Chinook Licensing DE LLC v. Facebook Inc.

Filing 1

COMPLAINT FOR PATENT INFRINGEMENT filed with Jury Demand against Facebook Inc. - Magistrate Consent Notice to Pltf. ( Filing fee $ 400, receipt number 311-1444250.) - filed by Chinook Licensing DE LLC. (Attachments: # 1 Exhibit A, # 2 Civil Cover Sheet)(dmp, )

Exhibit A

US007047482B1

(12) United States Patent

(10) Patent N0.:

Odom
(54)

(45) Date of Patent:

AUTOMATIC DIRECTORY

6,584,468 B1

SUPPLEMENTATION

2001/0039563 A1*
2003/0195877 A1*

(76) Inventor: Gary Odom, 15505 SW. Bulrush La.,
Tigard, OR (Us) 97223
(*)

Notice:

Subject to any disclaimer, the term of this
patent is extended or adjusted under 35
U.S.C. 154(b) by 623 days.

(22) Filed:

Tian ......................... .. 709/202
Ford et a1. ................... .. 707/3

(Aug. 2001).*

(200601)

(200601)

Uncertain and Changing Data” (Jun. 24-28, 2001).*

US. Cl. ............................. .. 715/500; 707/5; 707/3

Grasso, Antonietta et a1“ GROUP 99 in phoenix’ Arizona;

0f Classi?cation Search ........... ..

“Augmenting Recommender Systems

715/500; 707/3, 5; 709/202
See application ?le fOr complete Search hiSIOI'Y(56)

6/2003 Gabriel
11/2001
10/2003

Balabanovic, Marko et al., Communications of the ACM,
vol. 30, No. 3; “Content-Based, Collaborative Recommen
dation” Mar. 1997).*
Geisler, Gary et al., JCDL ’01 in Roanoke, Virginia; “Devel
oping Recommendation Services for a Digital Library With

Feb- 28: 2001

(51) Int‘ Cl‘
G06F 7/76
G06F 17/21

May 16, 2006

OTHER PUBLICATIONS
Maglio, Paul et al., Communications of the ACM, vol. 43,
No. 8; “Intermediaries Personalize Information Streams”
(Aug. 2000).*
Lieberman, Henry et al., Communications of the ACM, vol.
44, No. 8; “Exploring the Web With Reconnaissance Agents”

(21) Appl. No.: 09/796,235

(52)

US 7,047,482 B1

Inter

faces into Practices” (1999).*
Voss, Angi et al., GROUP 99 in Phoenix, Arizona; “Concept

Indexing” (1999).*

References Cited

* cited by examiner
U.S. PATENT DOCUMENTS

5,535,382
5,598,557
5,680,511
5,893,092
6,100,890
6,122,647
6,138,157
6,182,133
6,184,886
6,480,853
6,493,702

A
A
A
A *
A *

Primary ExamineriDoug Hutton

7/1996 Ogawa
1/1997 Doner
10/1997 Baker

(57)

4/1999 Driscoll ....................... .. 707/5
8/2000 Bates et a1. ..
715/826

A
A
B1
B1

9/2000
10/2000
1/2001 Horvitz
2/2001 Bates .... ..

B1 *
B1 *

11/2002
12/2002 Adar et a1

Horowitz

Welter

Jain

. . . . .

. . . . .

. . . ..

. . . ..

ABSTRACT

The present invention is computer software that automati

cally ?nds, saves, and displays links to documents topically

707/513

related to document links residing in a directory Without a
user having to search.

709/224

709/223
.. 345/357

. . . . . . . . . . . .

. . . ..

.................. ..

