WO2001048579A2

WO2001048579A2 - Information search and retrieval system

Info

Publication number: WO2001048579A2
Application number: PCT/IL2000/000851
Authority: WO
Inventors: Yossef Matias; Gadi Wolfman; Ehud Rivlin
Original assignee: Zapper Technologies Inc.
Priority date: 1999-12-23
Filing date: 2000-12-21
Publication date: 2001-07-05
Also published as: WO2001048579A3; AU2021701A

Abstract

A method for retrieving information from at least one information source (16). The method includes the steps of inputting words received from a user, receiving a user-selected specification (11) (selected from among a plurality of possible specifications), and retrieving the information. Each specification (11) at least provides details for the retrieval of information from the at least one information source (16). The retrieval of the information from the at least one information source (16), uses the words and the specification (11).

Description

INFORMATION SEARCH AND RETRIEVAL SYSTEM

FIELD OF THE INVENTION

The present invention relates generally to the search and retrieval of information from an information source

BACKGROUND OF THE INVENTION

The Internet contains an overwhelming amount of information about a multitude of topics and the information available continues to increase at a rapid rate However, the nature of the Internet, is that of an unorganized mass of information Navigation to a specific desired page requires either knowing the uniform resource locator (URL) for the site, having a bookmark to the site (which is actually a stored URL for the site), or successive requests for Web pages until the site is found

Hypertext transfer protocol (HTTP) is the communication protocol of the Internet Whenever a request for a World Wide Web ("Web") site is made, it is written in HTTP An HTTP request includes details such as the URL of the site, cookies (data stored on the computer of a user to track his patterns and preferences), and a referer field (the page from which the current request was made) Repeat requests are needed in most cases since, even when a user gets to a desired site, he must enter more information or choose between options a number of times, to get to the Web page containing the exact information he needs This is illustrated in Figs 1A and 1 B, schematics of Web pages, to which reference is now made In these figures, arrows denote data flow, dashed boxes denote the choice selected, and double boxes denote the button clicked

The first page, 101 , represents the US Patent and Trademark Office (USPTO) home page It shows three choices "general information", "searchable databases", and "PTO fees" Selecting "searchable databases" returns page 103, Search on the PTO Web Server This page has two choices "patent bibliographic and abstract database" and "trademark database with images" Selecting "patent bibliographic and abstract database" returns page 105, USPTO Web Patent Database. Under the heading Bibliographic Database are three choices: "boolean search", "advanced search", and "patent number search". Selecting "patent number search" returns page 107A, Patent Bibliographic Database. Filling in 5,123,456 in patent number box 108, as on page 107B, and clicking on the search box, returns page 109, Patent Bibliographic Database Results of Search, with the title of the patent "Banding Tool". If "Banding Tool" is clicked, page 111 containing the patent abstract is returned.

In this example, the user needs information about a specific U.S. patent, for example, US 5,123,456. If the user initially only has the URL of the home page of the USPTO, he has to make each of the following Web requests:

1. at the USPTO home page, select "searchable databases",

2. from the USPTO Search on the PTO Web Server page, choose "patent bibliographic and abstract database",

3. from the USPTO Web Patent Database, choose "patent number search", 4. on the Patent Bibliographic Database page, enter the patent number of interest, in the patent number box, and request a search, 5. from the Patent Bibliographic Database Results of Search page, if desired, select the patent title to see an abstract of the patent. If the user knows any of the URLs of the Web pages or has a bookmark for them, he can skip any of steps 1 - 3, as appropriate. He still has to complete step 4 of filling in patent number box 108 on page 107A and executing the search. Thus, two Web retrievals, the Patent Bibliographic Database page and the Patent Bibliographic Database Results of Search page, are required at a minimum as well as filling in patent number box 108. Many times though, the user is looking for information about a specific topic, but does not know where to find the information. In such a case, he may make use of a search engine to find sources for the information he needs. In the above example, the user could search, using any of the public domain search engines available on the Web, using criteria such as "patent" or "US patent" and would find one of the USPTO Web pages among the matching results. The more knowledgeable the user, the narrower he can make the search criteria, thus resulting in a finer tuned list of results. Additionally, there ex st some products that make shortcuts for specific searches. For example, using AltaVista Search, a search engine available from AltaVista Company at http://www.altavista.com/, the user enters "us patent 5123456" in the "Find This" box With a specific query of this type, AltaVista will return what they call an "Internet keyword", in this case U.S. patent 5123456RN, which, if clicked will return a page similar to page 111 , the USPTO Full Text and Image Database page containing the full text of the US patent. This search requires three Web retrievals, herein called "information requests", one to get the AltaVista Search page, the search itself, and the retrieval of the abstract page.

