FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(See section 10, rule 13)
1. Title of the invention: DOCUMENT COMPARISON
2. Applicant(s)
NAME NATIONALITY ADDRESS
ATA CONSULTANCY Indian Nirmal Building. 9th Floor, Nariman Point, SERVICES LIMITED Mumbai, Maharashtra 400021, India
3. Preamble to the description
COMPLETE SPECIFICATION
The following specification particularly describes the invention and the manner in which it is to be performed.

TECHNICAL FIELD
The present subject matter relates to content processing and, particularly but not exclusively, to comparison of documents.
BACKGROUND
Computer-generated documents, such as a hyper text markup language (HTML) file and a Microsoft Word™ document, are usually compared during, for example, a revision process, to identify differences in one document with respect to the other document. Generally, the comparison of the documents is done manually by proofreading the documents on paper or on an electronic display. With an increasing demand for fast and error-free comparison schemes, various machine implemented schemes are also now available. According to such conventional comparison schemes, bitmap image representations of the documents to be compared are rendered, and the corresponding bitmap images of the documents are compared to identify the regions where the bitmaps do not match. Few other conventional schemes exist that compare the documents byte-by-byte to identify the differences between the documents.
SUMMARY
This summary is provided to introduce concepts related to comparison of documents, and the concepts are further described below in the detailed description. This summary is neither intended to identify essential features of the claimed subject matter nor is it intended for use in determining or limiting the scope of the claimed subject matter. In one implementation, a method for document comparison is provided. The method includes obtaining content from a source document and a comparison document, fragmenting the obtained content into a plurality of word strings, where each of the plurality of word strings includes a predetermined number of words, comparing at least one of the plurality of word strings from the source document with the plurality of word strings from the comparison document to obtain candidate match word string pairs, and indicating differences in the comparison document with regard to the source document based on the comparison.
BRIEF DESCRIPTION OF THE DRAWINGS

The detailed description is described with reference to the accompanying figure(s).
In the figure(s), the left-most digit(s) of a reference number identifies the figure in which the
reference number first appears. The same numbers are used throughout the drawings to
reference like features and components.
Fig. 1 illustrates a computing system for document comparison, in accordance with
an implementation of the present subject matter.
Fig. 2 illustrates a method for document comparison, in accordance with an
implementation of the present subject matter.
DETAILED DESCRIPTION
Systems and methods for comparison of documents are described herein. The systems and methods can be implerented in a variety of Computing devices, such as laptops, desktops, workstations, tablet-PCs, smart phones, etc. Although the description herein is with reference to certain Computing systems, the systems and methods may be implemented in other electronic devices, albeit with a few variations, as will be understood by a person skilled in the art.
Conventionally, the process of comparing computer generated documents requires the documents to be rendered on paper, or on an electronic display and compared by a human. This is similar to proof reading. Automated methods evolving thereafter involve rendering bitmap image representations of the documents to be compared, and subsequently comparing them to identify the regions where the bitmaps do not match. Few other conventional schemes that compare the content of the documents byte by byte to identify the differences between the documents exist. The known methods focus more on Visual differences and are slow and error prone. Further, the discrepancies that may be present in the sequence of text characters and data may cause complications in the process of comparson Furthermore, in instances where there are differences in terms of production method, font, or document structure, a more efficient and adaptive comparative method is sought. For example in the known comparison methods, a comparison of documents with different pagination, such as a single line spaced document of two pages, would not be comparable to a double spaced document on three pages, even in the case of identical content.