707/5

20 Claims, 5 Drawing Sheets

707/3

10 ENABLE DIREcToRY SUPPLEMENTATION

10l SET BREADTH LEVEL

+
1 1 COLLATE KEYWORDS

9 DERIvE KEYwoRI:>(s)
l 1 1 COMPA RE DOCUMENT

KEYwoRDs

1 2 SEARCH

120 Fmo NEW PAGES

Ill
86 CULL DISCARDED
INKS

9 DERIVE KEYwoRD(s)

I
121 COMPARE KEYwoRDS

122 RANK NEw PAGES

+
6 SUPPLEMENT DIRECTORY
66 SIGNIFY LINK

88 DIREcToRY
KEYWORDS

U.S. Patent

May 16, 2006

US 7,047,482 B1

Sheet 1 0f 5

50 COMPUTER
51 CPU

52 STORAGE
53 MEMORY
54 RETENTION

DEvIcEs(s)
60 NETWORK COMPUTER

55 DISPLAY DEVICE

61 CPU

56 INPUT DEvICE(s)

62 STORAGE

57 POINTINO DEVICE

(E.G. MOUSE)

63 MEMORY
68 NETWORK

(CONNECTION)
58 KEYBOARD

64 RETENTION

DEvICEs(s)

59 NETWORK
CONNECTION DEVICE

69 NETWORK
CONNECTION DEVICE

FIGURE 1

U.S. Patent

May 16, 2006

Sheet 2 0f 5

US 7,047,482 B1

5 DIRECTORY
TITLE
3 DIRECTORY

2A DOCUMENT

2B DOCUMENT

FIGURE 2

U.S. Patent

May 16, 2006

Sheet 3 0f 5

2 DOCUMENT

US 7,047,482 B1

20D DOCUMENT TITLE

20 TITLE —

2GP PAGE PROPERTIES

21 HEADING

TITLE

22 BODY TEXT

—

21 A HEADING

23 MEDIA TEXT

23T MEDIA TITLE

22A BODY TEXT

23c MEDIA CAPTION

FIGURE 3

9 DERIvE KEYwORD(s)
9D DISCERN

KEYwORD(s)

9P RANK
KEYWORDS

FIGURE 4

U.S. Patent

May 16, 2006

Sheet 4 0f 5

US 7,047,482 B1

10 ENABLE DIRECTORY SUPPLEMENTATION
101 SET BREADTH LEVEL

1
11 COLLATE KEYWORDS

9 DERIvE KEYWORD(S)

110 DERIvE

+

TITLE

KEYWORDS

111 COMPARE DOCUMENT
KEYWORDS

112 RANK KEYWORDS

12 SEARCH
120 FIND NEW PAGES

r

86 CULL DISCARDED

w

LINKS

9 DERIvE KEYWoRD(s)

+
121 COMPARE KEYWORDS

V
4—

88 DIRECTORY

122 RANK NEW PAGES

6 SUPPLEMENT DIRECTORY

| 66 SIGNIFYLINK \

FIGURE 5

U.S. Patent

May 16, 2006

KING CRIMSON

Q

Sheet 5 0f 5

US 7,047,482 B1

3K DIRECTORY

MUSIC GUIDE — KING CRIMSON
1K KNOWN LINKS

6 DISCIPLINE GLOBAL MOBILE
13 OBSOLETE LINK

/® ~I’i-RUBTBQGNG-GRIMSON

/\® KING CRIMSON DISCOGRAPHY

6K DIRECTORY
SUPPLEMENTATION

66 SIGNIPY LINK BY \_/\®

ELEPHANT TALK

33 RELEVANCE

\//@ KING CRIMSON LIvE!
FIGURE 6

/
1P FOUND LINKs

US 7,047,482 B1
1

2

AUTOMATIC DIRECTORY
SUPPLEMENTATION

phological analysis reduces verbs and nouns to their base
form, providing a basis for direct Word matching. At least

TECHNICAL FIELD

one commercial product, LinguistX® from Inxight Soft
Ware, provides advanced natural language text analysis.
In spite of softWare sophistication, as every experienced

The present invention relates generally to information
retrieval systems, and more particularly, to automatically
?nding and displaying related document links Without user

Web user knoWs, user-initiated keyWord search can be

initiated searching.