SUMMARY OF THE INVENTION

An object of the present invention is to provide an improved information retrieval and searching system and method There is provided, in accordance with a preferred embodiment of the present invention, a system for retrieving information from at least one information source The system includes an input unit, a specification unit, and a retrieval engine The input unit receives words from a user The specification unit receives a user-selected specification which provides details for the retrieval of information from the at least one information source The retrieval engine retrieves the information from the at least one information source using the words and the specification

Additionally, in accordance with a preferred embodiment of the present invention, the retrieval engine includes a unit for generating at least one information request to the at least one information source

Furthermore, in accordance with a preferred embodiment of the present invention, the specification is a script for generating information requests for at least one specific information source The specification includes variables and the words are the values for the variables used to generate the information request

Moreover, in accordance with a preferred embodiment of the present invention, the specification unit includes a unit for updating the specification from an external information source

Further, in accordance with a preferred embodiment of the present invention, the selected specification belongs to a group of specifications and the specification unit includes a unit for updating the group of specifications from an external information source Alternatively, the specification group generally forms a related unit with a common theme

Still further, in accordance with a preferred embodiment of the present invention, the system for retrieving information also includes an interface unit connected to the input unit and to the specification unit The interface unit includes the input unit, a plurality of user selectable references each referring to a different specification, and a unit for providing the specification corresponding to the selected reference to the specification unit.

Additionally, in accordance with a preferred embodiment of the present invention, an information request includes instructions to an external application. Alternatively or in addition, the external application includes a unit for manipulating the information request into a new information request and a unit for sending the new information request to the one information source.

Moreover, in accordance with a preferred embodiment of the present invention, the external application includes a unit for generating multiple information requests, a unit for manipulating the output result of the information requests of each information source, and a unit for sending the manipulated output to the user.

There is provided, in accordance with a preferred embodiment of the present invention, an interface unit operable with a system for retrieving information from at least one information source. The interface unit includes an input area where words may be entered, a plurality of user selectable references, and a unit for providing the words and the specification corresponding to the selected reference to the system for retrieving information. Each user selectable reference refers to a different specification and each specification provides details for the retrieval of information from one different information source.

Furthermore, in accordance with a preferred embodiment of the present invention, the one information source is a World Wide Web ("Web") site.

Moreover, in accordance with a preferred embodiment of the present invention, the one information request is a hypertext transfer protocol (HTTP) request.

Still further, in accordance with a preferred embodiment of the present invention, the external application is accessible on an external information source. There is also provided in accordance with a preferred embodiment of the present invention, a method for retrieving information from at least one information source. The method includes the steps of inputting words received from a user, receiving a user-selected specification, and retrieving the information. The user-selected specification at least provides details for the retrieval of information from the at least one information source. The retrieval of the information from the at least one information source, uses the words and the specification.

There is also provided in accordance with a preferred embodiment of the present invention, a system for retrieving information from at least one information source. The system for retrieving information includes an input unit for receiving words from a user, at least one specification, selectable by a user, wherein each specification at least provides details for the retrieval of information from the at least one information source, a specification unit for receiving the user-selected specification and a retrieval engine at least for retrieving the information from the at least one information source using the words and the specification.

There is also provided in accordance with a preferred embodiment of the present invention, a method for retrieving information from at least one information source. The method for retrieving information includes the steps of having at least one specification, wherein each specification at least provides details for the retrieval of information from the at least one information source, inputting words received from a user, receiving a user-selected one of the at least one specification, and retrieving the information from the at least one information source using the words and the specification.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention will be understood and appreciated more fully from the following detailed description taken in conjunction with the appended drawings in which:

Fig. 1 is a schematic illustration of prior art searches and retrievals of information using the Internet;

Fig. 2 is a block diagram illustration of an information search and retrieval system, constructed and operative in accordance with a preferred embodiment of the present invention;

Fig. 3 is a block diagram illustration of the engine of Fig. 2;

Fig. 4 is a schematic illustration of the service selector of Fig. 2, constructed and operative in accordance with a preferred embodiment of the present invention; Fig. 5 is a schematic illustration of a graphical user input device for an information search and retrieval system, constructed and operative in accordance with a preferred embodiment of the present invention; and

Figs. 6A, 6B, and 6C are illustrations of data flow during the execution of an information request system of the invention.