The present subject matter discloses systems and methods for document comparison. In one implementation, the method may identify differences in the content of a comparison document with respect to the content of a source document. The source document is a document containing information of interest. The comparison document is the document that is required to be compared with the source document in order to identify differences therein with reference to the source document, and then establish the correspondence between the differences. Hereinafter, the source document and the comparison document may collectively be referred to as the documents. According to the present subject matter, the systems and methods reduce time required to assess new documentation and also increase the rate of error detection.
According to the present subject matter, the systems and methods disclosed herein enable a user to focus on error correction rather than error detection. Furthermore, effort required to carry out regression testing, i.e , where no changes are expected, is significantly reduced. Moreover, original document formatting and presentation is maintained post comparison. Furthermore, according to the present subject matter, the comparison is focused on differences in content, rather than differences, such as text formatting, line spacing, font differences, column structure, and the like. Therefore, even content presented using dissimilar media, such as a newspaper and a web browser may be compared using the method of the present subject matter.
According to the present subject matter, the method can also identify differences in punctuation in the source document and the comparison document. Differences in punctuation include missing full stops (period or decimal point), commas, and the like. Many a times the error in punctuation could change the meaning of the sentence, for example, in legal or accounting documents, and so the present method helps in detecting and correcting such errors.
In one implementation, a computing system for document comparison includes a processor and a memory coupled to the processor. The memory includes an extraction module, a fragmentation module and a comparison module. In said implementation, the extraction module is configured to extract content from the source document and the comparison document. Furthermore, the extraction module is configured to parse the extracted content from the source document and the comparison document.

In an implementation, the fragmentation module is configured to fragment the extracted content from the documents into a plurality of overlapping sets of word strings having 'n' words, where 'n' can be any number of words. Furthermore, the comparison module is configured to compare at least one of the plurality of word strings of the source document with the plurality of word strings from the comparison document, determine one or more candidate match word string pairs from the comparison, compute a relative location of the candidate match word string pairs with reference to a datum point, retain those word string pairs with the closest relative location, and tag the rest as non-matching. While aspects of described systems and methods for the comparison of documents can be implemented in any number of different computing systems, environments, and/or configurations, the embodiments are described in the context of the following exemplary system(s).
Fig. 1 illustrates a computing system 100 for comparing a plurality of documents, as per one implementation of the present subject matter. In said implementation, the computing system 100 includes one or more processor(s) 102, interface(s) 104, and a memory 106 coupled to the processor 102. The processor i 02 can be a single processing unit or a number of units, all of which could also include multiple computing units. The processor 102 may be implemented as one or more microprocessors, microcomputers, microcontrollers, digital signal processors, central processing units, state machines, logic circuitries, and/or any devices that manipulate signals based on operational instructions. Among other capabilities, the processor 102 is configured to fetch and execute computer-readable instructions and data stored in the memory 106.
The interfaces 104 may include a variety of software and hardware interfaces, for example, interface for peripheral device(s), such as a keyboard, a mouse, an external memory, and a printer. Further, the interfaces 104 may enable the computing system 100 to communicate with other computing devices, such as web servers and external databases in the communication network (not shown in the figure). The interfaces 104 may facilitate multiple communications within a wide variety of protocols and networks, such as a network, including wired networks, e.g., LAN. cable, etc.. and wireless networks, e.g., WLAN, cellular, satellite, etc. The interfaces 104 may include one or more ports for connecting the computing system 100 to a number of computing devices.

The memory 106 may include any computer-readable medium known in the art including, for example, volatile memory such as static random access memory (SRAM) and dynamic random access memory (DRAM), and/or non-volatile memory, such as read only memory (ROM), erasable programmable ROM, flash memories, hard disks, optical disks, and magnetic tapes. The memory 106 also includes module(s) 108 and data 110. The module(s) 108 include routines, programs, objects, components, data structures, etc., which perform particular tasks or implement particular abstract data types. In one implementation, the module(s) 108 includes an extraction module 112, a fragmentation module 114, a comparison module 116, and other module(s) 118, The other module(s) 118 may include programs or coded instructions that supplement applications and functions of the computing system 100.
On the other hand, the data 110, inter alia serves as a repository for storing data processed, received, and generated by one or more of the module(s) 108. The data 110 includes for example, fragmentation data 120, extraction data 122, comparison data 124, and other data 126. The other data 126 includes data generated as a result of the execution of one or more modules in the module(s) 108.
In one implementation, the computing system 100 compares content from a source document with content of a comparised document. The source document is a document containing information of interest. The comparison document is the document that is required to be compared with the source document in order to identify the differences therein with reference to the source document and then establish the correspondence between them. Further, the word content as used herein relates to human-readable text and data, i.e., the information content irrespective of the formatting, such as tags in an HTML file. In an implementation, a plurality of documents, such as an HTML file and a Microsoft Word™ document, are compared to identify differences in the Microsoft Word™ document with respect to the content in the HTML file. A person skilled in the art will appreciate that the scope of the present subject matter extends to any form or type of the source document and the comparison document.
In one implementation, the extraction module 112 extracts the content from the source document and from the comparison document. Further, the extraction module 112 is configured to extract content from different types of documents. For example, where the

