System, method, and product for dynamically aligning translations in a translation-memory system

• US 6,345,244 B1
• Filed: 05/27/1998
• Issued: 02/05/2002
• Est. Priority Date: 05/27/1998
• Status: Expired due to Term

• Current Assignee: Lionbridge US, Inc. (Lionbridge Technologies Incorporated)
• Original Assignee: Lionbridge Technologies Incorporated
• Inventors: Clark, Jonathan P.
• Primary Examiner(s): EDOUARD, PATRICK NESTOR
• Application Number:
  US09/085,471
• Time in Patent Office: 1,350 Days
• Field of Search: 704/2-8, 704/9-10, 704/275, 704/257, 704/531, 707/530, 707/536
• US Class Current: 704/2
• CPC Class Codes: G06F 40/47 Machine-assisted translation, e.g. using translation memory

Abstract

A computer-implemented system, method, and product associate words, phrases, or other characters (“words”) to be translated with previous translations of such words, if a previous translation exists. The words to be translated, and their identifying attributes (such as, for example, formatting information), are extracted from one or more source files. The previous translations, and their identifying attributes, are extracted from one or more target files. If a previous translation does not exist, the words to be translated, and their attributes, are copied to create a pseudo-previous translation, with attributes. Each word to be translated is associated with a corresponding previous translation based upon commonalities of their attributes. Such words may also be associated based upon their respective locations in their respective files. Each word to be translated, and its attributes, may be stored with a previous translation in a source-target pair record of a source-target pair database. Each source-target pair record may include a propagation flag identifying whether the previous translation stored in the source-target pair record is to be propagated to other occurrences of the associated word to be translated in the source-target pair database. Each source-target pair record may also include a pointer to a page of an occurrence book having pages, wherein each page includes pointers to a common word to be translated in records of the source-target pair database.

Claims

1. A computer program product comprising a computer-readable medium and instructions stored thereon, said instructions, when executed by an appropriate digital computer causing the digital computer to operate as an aligner which associates translatable source segments from one or more source files having a first format with corresponding target segments extracted from one or more target files having the first format, such association being based upon commonality of attributes associated with the segments, wherein at least one such attribute is selectable, whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and at least some times excludes an attribute from a commonality of attributes necessary to achieve a match.

2. The computer program product of claim 1, wherein the aligner is further configured to associate the source and target segments based upon respective locations in their respective files.

3. The computer program product of claim 1, wherein the aligner stores each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

4. The computer program product of claim 3, wherein each source-target pair record comprises:

a propagation flag identifying whether the corresponding target segment stored in the source-target pair record is to be propagated to other occurrences of the associated translatable source segment in the source-target pair database.

5. The computer program product of claim 3, wherein each source-target pair record comprises:

a pointer to a page of an occurrence book comprising pages, each page comprising pointers to a common translatable source-segment in records of the source-target pair database.

6. The computer program product of claim 1,

wherein each of the extracted translatable source segments are stored in a record of a source database and each of the extracted corresponding target segments are stored in a record of a target database, and

wherein the aligner associates each translatable-source-segment record with one corresponding-target-segment record.

7. The computer program product of claim 1, wherein the aligner is configured to assign a unique identifier to each translatable source segment, wherein the unique identifier is generated based on the attributes of the translatable source segment.

8. The computer program product of claim 7, wherein the unique identifier is generated based upon a location of each translatable source segment in its source file.

9. The computer program product of claim 6, wherein the aligner is configured to assign a unique identifier to each translatable source segment, wherein the unique identifier is generated based on the attributes of the translatable source segment.

10. The computer program product of claim 9, wherein the unique identifier of each translatable source segment is stored in the source-segment record and wherein the unique identifier of each corresponding target segment is stored in the target-segment record.

11. The computer program product of claim 6, wherein the association is based upon a matching of the unique identifier of the source and target segments.