DETAILED DESCRIPTION OF THE PRESENT INVENTION

The present invention is a system and method for information retrieval and for searching in an information source This invention includes a number of features, for example, allowing a user to successfully retrieve information without knowledge about the resources available for use with a specific information source and, bypassing intermediate search and retrieval requests This is true for all types of information retrieval, though for the sake of clarity, examples using the World Wide Web ("Web") will be described

Reference is now made to Fig 2, which is a block diagram illustration of an information search and retrieval system, constructed and operative in accordance with a preferred embodiment of the present invention The system comprises a service selector 10, a text handler 12, and an information search and retrieval engine 14 hereinbelow referred to as the "engine" Service selector 10 contains a plurality of specifications 11 Service selector 10 passes a selected specification 11 , describing a user-selected information source, to engine 14 Selected specification 11 contains details particular to an information source 16, so that an appropriate information request can be built for the information source In addition, text handler 12 passes user-provided words defining a search to be done, to engine 14 Engine 14 then uses these two inputs to build an information request defined by the words of the user and appropriate to the user-selected information source 16 Information source 16 performs the search or retrieval and provides the results back to the user

As described hereinbelow in more detail, selected specification 11 describes the details for retrieval of the information the user wants Thus, the user does not need to know what Web page is an appropriate data source, the uniform resource locator (URL) of the page, or how to fill in the necessary data in the form on the page

Referring to the exemplary search in the US Patent and Trademark Office (USPTO) related in the Background, a useful specification 11 is one describing a patent number search using the Patent Bibliographic Database page 107 shown in Fig 1A The appropriate words would be "5,123,456" Engine 14 takes the two inputs and creates an information request which, in effect, fills in patent number box 108 on the Patent Bibliographic Database page 107A shown in Fig 1A and sends the hypertext transfer protocol (HTTP) request to the Web, all in one step The result, Patent Bibliographic Database results page 111 in Fig 1 B, is returned In the present invention, all intermediate request and retrieval steps are skipped, and the actual filling in of the search box is done by engine 14 Thus, only one Web request is issued The user does not have to know any details about the USPTO Web site or the hierarchy of pages normally traversed in an information request This is opposed to the prior art of using a bookmark, in which the user first selects the bookmark for page 107, retrieves the page, and then must fill in the search box on the retrieved page Only then can the user request the desired search

Service selector 10 contains up to N specifications 11 that the user chooses from The specifications 11 contain details telling engine 14 how to build the information request Specification 11 provides a definition of where to search, what specific resource(s) to use, and the parameters needed for the resource being used Thus, an exemplary specification 11 for patent searches of the Web, refers to a Web page, for example page 107 described hereinabove The user need not know anything about the details of the site or its specification 11 , only that he wants a certain type of data source and search

The user chooses the specification 11 , appropriate for the desired information, from among specifications 11 contained in service selector 10 Other specifications 11 could be a dictionary search or a search of Broadway show listings Specification 11 for a dictionary search would include the URL of an on-line dictionary and would include code for creating a request for the definition of a specific word Specification 11 for a Broadway show listing would include the URL of a site listing Broadway shows and code for selecting the requested show and finding its Web page