source document is a newspaper article, i.e., in a newspaper type format, the extraction module 112 extracts the content column-wise. Therefore, the extraction module 112 provides and recognizes a start and a finish datum point for a particular column and subsequently, moves to the next column to extract the content.
Furthermore, the extraction module 112 is configured to extract content from documents having various types of content formatting. For example, in a case of a 2-column layout, the extraction module 112 identifies a margin between the two columns and treats them effectively as two pages.
According to the present subject matter, the extraction module 112 then parses the extracted content from the source document and the comparison document. In an example, the extraction module 112 determines a relative location of each word of the extracted content with respect to a datum point in the respective document. In an example, this datum point may be the start datum point provided during the extraction. In another example, the datum point may be provided afresh by the extraction module during the parsing. Furthermore, the extraction module 112 may parse the extracted content based on a pre-defined criterion. For example, the extraction module 112 may parse the content from top to bottom, or left to right, or any other combination thereof.
In one implementation, the data extracted by the extraction module 112 is stored in the extraction data 122. The extraction data 122 further includes rules regarding the extraction technique to be used depending on the type of document, as discussed earlier. Further, in one implementation, the fragmentation module 114 fragments the extracted and parsed content from the documents into overlapping sets of word strings having 'n' words, where 'n' can be any number of words. For example, in a case where a 1st word string has 3 words, a second word string may contain the 2nd word of the lst word string, followed by the 3rd word of the lst word string and a subsequent word from the extracted content. In an example, the subsequent word may be the subsequent word in the sentence or in the case where the sentence ends, the first word in the subsequent sentence. This may be further understood with the help of the following example:
Consider the sentence "The brown fox jumps over the grey fence". In this case, if a word string is configured to have 3 words, i.e., n=3, the fragmentation module 114 will fragment the sentence in the following manner:

The brown fox...
.. .brown fox jumps...
...fox jumps over..., etc. In one implementation, the fragments generated by the fragmentation module 114 are stored in the fragmentation data 120. Further, the fragmentation data 120 includes rules for fragmenting the content into the word strings for example the various word string configurations (where n=3, n=4 and so on).
Subsequent to the fragmentation process, the comparison module 116 compares each of the fragmented word strings form the extracted content of the source document, with each of the fragmented word strings from the extracted content of the comparison document to identify candidate match word string pairs. In an implementation, the comparison module 116 compares keywords, for example, the words shown above in italics, of each of the word strings from the source document with the keywords of each of the word strings of the comparison document within a grammatical context. In an example, the grammatical context of a keyword is assessed by the word(s) preceding the keyword and the word(s) following the keyword. In cases where the number of words is greater than 3, i.e., n>3, there may be more than one word preceding and following the keyword. Therefore, the keyword of a word string may be chosen such that the keyword is preceded and succeeded by words in the word string in order to effectively assess the grammatical context of the keyword. In the earlier example, in the word string "The brown fox", the keyword "brown" is preceded by the word "the" and followed by the word "fox". Therefore, for all instances of the keyword ''brown" in the comparison document, the comparison module 116 ensures that each instance is a match, only when preceded by the word "the" and followed by the word "fox", i.e., in the same grammatical context. The word string? that are identified as matches in the above described manner are referred to as candidate match word string pairs.
In an embodiment, the comparison module 116 calculates relative locations of the candidate match word strings of the candidate match word string pairs. The comparison module 116 ignores differences in types of formatting in the source document and the comparison document during the comparison process, but rather focuses on the information contained in the documents therein. However, the comparison module 116 may utilize the