example, thinking time pieces, and you’ll likely end up With

vexing: searches commonly return a plethora of pages, many
unrelated to the desired topic. Search for ‘Watch’, for
a bushel of pages about voyeurism. Careful application of
search terms yields more relevant links, but the process and
results are problematic: beyond searching for “this ‘and’

BACKGROUND OF THE INVENTION
The Internet has become the World’ s information retrieval

intranet) documents is the use of embedded document links.

that”, search Boolean logic is not exactly intuitive; different
search engines have different syntaxes for search Boolean
logic, and different Ways to apply it, making that bit of

Such a link is a portion of a source document that links to
a target document: another document, or a different section
of the same document. The other document may be on any

business even less amenable; a bit of search pruning still
leaves an abundance of junk, While a search result leaving
out the chaff probably leaves out a good bit of Wheat too.

system. One of the distinguishing features of Internet (and

computer system on a netWork supporting the appropriate
communication protocols. Selecting a link navigates from

The technology of document linking, search, and soft
20

the source document to the target document.
A Web site is a collection of linked documents accessible

through the World Wide Web, a part of the Internet. Such
documents are commonly called Web pages. Typically a Web
site has a “home page” that is the entry document into the
site. The World Wide Web is commonly referred to as “the
Web”.
Web pages commonly use a description language such as

HTML (hypertext markup language) or XML (extensible
markup language) to embed links and provide document

Ware-based linguistic analysis are Well established. Recent
advances enhance utility in locating desired information. For

example, the subject of US. Pat. No. 6,122,647 is dynami
cally linguistically analyZing the text of a user-selected
portion of a target document and generating neW links to
25

related documents. The subject ofU.S. Pat. No. 6,184,886 is
alloWing a user to generate and maintain a list of prioritized
bookmarks (links) that alloW later access to selected sites

(documents). The subject of US. Pat. No. 6,182,133 is
pre-fetching pages for later vieWing, thus saving a user time
30

retrieving documents.

formatting.
SUMMARY OF THE INVENTION

A link on a Web page is by convention expressed as a
uniform resource locator (URL). A link is often associated
With a Word or phrase in a source document, hence the
common nomenclature: hypertext link. But a link may also

The present invention automatically ?nds, saves, and
35

be associated With images, or controls such as buttons,
menus, and the like.
A Web broWser is a program for displaying Web pages.

displays links to documents topically related to a set of
documents Without a user having to search or specify search

terms. An incidental aspect of the invention is automatically

signifying links by their status.

Examples of popular Web broWsers include Microsoft Inter

net Explorer and Netscape Navigator.

40

BRIEF DESCRIPTION OF THE DRAWINGS

Web broWsers alloW users to create and maintain direc

FIG. 1 is a block diagram of computers suitable for

tories of Web page links. Such directories are commonly
represented as folders or, sometimes, tabs.
NeW Web pages or Web sites are commonly found by links

in knoWn documents, or by keyWord search. Users typically

practicing the invention.
45

topically group links to related documents in self-titled

directories, the directory title being the common topic of
links Within it.

FIG. 5 depicts the directory supplementation process.
FIG. 6 depicts an example of directory supplementation.

Web sites are often extensive enough (so many pages) that
a site typically offers a search facility for the site; commer
cial Web sites almost alWays offer site search. Search refers
to inquiry based upon one or more keyWords (search terms).
Search engines that search a multitude of sites abound on the

Web. A good search engine provides a commercial advan
tage. Some search engines, and some commercial products,
such as Copemic® from Copemic Technologies, tap into

50

DETAILED DESCRIPTION OF THE
INVENTION

55

meaning of Words by examining Word relationships. Mor

FIG. 1 is a block diagram of a computer 50 connected to
a netWork computer 60 through a netWork 68. A computer 50

comprises at least a CPU 51; storage 52, Which comprises

multiple search engines to conglomerate searches.
Based upon keyWords, quality search engines glean the
most probably related pages using a con?uence of linguistic
analysis methods. Word location analysis is based upon the
assumption that the topic of a document is speci?ed in the
title, headings, or the early paragraphs of text. Word fre
quency analysis counts the number of times search terms
appear in a document. Syntactic analysis processes the
grammatical structure of a document, serving to indicate
nouns and verbs. Semantic analysis interprets the contextual