Specifications 11 can be updated as changes are made to the underlying Web pages, but this does not affect the user, since these details are hidden within the specifications 11 Using text handler 12, the user specifies the information he wants. In the USPTO patent search example, the words would minimally be a patent number, whereas in the dictionary example they would include the word(s) whose definition the user wants. Additionally, the information could be taken from a document and could include words alluding to the context within which the information is described. For example, in the dictionary example, if the unknown word is "negotiator", other information could limit the definition to the field of law. An exemplary system which utilizes context is described in the US provisional patent application 60/161 ,712, filed October 27, 1999 and entitled "Heuristic Context-Directed Search", and in the US provisional patent application 60/171 ,586, filed December 23, 1999 and entitled "Autonomous Context-Driven Search", both owned by the common assignee of the present invention and incorporated herein by reference. These applications have been converted to US patent applications 09/524,569 and 09/568,988. Text handler 12 outputs words that describe the information wanted. Text handler 12 can receive input from any application integrated into the computer desktop. Among other applications, it could be a document in the Microsoft Word editor (available from Microsoft Corporation of Redmond, WA, USA), a Web page, email, or text entered manually in the user interface of text handler 12. Fig. 3, to which reference is now made, is a block diagram illustration of engine 14 of Fig. 2. Engine 14 comprises a specification interpreter 20, a request engine 24, and a request object 26.

Specification interpreter 20 manages engine 14 and receives words and selected specification 1 1. Specification interpreter 20 interprets the lines of selected specification 1 1 and provides the relevant information to request engine 24. This is described in detail with respect to an exemplary specification 11 hereinbelow.

The instruction code contained in specification 1 1 may be written in any language that specification interpreter 20 understands. In a preferred embodiment of the current invention, code including "key words" in an Extensible Markup Language (XML) wrapper around JavaScript is used. The code describing the instructions for the exemplary specification 11 "patent number" is the following:

Code 1

1. <guidelet title="Patent Number Search" url=" http://www.guidelet.eom/library/patent_guidepack#Patent.Number.Search">

2. <desc>

3. Patent Number Search

4. </desc>

5. <script lang="javascript"><![CDATA[function guidelet(text,context) 6. {

7. if (! text.search(/^{Λ *}$/))

8. return new

HttpRequest("http://128.109.179.23/access/search-num.html"); 9. 10. var req = new

HttpRequest("http://128.109.179.23/cgi-bin/num_srch4?INDEX 0");

11. req.setMethod("POST");

12. req.addParam("DBSELECT2=" + "ALL");

13. req.addParam("RANKTYPE=" + "CHRON"); 14. req.addParam("ELEMENT_SET=" + "OF");

15. req.addParam("QUERY=" + text);

16. return req;

17. }]]></script>

18. </guidelet>

Lines 1 - 4 of Code 1 are used by specification interpreter 20. The title is used to designate specification 11 to the user and the URL of line 1 refers to the site from which specification 11 will be refreshed, as described hereinbelow.

Specification interpreter 20 passes the instructions, in lines 5 - 16 of the code, and the words, as the text and context parameters of line 5, to request engine 24 to compose request object 26, which is an abstract object encapsulating the request details. Request engine 24 can use any available scripting engine. In a preferred embodiment of the present invention, request engine 24 interprets Java Script. It interprets and runs the function "guidelet" referred to in line 5, instantiating, in line 10, a "req" object that is an instance of an HttpRequest object, and adding its parameters as defined in lines 1 1 - 15. The resulting "req" object is request object 26 that is returned to specification interpreter 20 in line 16 of the code.

A new request object 26 is instantiated each time specification interpreter 20 is run, making use of the specific words and selected specification 1 1. In the above example, request engine 24 creates an object making use of parameters DBSELECT2, RANKTYPE, ELEMENT_SET, and QUERY.

Specification interpreter 20 then uses the contents of request object 26 to construct an information request appropriate to the information source. For example in the case of the Web, the information request would be written in HTTP. The resulting HTTP code for the patent number example would be:

Code 2

19. POST http://128.109.179.23/cgi-bin/num_srch4?INDEX+0 HTTP/1.0

20. Referer: http://128.109.179.23/access/search-num.html

21. Proxy-Connection: Keep-Alive 22. User-Agent: Mozilla/4.7 [en] (WinNT; I)

23. Host: 128.109.179.23

24. Accept: image/gif, image/x-xbitmap, image/jpeg, image/pjpeg, image/png, ^*/^*

25. Accept-Encoding: gzip 26. Accept-Language: en,pdf

27. Accept-Charset: iso-8859-1 ,^*,utf-8

28. Content-type: application/x-www-form-urlencoded

29. Content-length: 57 30. 31. DBSELECT2=ALL&RANKTYPE=CHRON&ELEMENT_SET=OF&QUER

Y=5123456 The URL of line 10 (Code 1 ) and the parameter of line 1 1 of specification 11 appear in line 19 (Code 2) of the HTTP information request. The various parameters of lines 12 - 15 (Code 1 ) of specification 1 1 are filled in as well, appearing in line 31 (Code 2) of the HTTP information request. Note that the request includes ' "QUERY- ' + text' from line 15 (Code 1 ) of specification 1 1 in which "text" is substituted with "5123456", the words specification interpreter 20 receives as input.