12. The computer program product of claim 1, wherein the source and target files include files in a file system.

13. The computer program product of claim 1, wherein the file system is a local file system.

14. The computer program product of claim 1, wherein the file system is a remote file system.

15. The computer program product of claim 12, wherein the aligner further comprises:

a project identifier that selects the source and target files from the files in the file system.

16. The computer program product of claim 15, wherein said selection is accomplished under the direction of user-translator.

17. The computer program product of claim 15, wherein said selection is accomplished based upon predetermined selection criteria.

18. The computer program product of claim 15, wherein legacy files exist and wherein said project identifier comprises:

means for identifying existing legacy files associated with one or more of the source and target files.

19. The computer program product of claim 1, further comprising:

a parser-extractor for extracting each translatable source segment and its attributes from the one or more source files and for extracting each corresponding target segment from the one or more target files.

20. The computer program product of claim 19, wherein the parser-extractor comprises:

a syntactic customizer that generates a customized syntactical description of the first format.

21. The computer program product of claim 20, wherein the customized syntactical description comprises a syntactic rule for identifying the source segments in the one or more source files and the target segments in the one or more target files.

22. The computer program product of claim 21, wherein the syntactic rule is in a BNF form.

23. The computer program product of claim 21, wherein the syntactic rule is determined manually.

24. The computer program product of claim 20, wherein the customized syntactical description comprises a tagged syntactical element for uniquely identifying source and target segments.

25. The computer program product of claim 24, wherein the tagged syntactical element comprises a tag that is an extension to a conventional BNF notation.

26. The computer program product of claim 24, wherein the tag is generated by comparing an exemplar of a file having the first format to a syntactical rule in BNF form for the first format.

27. The computer program product of claim 24, wherein the tag is generated by determining a degree of variability across a plurality of samples of files of the first format type.

28. The computer program product of claim 19, wherein the parser-extractor further comprises:

means for parsing the source files to generate the translatable source segments and their attributes;

means for parsing the target files to generate the corresponding target segments and their attributes;

means for extracting the translatable source segments and their attributes;

means for extracting the corresponding target segments and their attributes;

means for storing the translatable source segments and their attributes in a source segment and attribute list; and

means for storing the corresponding target segments and their attributes in a target segment and attribute list.

29. The computer program product of claim 28, wherein the parser-extractor further comprises:

means for identifying a pre-existing target file corresponding to each source file; and

means for generating a target file when the pre-existing target file does not exist.

30. The computer program product of claim 19, wherein the parser-extractor further comprises:

a conflict resolver that determines whether the attribute identifier of each translatable source segment and each corresponding target segment is a unique attribute identifier and, if not, assigns a unique attribute identifier.

31. The computer program product of claim 30, wherein the unique attribute identifier includes hashed representations of identifying attributes of each translatable source segment.

32. The computer program product of claim 30, wherein the unique attribute identifier includes hashed representations of identifying attributes of each corresponding target segment.

33. The computer program product of claim 19, wherein the parser-extractor employs morpho-syntactic analysis to identify the source and target segments.

34. The computer program product of claim 1, wherein the first format is a resource file format.

35. A method for associating translatable source segments extracted from one or more source files having a first format with corresponding target segments extracted from one or more target files having the first format, comprising the steps of:

(1) determining identifying attributes of each translatable source segment, wherein at least one such attribute is selectable;

(2) generating a unique attribute identifier for each translatable source segment based upon its identifying attributes;

(3) determining identifying attributes of each corresponding target segment, wherein at least one such attribute is selectable;

(4) generating a unique attribute identifier for each corresponding target segment based upon its identifying attributes;

(5) comparing the unique attribute identifiers of the translatable source segments and corresponding target segments; and

(6) associating a translatable source segment with a corresponding target segment when they have the same unique attribute identifier;

whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and an attribute may be excluded from a commonality of attributes necessary to achieve a match.

36. The method of claim 35, wherein step (1) comprises the steps of:

(a) identifying a first type of file of the first format; and

(b) searching for identifying attributes based on a syntactical description of the first type of file.