FIG. 2 depicts a directory of links.
FIG. 3 depicts a document.
FIG. 4 depicts the process to derive keyWords from a
document.

memory 53 and optionally one or more devices With reten

tion medium(s) 54 such as hard disks, diskettes, compact
disks, or tape; an optional display device 55; and optionally
60

one or more input devices 56, examples of Which include but
are not exclusive to, a keyboard 58, and/or one or more

pointing devices 57, such as a mouse. A computer 50 also
optionally includes a device for connection to a netWork 59.

AnetWork computer 60 comprises at least a CPU 51; storage
65

52, Which comprises memory 53 and optionally one or more
devices With retention medium(s) 64 such as hard disks,
diskettes, compact disks, or tape; and a device for connec

US 7,047,482 B1
3

4

tion to a network 59. In one embodiment, a computer 50 is
a client to a network computer 60 that is a server. A
client-server environment is a setup whereupon one or more

a web page 2. In such an instance, software linguistic
analysis of the noti?cation can determine document 2
removal.

clients 50 are connected to one or more servers 60 through

Titles 20, including document title 20D, and associated
page properties title for web pages 20P, media object titles
23T, and headings (section titles) 21 are prime fodder for

a network 68. A client 50 in a client-server environment

primarily receives data. A server 60 primarily transmits data
to be received by one or more clients 50. A peer-to-peer
network is a setup whereupon one or more computers 50 are

keywords. For a document 2 with a link 1, the link title 7

should also be considered for keyword derivation 9. Titles
may be considered highly indicative of document topics/
keywords 8. Likewise document headings 21, which can be

connected to one another 60 with or without a server on the

network 68. A computer 50 in a peer-to-peer environment
shares data with other computers 60. A network 68 may be

identi?ed by location, possibly font formatting, and isolation

any means by which one or more computers 50 are con

from body text 22; headings 21 in HTML documents are

nected to one or more other computers 60 for data transfer.

most always distinguished by font formatting, hence, easily

As depicted in FIG. 2, a directory 3, if not empty,

identi?ed.
Body text 22 may provide the bulk of information upon
which keywords 8 are derived 9. A common technique is to

comprises a set of documents 2, or a set of links 1 to
documents 2, or a combination of documents 2 and links 1.
A link 1 is a reference to a document 2. A user-determined

highly regard the ?rst paragraph of body text 22 (and the

directory title 5 may provide concise topic indication.

body text 22 immediately following headings 21) for key

FIG. 3 depicts a document 2 to which a link 1 may refer,

and document 2 components. A document 2 comprises at

word derivation 9, as the topic of a document 2 or section is
20

least a passage of text 22, and may optionally include one or

typically revealed in the ?rst paragraph (academically

more titles 20, section headings 21, or adjunctive text such

known as the “topic paragraph”).
Once a document 2 has been analyZed and keywords

as media titles 23T or captions 23c. A document 2 may

discerned 9D, document 2 keywords 8 can be rated or ranked

comprise other components besides text, such as media
objects. A media object is a non-text software entity,
examples of which include a picture, video, or sound. Text

25

ondarily in headings 21; prominence and frequency in topic

related to a media object is media text 23.

paragraphs and media text 23. Otherwise, word frequency
may be a primary keyword 8 indicator. A suggested method

FIG. 4 depicts keyword derivation 9. A keyword 8 is one
or more words used as an indication of the contents of a

document. A keyword 8 may be a combination of words: for
example, the Grateful Dead are signi?cantly different than

being either grateful or dead.
Various linguistic analysis methods may be applied to
documents 2 for keyword 8 derivation: lexical, word fre
quency, word placement, syntactic, semantic, or morpho

to rank keywords 9P is to use a point system to weigh relative
30

quency one-third. Keyword 8 relevancy rating schemes 9P
are known to those skilled in the art.
35