Code lines 1 - 18 of Code 1 hereinabove describe the instructions for a simple specification 11. In a further preferred embodiment of the present invention, code lines 32 - 82 of Code 3 below, describe the instructions for a specification 1 1 in which specification interpreter 20 first manipulates the information on the client machine. Specification interpreter 20 then sends an information request to information source 16 incorporating the results of the manipulations. In the exemplary specification 1 1 , the domain name is extracted from a user given URL, and an information request for information about this domain name is made.

Code 3 32. <guidelet title="" url="http://www.guidelet.com/library/domain_guidepack#Domain.Search"> 33. <script lang="javascript"><![CDATA[ function getNameDomain( url )

34. {

35. var text = new String( url );

36. var textl = new Sthng(); 37. var result = new Object();

38. var supported_domains = new Array( "com", "org", "net" ); 39.

40. var re = /[ \t\n]+/;

41. text = text.replace( re, "" ); // get rid of white spaces 42. var re = /.^*W/;

43. text = text.replace( re, "" ); // get rid of everything before "//" 44. re = Λ/.^*/;

45. text = text.replace( re, "" ); // get rid of everything after "/"

46. textl = text.toLowerCaseO;

47. var Arr = textl .split("."); 48. result.name = "";

49. result.domain = "";

50.

51. if ( Arr[0] != "" ){ // not empty string

52. if ( Arr.length < 2 ){ // single word only 53. result.name = Arr[0];

54. result.domain = "com";

55. }

56. else { // Arr.length >= 2

57. var domain = Arr[ Arr.length - 1 ]; 58. result.domain = "";

59. for ( i = 0 ; i < supported_domains. length ; i++ ){

60. if ( domain == supported_domains[i] ){

61. result.domain = domain;

62. } 63. }

64. result.name = ( result.domain != "" )? Arr[ Arr.length - 2 ] : "";

65. }

66. } 67. return result;

68. } 69.

70. function guidelet(text,context)

71. { 72. var req = new

HttpRequest("http://www.dctcomdirectory.com/nsi/domainsearch.hm"); 73. req.setMethod("GET"); 74. req.addParam("st=" + "basic");

75. var result = getNaneDomain( text );

76. req.addParam("Mc inDom=" + result.name );

77. req.addParam("Ex:Dom=" + result.domain );

78. req.addParam("image2.x=" + "19");

79. req.addParam("image2.y=" + "4");

80. return req;

81. }]]></script>

82. </guidelet>

Only code that describes the use of the given URL to find the domain name will be discussed. On lines 33 - 68 a new function "getNameDomain" with a single parameter, the URL, is introduced. A list of supported generic top level domains appears on line 38. The URL is split into its various components on lines 40 - 50. In lines 51 - 66, the member data of the result object are filled with the Web site name and the domain.

Lines 70 - 80 are similar to the function guidelet in Code 1, line 5 hereinabove. The function "getNameDomain" is called on line 75 with the text parameter set to the URL the user entered. Using the returned object "result" from line 75, new parameters, such as, "result.name" and "result.domain" can be added to "req", request object 26, on lines 76 and 77.

It will be appreciated that the complexity of the manipulation of the text and context inputs on the client can be as complicated as the writer of specification 11 makes it. There is no inherent limit in specification interpreter 20 regarding the complexity of specification 11.

It will be further appreciated, that in a further preferred embodiment of the present invention, parameter and result manipulation can be performed on the server machine as described hereinbelow with respect to Figs. 6A - 6B.

Reference is now made to Fig. 4, a schematic illustration of service selector 10 of Fig. 2, constructed and operative in accordance with a preferred embodiment of the present invention. Service selector 10 comprises a specification handler 32 and an interface unit 40. Interface unit 40 comprises one or more file references 42, each file reference 42 comprising one or more specification references 44. File reference 42 refers to a file containing related specifications 11. Specification reference 44 refers to the section of code within the file that encapsulates details of one information request, in other words, one specification 11.