formatting type when calculating the relative locations of the candidate match word string pairs in the documents.
In one implementation, the comparison module 116 calculates a physical distance of a printed position of the keyword of the candidate match word string in the source document from a datum point, and of a printed position of the keyword of the candidate match word string in the comparison document. Further, the comparison module 116 also takes into account a distance traversed in parsing any intervening words.
In one implementation, the comparison module 116 is configured to compare differences in punctuation, such as missing full stops (period or decimal point) in the comparison document with respect to the source document. In an example, the punctuation symbols may be regarded as characters of the word string. For example, the word "investment" in quotes may be considered as a single word in a word string including the quotes as characters of the single word. Therefore, when compared with a word string having a word 'investment', i.e., in single quotes, the result will be tagged as non-matching as the words themselves will be considered as different.
In another example, punctuation symbols may be regarded as white space delimiters as well as individual words. In said example, the word "investment" may be considered as a word string with three individual words, i.e., a quote; followed by the word investment; followed by another quote. In said example, the comparison module 116 may ascertain the grammatical context of a keyword in a candidate match word string, by including the punctuation symbols as separate words in the word string. Therefore, in an example where the word string "investment" is compared with another word string 'investment', this will result in a keyword match (keyword being investment), but will be tagged as non-matching when compared in grammatical context.
In yet another example, punctuation symbols may be regarded as white space delimiters only, and disregarded from the comparison. In said example, the word "investment" may be considered as identical to the word investment without the punctuation symbols. In an implementation, the rules for regarding the punctuation symbols in the examples provided above may be configured and stored in the other data 126. In a case where the script is of the Western type, the calculation of physical distance of a printed position of the keyword of the candidate match word string in the source

document from a datum point, and of a printed position of the keyword of the candidate match word string in the comparison document, would include total left-to-right and top-to-bottom distance covered before arriving at the printed position of the keyword of the candidate match word string. Candidate match word string pairs with the smallest matching complementary proximities are retained, and all the other candidate match word string pairs are tagged as non-matching.
In an implementation where the documents are lengthy or complex, the optimal candidate match word string pairs are determined by resetting the datum point to zero at recognizable boundaries in the documents. For example, the datum point provided during the parsing can be reset at the start of a new section or at a customer identifier, such as customer names, customer ID, policy ID, policy number, etc. Further, areas of no interest or where known differences occur, can be masked. These areas include, for example, marketing messages or production dates, etc. Such masked areas will be disregarded during the comparison process.
Further, in one implementation, the final result of all differences in the comparison document can be displayed as markup in the source document and comparison document or in a tabulation report. The computer-generated documents may be presented in their original form without any modifications, other than the presence of markup indicating the location of differences between the documents.
In one implementation, comparison data generated by the comparison module 116 is stored in the comparison data 124, such as the candidate match word strings and the non-matching word strings.
In one implementation, the relative datum points, the datum points, and rules for the parsing of the documents such as the customer identifier sections, customer name, the customer ID, the policy ID, the pollyunumber etc., are stored in the other data 126. Furthermore, the other data 126 may also be configured to temporarily store the candidate match word strings generated by the comparison module 116.
Fig. 2 illustrates a method 200 for document comparison, according to one embodiment of the present subject matter. The method 200 may be implemented in a variety of computing systems, mentioned in description of fig. 1, in several different ways. For

