lexical items, and examples of using said lexical items, and
• a machine translation domain model which contains information which is required by only said machine translation system, said machine translation domain model includes a hierarchy of concepts used for unambiguous mapping and semantic verification in translation;
• (4) checking for syntactic grammatical errors and semantic ambiguities in said constrained text by consulting said domain model;
• (5) providing interactive feedback to said author to remove said syntactic grammatical errors and said semantic ambiguities in said constrained source text to produce at least one unambiguous constrained information element;
• (6) saving said at least one unambiguous constrained information element for later use.
• (7) translating with said machine translation system said at least one unambiguous constrained information element into a foreign language.