Code lines 83 - 134 (of Code 4 below) represent the contents of one exemplary file containing two different specifications 11 , one titled "Mega Search" and the other "Golden Retriever".

Code 4 83. <?xml version="1.0" encoding="utf-8"?>

84. <!DOCTYPE guidepack SYSTEM "guidepack.dtd"> 85.

86. <guidepack version="2.1"

87. title="Server Guidepack" 88. url="http://dir.guidelet.com/services/Provide?name=guidelet">

89.

90. <html src=""/>

91.

92. <hier> 93. <glref label- 'Mega Search"

94. url="Combined.Search'7>

95. <glref label- 'Golden Retriever"

96. url="Golden.Retriever"/>

97. </hier> 98.

99. <guidelet title- 'Combined Search" url="Combined.Search">

100. <desc>

101. Queries a combination of search engines,

102. using your text and context. 103. </desc>

104. <script lang="javascript"><![CDATA[function guidelet(text,context,location) 105 {

106 var req = new

HttpRequest("http //dir guidelet com/servιces/AlgStage1 "),

107 req setMethod("POST"),

5 108 req addParam("text=" + text),

109 req addParam("context=" + context),

1 10 req addParam("locatιon="+ location),

1 1 1 req addParam("engιne=mιxed"),

1 12 req addParam("form=ιnιt"), 10 1 13 return req,

1 14 }]]></scrιpt>

1 15 </guιdelet>

1 16 <guιdelet title- 'Golden Retriever" url="Golden Retrιever">

1 17 <desc> i s 1 18 We select an engine, according to your

1 19 word(s) and context, and search it

120 </desc>

121 <scrιpt lang="javascπpt"^χi[CDATA[functιon guιdelet(text,context,locatιon) 0 122 {

123 var req = new

HttpRequest("http //dir guidelet com/servιces/AlgStage1 "),

124 req setMethod("POST"),

125 req addParam("text=" + text), 5 126 req addParam("context=" + context),

127 req addParam("locatιon="+ location),

128 req addParam("engιne=sr"),

129 req addParam("form=ιnιt"),

130 return req, 0 131 }]]></scrιpt>

132 </guιdelet> 133 134. </guidepack> 135.

In the code above, code lines 83 - 98 and 134 are used to describe file reference 42. Line 87 gives the title of file reference 42 used to designate this grouping of specifications 1 1 to the user. Each specification reference 44 refers to a specific section of the code. Lines 92 - 97 show the hierarchy of specification references 44 included in the file. In this example, there are two specification references 44: "Mega Search" and "Golden Retriever", as given by lines 93 and 95. Lines 94 and 96 give a "url" value which is matched in lines 99 and 1 16 to find the specification 1 1 code relating to the specific specification reference 44. Specification 1 1 , "Mega Search" is described on lines 99 - 1 15 whereas specification 1 1 , "Golden Retriever" is described on lines 1 16 - 133. This grouping of related specifications 1 1 into files enables specification handler 32 to manage groups of specifications 1 1 as a unit as described hereinbelow.

When a user chooses a specification reference 44 using interface unit 40, it passes the selected specification reference 44 to specification handler 32 as shown by the arrow labeled "selected reference". Specification handler 32 requests specification 11 corresponding to specification reference 44 from a local storage, as shown by arrow 34. The local storage contains a plurality of local specifications 35 and returns the local specification 35 corresponding to specification reference 44, as shown by arrow 38. The returned local specification 35 is used in specification handler 32 as selected specification 1 1.

Specifications 1 1 can require management, for example, to update or add to currently available specifications 1 1. For example, specification 1 1 must be updated whenever changes are made to its associated information source. Further, if specification 1 1 refers to a broad type of search for which multiple sites exist (like the dictionary example), as better resources become available, the associated specification 1 1 can change to use these resources instead. There are two modes for updating specification 1 1. It can be done whenever a user chooses a specification 1 1 using interface unit 40 or it can be done in the background, by a process invoked by specification handler 32. Updates are retrieved from an information source that contains a multiplicity of remote specifications 37.