example, the method 200, described herein, may be implemented using the computing system 100, as described above.
The method 200, completely or partially, may be described in the general context of computer executable instructions. Generally, computer executable instructions can include routines, programs, objects, components, data structures, procedures, modules, functions, etc., that perform particular functions or implement particular abstract data types. A person skilled in the art will readily recognize that steps of the method can be performed by programmed computers. Herein, some embodiments are also intended to cover program storage devices, e.g., digital data storage media, which are machine or computer readable and encode machine-executable or computer-executable programs of instructions, wherein said instructions perform some or all of the steps of the described method.
The order in which the method 200 is described is not intended to be construed as a limitation, and any number of the described method blocks can be combined in any order to implement the method, or an altermative method, Additionally, individual blocks may be deleted from the method without departing from the spirit and scope of the subject matter described herein. Furthermore, the methods can be implemented in any suitable hardware. software, firmware, or combination thereof. It will be understood that even though the method 200 is described with reference to the computing system 100, the description may be extended to other systems as well.
The method 200 is initiated at a block 202, where the content is extracted from the source document and from the comparison document. According to the present subject matter, the manner in which the extraction of the content is facilitated depends on the formatting of the content in the document. For example, the source document may be a newspaper article in multi-column form, and the comparison document may be a Microsoft Word™ document in single column form.
According to the present subject matter, either of the source document and the comparison document may be a Rich Text Format ( RTF) document, an Advanced Function Presentation (AFP) format document, a Hyper Text Markup Language (HTML) document, a Portable Document Format (PDF) document, a spreadsheet, a plain text document, or any other document format conventionally known in the art.

The content from the documents is extracted based on a start and a finish datum point provided in the source document and the comparison document. In an example where the source document is a newspaper anicle in multi-column form, the start datum point may be provided at the beginning of a first sentence of the first column, and the finish datum point may be provided at the end of the last sentence of the first column. In this manner, the content may be extracted as if each column were a page. Similarly, in an example where the source document is a Microsoft Word™ document, a start datum point may be provided at the beginning of a first sentence of a first paragraph of a first page, and a finish datum point may be provided at an end of the last sentence of the first paragraph of the first page. At block 204, the extracted content is parsed. In an example, the extracted content may be parsed depending on predefined criteria, such as left to right, top to bottom, or any other combination. In another implementation, the parsing includes determining a location of each word of the extracted content with reference to a datum point in the document. In an example, this datum point may be the start datum point provided during the extraction. In another example, a new datum point may be provided during the parsing. In this manner, the method according to the present subject matter allows for an effective comparison between any two types of documents, independent of the type of formatting of the content in the document.
At block 206, the extracted and parsed content is fragmented into a plurality of word strings having 'n' words, where 'n' can be any number of words. In an example, the extracted and parsed content is fragmented into overlapping word strings in a manner described previously. Depending on the content, and a level of accuracy desired, a user may specify the number of words in the word string, i.e., 'n' can be any number. Of the 'n' number of words, one of the words may be specified as a keyword. In an example, where n=3, the keyword may be selected as the word occurring in the middle.
At block 208, at least one of the word strings in the source document is compared with the plurality of word strings in the comparison document to obtain candidate match word strings. In an implementation, the comparison includes comparing the keywords of each of the word strings in the source document, with the keywords of each of the word strings in the comparison document in the same grammatical context as described previously. The comparison results in a candidate match word string pair, when a keyword of a word string in

the source document, matches a keyword of a word string in the comparison document in the same grammatical context. As mentioned previously, the grammatical context of a keyword is ascertained by the word(s) preceding the keyword, and the word(s) following the keyword. At block 210, a relative location of the candidate match word string in the source document as well as the comparison document is calculated. In an implementation, the location of the word string may be determined as the physical distance of the keyword of the word string from the datum point. In an example, the datum point provided during the parsing may be utilized in this regard. In another example, another datum point may be provided, from which the distance of the keyword may be ascertained. In this example, a datum point may be provided by re-setting the datum point to zero at recognizable boundaries in the documents. For example, the datum point provided during the parsing can be reset to the start of a new section or at a customer identifier, such as customer names, customer ID, policy ID, policy number, etc.
Further, areas of no interest or where known differences occur, can be masked. These areas include, for example, marketing messages or production dates, etc. Such masked areas will be disregarded during the comparison process.
At block 212, the differences in the comparison document with respect to the source document are indicated. In an example, the candidate match word string pairs having the closest relative locations are retained, while the remaining word strings are tagged as non-matching.
The method may be repeated as desired and the final output of all differences is presented in the comparison document or in a new document. In an example, the non-matching word string pairs may be utilized to indicate differences in the comparison document with respect to the source document in, for example, a mark-up document or as a tabulation of differences.
Although implementations of content comparison have been described in language specific to structural features and/or methods, it is to be understood that the present subject matter is not necessarily limited to the specific features or methods described. Rather, the specific features and methods are disclosed as implementations for content comparison.