37. The method of claim 35, wherein step (1) comprises the steps of:

(a) identifying a first type of file of the first format;

(b) customizing a syntactical description of the first type of file;

(b) searching for identifying attributes based on the customized syntactical description.

38. The method of claim 37, wherein step (1)(b) comprises the step of:

(i) tagging a syntactical element with a tag that is an extension to a conventional BNF notation.

39. The method of claim 35, wherein step (2) comprises the step of:

(a) further generating the unique attribute identifier of each translatable source segment based upon its locations in a source file.

40. The method of claim 35, wherein step (3) comprises the step of:

(a) further generating the unique attribute identifier of each corresponding target segment based upon its locations in a target file.

41. The method of claim 35, further comprising the step of:

(7) storing each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

42. The method of claim 41, further comprising the step of:

(8) storing in each source-target pair record a propagation flag identifying whether the corresponding target segment stored in the source-target pair record is to be propagated to other occurrences of the associated translatable source segment in the source-target pair database.

43. The method of claim 41, further comprising the step of:

(8) storing in each source-target pair record a pointer to a page of an occurrence book comprising pages, each page comprising pointers to a common translatable source-segment in records of the source-target pair database.

44. The method of claim 35, wherein the source and target files include files in a file system and further comprising the step of:

(7) selecting the source and target files from the files in the file system.

45. The method of claim 35, further comprising the step of:

(7) identifying existing legacy files associated with one or more of the source and target files.

46. The method of claim 35, wherein step (4) comprises the steps of:

(a) generating hashed representations of identifying attributes of each translatable source segment; and

(b) generating hashed representations of identifying attributes of each corresponding target segment.

47. A computer system having at least one central processing unit (CPU), an operating system, at least one memory unit, and an aligner configured to cooperate with the at least one CPU and the operating system to associate translatable source segments extracted from one or more source files having a first format with corresponding target segments extracted from one or more target files having the first format, such association being based upon commonality of attributes of the segments, wherein at least one such attribute is selectable, whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and an attribute may be excluded from a commonality of attributes necessary to achieve a match.

48. The computer system of claim 47, wherein the aligner comprises:

(1) means for determining identifying attributes of each translatable source segment;

(2) means for generating a unique attribute identifier for each translatable source segment based upon its identifying attributes;

(3) means for determining identifying attributes of each corresponding target segment;

(4) means for generating a unique attribute identifier for each corresponding target segment based upon its identifying attributes;

(5) means for comparing the unique attribute identifiers of the translatable source segments and corresponding target segments; and

(6) means for associating a translatable source segment with a corresponding target segment when they have the same unique attribute identifier.

49. The computer system of claim 47, further comprising:

(7) means for storing each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

50. The computer system of claim 49, further comprising:

(8) means for storing in each source-target pair record a propagation flag identifying whether the corresponding target segment stored in the source-target pair record is to be propagated to other occurrences of the associated translatable source segment in the source-target pair database.

51. The computer system of claim 49, further comprising:

(8) means for storing in each source-target pair record a pointer to a page of an occurrence book comprising pages, each page comprising pointers to a common translatable source-segment in records of the source-target pair database.

52. Storage media that contains software that, when executed on an appropriate computing system having at least one central processing unit (CPU), an operating system, and at least one memory unit, performs a method to associate translatable source segments extracted from one or more source files having a first format with corresponding target segments extracted from one or more target files having the first format, such association being based upon commonality of attributes of the segments, the method comprising the steps of:

(1) determining identifying attributes of each translatable source segment, wherein at least one such attribute is selectable;

(2) generating a unique attribute identifier for each translatable source segment based upon its identifying attributes;

(3) determining identifying attributes of each corresponding target segment, wherein at least one such attribute is selectable;

(4) generating a unique attribute identifier for each corresponding target segment based upon its identifying attributes;

(5) comparing the unique attribute identifiers of the translatable source segments and corresponding target segments; and

(6) associating a translatable source segment with a corresponding target segment when they have the same unique attribute identifier;

whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and an attribute may be excluded from a commonality of attributes necessary to achieve a match.