In the first update mode, wherein an update is performed whenever a user chooses a specification 11 , after retrieving local specification 35 as shown by arrow 38, specification handler 32 queries the original information source, typically through the Web, as shown by arrow 36. If a corresponding remote specification 37 does not have the same version number as selected specification 11 , remote specification 37 is retrieved, as shown by arrow 39, and replaces selected specification 11. The retrieved remote specification 37 is then stored in the local storage as well, as shown by arrow 34, replacing the appropriate local specification 35. This updated selected specification, now labeled 11 , is the input to specification interpreter 20 of engine 14.

In the second update mode, wherein the update is done in the background, at a specified time, specification handler 32 performs a process similar to that just described. For each file being updated, the version number of its associated remote specification 37 is checked against the version number of local specification 35 and replaced as necessary.

It is noted that updates can be performed on files as a unit, updating all specifications 11 contained in a given file at once. This provides a relatively simple update process.

Specification 11 may be written for an individual user, for a local group of users, or for a remote group of users. Specifications 11 can be obtained locally or from other users or suppliers, for example, from a Web site. When a user finds a specification 11 of interest, it is stored by specification handler 32 as local specification 35.

Fig. 5, to which reference is now briefly made, is a schematic illustration of a graphical user input device 50, useful in the system of the present invention. Graphical user input device 50 consists of an input area 51 , a graphical interface unit 40', and a description box 55. Input area 51 is available to receive the words the user has selected or has typed. Graphical interface unit 40' has a number of graphical file references 42' listed. Those graphical file references 42' containing several graphical specification references 44' contain an arrow 53. Selecting one of graphical file references 42' containing arrow 53, for example "dictionaries", causes a pull-down list of its graphical specification references 44' to appear. When a specific graphical specification references 44' is selected, its description box 55 appears. Description box 55 provides a "descriptive tip" to the user, a short sentence describing the use of graphical specification reference 44'. For example, description box 55 might read "Enter word/phrase. Get meaning" for an exemplary graphical specification reference 44' called "Dictionary.com".

Figs. 6A, 6B, and 6C, to which reference is now made, are illustrations of data flow during the execution of an information request in the information search and retrieval system of the invention. Depending on the type of the search required, as defined in specification 1 1 , there are three possible data flow paths. Some searches make use of a server 47 on a remote information source. Server 47 performs special operations and manipulates query parameters creating new information request queries. Fig. 6A illustrates the simplest type of information request, such as the patent search described hereinabove. In this search, all manipulation and interpretation of the text, context, and specification can be accomplished within engine 14. Thus, the information request is sent directly to a specific information resource 46 and the results are returned. Figs. 6B and 6C illustrate other types of information requests.

Refinement of the request requires information on server 47, which is accessed on a remote information source. Lines 106 and 123 of code section 4 instruct engine 14 to send the information request to server 47 first. Examples requiring such a data flow model would be the use of a history file (which, for example, keeps a record of instructions on a computer) or a database relating to the context of the search being used to cause a search redirection or a more refined search.

In Fig. 6B, engine 14 produces an information request, which is sent to server 47. Server 47 uses the information available to it locally, for example the history file, to create a new information request that is sent to an external information resource 46. Information resource 46 then outputs the results. Specification reference 44, "Golden Retriever", detailed in lines 1 16 - 133 of Code 4 above, describes this type of search. Line 128 of the code indicates to the server what type of manipulation of the parameters is necessary.

Fig. 6C illustrates the most complex of the data flows. This is an example of meta searching. Such an algorithm is described in the above mentioned US provisional patent application 60/171 ,586, entitled "Autonomous Context-Driven Search". In this case, the information request is passed to server 47, in a manner similar to that explained for Fig. 6B. Server 47 creates one or more new information requests to be executed on one or more different information resources 46 as shown by arrows 48. The results of the information request(s) are returned to server 47, shown by arrows 49, and further manipulated, possibly resulting in further information requests to the same or different information resources 46. When all information requests are completed, server 47 manipulates or combines the intermediate search results and returns the final results to the user. An example is a program on server 47 that uses the text and context to perform several searches, and combines the results. Specification reference 44, "Mega Search" described on lines 99 - 115 of Code 4 above describes this type of search. Line 1 1 1 of the Code 4 indicates to the server what type of manipulation of the parameters is necessary and is different than in line 128 of Code 4. It will be appreciated by persons skilled in the art that the present invention is not limited by what has been particularly shown and described herein above. Rather the scope of the invention is defined by the claims that follow:

Claims

1. A system for retrieving information from at least one information source, the system comprising: an input unit for receiving words from a user; specification means for receiving a user-selected specification at least providing details for the retrieval of information from said at least one information source; and a retrieval engine at least for retrieving said information from said at least one information source using said words and said specification.