I/We Claim:
1. A method for document comparison, the method comprising:
obtaining content from a source document and a comparison document;
fragmenting the obtained content into a plurality of word strings, wherein each of the plurality of word strings comprises a predetermined number of words;
comparing at least one of the plurality of word strings from the source document with the plurality of word strings from the comparison document to obtain candidate match word string pairs; and
indicating differences in the comparison document with regard to the source document based on the comparison.
2. The method as claimed in claim 1, wherein the source document is at least one of a Rich Text Format (RTF) document, an Advanced Function Presentation (AFP) format document, a Hyper Text Markup Language (HTML) document, a Portable Document Format (PDF) document, a spreadsheet, and a plain text document.
3. The method as claimed in claim 1. wherein the comparison document is at least one of a Rich Text Format (RTF) document, an Advanced Function Presentation (AFP) format document, a Hyper Text Markup Language (HTML) document, a Portable Document Format (PDF) document, a spreadsheet, and a plain ;ext document.
4. The method as claimed in claim 1, wherein the obtaining the content further comprises:
extracting the content based on a start datum point and a finish datum point in the source document and the comparison document; and parsing the extracted content.
5. The method as claimed in claim 4, wherein the start datum point and the finish datum point are configured based on a start of a content column and a finish of the content column.
6. The method as claimed in claim 4, wherein the parsing further comprises determining a distance of each extracted word from a datum point in the source document and the comparison document, and wherein a position of the datum point is configurable in the source document and the comparison document.
7. The method as claimed in claim 1, wherein the fragmenting further comprises fragmenting the content into overlapping word strings.

8. The method as claimed in claim 1, wherein the comparing comprises determining whether a keyword of a word string in the source document matches a keyword of a word string in the comparison document in a same grammatical context.
9. The method as claimed in claim 1, further comprising:
measuring a relative distance of the word string of the candidate match word string pair from the datum point in the source document, and a relative distance of the word string of the candidate match word string pair from the datum point in the comparison document;
comparing the relative locations of the word strings; and
retaining those candidate match word string pairs with closest relative location.
10. The method as claimed in claim 9, further comprising tagging non-matching candidate
match word string pairs as non-matching.
11. The method as claimed in claim 1, wherein the indicating comprises indicating
differences in at least one of a source document mark up, a comparison document mark up,
and a tabulation report.
12. A computing system (100) for comparing a source document and a comparison
document, the system (100) comprising:
a processor (102); and
a memory (106) coupled to the processor (102), the memory (106) comprising:
a fragmentation module (114), configured to fragment content extracted from the source document and the comparison document, into a plurality, of word strings comprising a predetermined number of words; and
a comparison module (116), configured to compare the word strings.
13. The system (100) as claimed in claim 12, further comprising an extraction module
(112), configured to:
extract the content based on a start datum point and a finish datum point in the content of the source document and the comparison document; and parse the extracted content.
14. The system (100) as claimed in claim 13, wherein the extraction module (112) is
further configured to measure a relative location of each word of the extracted content from a
datum point.

15. The system (100) as claimed in claim 12, wherein the fragmentation module (114) is further configured to fragment the content into overlapping word strings having a predetermined number of words.
16. The system (100) as claimed in claim 12, wherein the comparison module (116) is further configured to compare the word strings based on a grammatical context.
17. The system (100) as claimed in claim 12, wherein the comparison module (116) is further configured to compare the word strings based on a relative location of the word strings from a datum point.
18. A computer-readable medium having embodied thereon a computer program for executing a method comprising:
extracting and parsing content from a source document and a comparison document;
fragmenting the obtained content into a plurality of word strings comprising a predetermined number of words;
comparing at least one of the plurality of word strings from the source document with the plurality of word strings from the comparison document to obtain candidate match word string pairs; and
indicating differences in the comparison document with regard to the source document based on the comparison.