53. The storage media of claim 52, wherein the method further includes the step of:

(7) storing each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

54. The storage media of claim 52, wherein the method further includes the step of:

(7) storing in each source-target pair record a propagation flag identifying whether the corresponding target segment stored in the source-target pair record is to be propagated to other occurrences of the associated translatable source segment in the source-target pair database.

55. The storage media of claim 54, wherein the method further includes the step of:

(8) storing in each source-target pair record a pointer to a page of an occurrence book comprising pages, each page comprising pointers to a common translatable source-segment in records of the source-target pair database.

56. A computer program product for use with an appropriate computing system having at least one central processing unit (CPU) and at least one memory unit, the computer program product comprising a computer usable medium having embodied therein computer readable program code method steps to associate translatable source segments extracted from one or more source files having a first format with corresponding target segments extracted from one or more target files having the first format, such association being based upon commonality of attributes of the segments, the method steps comprising:

(1) determining identifying attributes of each translatable source segment, wherein at least one such attribute is selectable;

(2) generating a unique attribute identifier for each translatable source segment based upon its identifying attributes;

(3) determining identifying attributes of each corresponding target segment, wherein at least one such attribute is selectable;

(4) generating a unique attribute identifier for each corresponding target segment based upon its identifying attributes;

(5) comparing the unique attribute identifiers of the translatable source segments and corresponding target segments; and

(6) associating a translatable source segment with a corresponding target segment when they have the same unique attribute identifier;

whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and an attribute may be excluded from a commonality of attributes necessary to achieve a match.

57. The computer program product of claim 56, wherein the method steps further comprise:

(7) storing each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

58. The computer program product of claim 56, wherein the method steps further comprise:

(7) storing in each source-target pair record a propagation flag identifying whether the corresponding target segment stored in the source-target pair record is to be propagated to other occurrences of the associated translatable source segment in the source-target pair database.

59. The computer program product of claim 58, wherein the method steps further comprise:

(8) storing in each source-target pair record a pointer to a page of an occurrence book comprising pages, each page comprising pointers to a common translatable source-segment in records of the source-target pair database.

60. A translation memory system for translating translatable source segments in one or more source files of a source project, the system having an aligner constructed and arranged to associate each of at least one translatable source segment extracted from at least one source file having a first format with a corresponding target segment extracted from at least one target file having the first format, such association being based upon commonality of attributes of the segments, wherein at least one such attribute is selectable, whereby said commonality of attributes allows an imperfect match between sequences of attributes associated with the segments and an attribute may be excluded from a commonality of attributes necessary to achieve a match.

61. The translation memory system of claim 60, wherein the aligner is further constructed and arranged to associate the source and target segments based upon respective locations in their respective files.

62. The translation memory system of claim 60, wherein the aligner stores each translatable source segment and its attributes with its corresponding target segment in a source-target pair record of a source-target pair database.

63. The translation memory system of claim 60, wherein the aligner is constructed and arranged to assign a unique identifier to each translatable source segment, wherein the unique identifier is generated based on the attributes of the translatable source segment.

64. The translation memory system of claim 63, wherein the unique identifier is generated based upon a location of each translatable source segment in its source file.

65. The translation memory system of claim 60, wherein the aligner comprises:

a parser-extract for extracting each translatable source segment and its attributes from the one or more source files and for extracting each corresponding target segment from the one or more target files.

66. The translation memory system of claim 65, wherein the parser-extractor comprises:

a syntactic customizer that generates a customized syntactical description of the first format.

67. The translation memory system of claim 66, wherein the customized syntactical description comprises a syntactic rule for identifying the source segments in the one or more source files and the target segments in the one or more target files.

68. The translation memory system of claim 67, wherein the syntactic rule if in a BNF form.