2. A system according to claim 1 and wherein said retrieval engine comprises means for generating at least one information request to said at least one information source.

3. A system according to claim 1 and wherein said specification is a script for generating information requests for at least one specific information source, wherein the specification has variables therein and said words are the values for said variables used to generate an information request.

4. A system according to claim 1 wherein said specification means includes means for updating said specification from an external information source.

5. A system according to claim 1 wherein said selected specification belongs to a group of specifications and wherein said specification means includes means for updating said group of specifications from an external information source.

6. A system according to claim 5 and wherein said specification group generally forms a related unit with a common theme.

7. A system according to claim 1 and also comprising an interface unit connected to said input unit and to said specification means, the interface comprising: said input unit; a plurality of user selectable references each referring to a different specification; and means for providing the specification corresponding to a selected reference to said specification means.

8. A system according to claim 2 and wherein said at least one information requestjncludes instructions to an external application.

9. A system according to claim 8 and wherein said external application includes means for manipulating said information request into a new information request and means for sending said new information request to said at least one information source.

10. A system according to claim 9 and wherein said external application includes: means for generating multiple information requests; means for manipulating the output result of said information requests of each said information source; and means for sending said manipulated output to said user.

1 1. An interface unit operable with a system for retrieving information from at least one information source, the interface unit comprising: an input area where words may be entered; a plurality of user selectable references each referring to a different specification and each specification providing details for the retrieval of information from one different information source; and means for providing said words and the specification corresponding to said selected reference to said system for retrieving information.

12. A system according to claim 1 and wherein said at least one information source is a World Wide Web ("Web") site.

13. A system according to claim 2 and wherein said at least one information request is a hypertext transfer protocol (HTTP) request.

14. A system according to claim 8 and wherein said external application is accessible on an external information source.

15. A system according to claim 11 and wherein said at least one information source is a Web page.

16. A method for retrieving information from at least one information source, the method comprising the steps of: inputting words received from a user; receiving a user-selected specification at least providing details for the retrieval of information from said at least one information source; and retrieving said information from said at least one information source using said words and said specification.

17. A method according to claim 16 and wherein said step of retrieving comprises the step of generating at least one information request to said at least one information source.

18. A method according to claim 16 and wherein said user-selected specification is a script for generating information requests for at least one specific information source wherein the specification has variables therein and said words are the values for said variables used to generate said at least one information request.

19. A method according to claim 16 and wherein said step of receiving includes the step of updating said specification from an external information source.

20. A method according to claim 16 and wherein said selected specification belongs to a group of specifications and wherein said step of receiving includes steps of updating said group of specifications from an external information source.

21. A method according to claim 16 and wherein said specification group generally forms a related unit with a common theme.

22. A method according to claim 18 and wherein said information request includes instructions to an external application.

23. A method according to claim 22 and including the steps of manipulating said information request in said external application into a new information request and sending said new information request to said at least one information source.

24. A method according to claim 22 further including the following steps performed by said external application: generating multiple information requests, each to one said information source; manipulating the output of each said information source; and sending said manipulated output to said user.

25. A method according to claim 16 and wherein said at least one information source is a World Wide Web (Web) site.

26. A method according to claim 17 and wherein said at least one information request is a hypertext transfer protocol (HTTP) request.

27. A method according to claim 22 and wherein said external application is accessible on an external information source.

28. A system for retrieving information from at least one information source, the system comprising: an input unit for receiving words from a user; at least one specification, selectable by a user, wherein each specification at least provides details for the retrieval of information from said at least one information source; specification means for receiving said user-selected specification; and a retrieval engine at least for retrieving said information from said at least one information source using said words and said specification.

29. A method for retrieving information from at least one information source, the method comprising the steps of: having at least one specification, wherein each specification at least provides details for the retrieval of information from said at least one information source; inputting words received from a user; receiving a user-selected one of said at least one specification; and retrieving said information from said at least one information source using said words and said specification.