FIG. 5 depicts the directory supplementation 6 process.
Directory supplementation 6 must be enabled 10. Directory
supplementation 6 may be enabled 10 by default, by soft
ware-determined protocol, or by user determination. Auto

link 1 of a found document 2 without a user having to
manually add a link 1 to a directory 3.

matically supplementing a directory 6 refers to adding links
1 or documents 2 to a directory 3 without a user having to
40

search 12 or manually add links 1 to that directory 3.
Optionally, a breadth threshold level may be set 101. A
breadth threshold level is intended as user-determined set

ting that possibly adjusts the number and potential relevance
45

discovered link 1, or an obsolete link 13.
Attempting to retrieve a document 2 from a link 1

sometimes reveals that the link 1 is no longer valid: the
document 2 is gone, having been moved or removed. In this
instance, the link 1 should be signi?ed 66 as obsolete 13 if
its document 2 has certainly been removed, or, if a link 1 to
a moved document 2 can be ascertained, the stored link 1
should be updated to re?ect the new document’s 2 location.
Pages 2 or sites that have moved often temporarily leave a
notice behind telling where the site or page 2 has moved to.

prominence and frequency, where, for example, prominence
may comprise two-thirds of a keyword’s 8 score and fre

logical. Such methods are known to those skilled in the art.
Automatically displaying a link 1 refers to displaying a

Signifying a link 66 refers to visibly indicating the current
status of a link 1. Examples of visible indication include
color coding or other visible distinction of link 1 text, such
a font style; or striking icon 4: either the usual icon 4 color
coded, or icons 4 indicating status. Examples of status
include a newly found link 1, a level of relevance for a newly

9P. Factors esteeming a keyword 8 include the following:
prominence and frequency primarily in titles 20 and sec

of accepted documents 2. Greater breadth casts a wider net:
more links 1 or documents 2 are retained, and vice versa. If
a user desires closely related documents 2 as a product of

directory supplementation 6, set a low breadth level 101.
A relation threshold level would the mirror image equiva
50

lent to a breadth threshold level 101: a higher setting would

be indication to limit directory supplementation 6 to closely
related documents 2, and vice versa. Level indication 101
may be ordinal or numeric, such as percentage.
In an embodiment where breadth level setting 101 is
55

employed, the setting 101 may be applied before and/ or after

In such an instance, software linguistic analysis of the

search 12. A search 12 may use a broader setting 101 than

noti?cation can glean the new link 1.

the user speci?ed. If then directory supplementation 6 pre

Document 2 inaccessibility does not necessarily mean
link obsolescence 13: other possible causes exist, such as,
for example, temporary server problems at the document’s
2 home site. A link 1 should be signi?ed 66 obsolete 13 only

sents sparse results, a user may want to adjust to a broader

setting 101: if broader documents 2 have already been
60

retrieved, the outcome of a broadened search may appear to

the user immediately (with presentation of additional links

if document 2 removal can be veri?ed: inaccessibility over

1).

a prolonged period of time would be indicative. For

Documents 2 in a directory 3 are analyZed 9 for keywords
8. Derived keywords 8 and attendant data may be stored to
avoid repetition of the process 9. Attendant keyword data 8

example, by keeping track of attempted access times, link
obsolescence 13 may be concluded given document 2 inac
cessibility at different times of the day for over a period of
a week or so. Sometimes, document 2 removal is noted on

65

may include keyword 8 rating data, such as keyword fre
quency and prominence in a document 2.

US 7,047,482 B1
6

5

a user had speci?ed via breadth level setting 101 only
displaying links 1 level 2 or better, the “King Crimson
Live!” link 1F Would not be displayed.
The invention claimed is:
1. A computer-implemented method for augmenting a

Though titles are necessarily terse, that very terseness
makes directory 5 and link titles 7 an esteemed source of

keywords 8. If directories 3 are hierarchical, topical infor
mation regarding a nested (loWer level) directory 3 may be

gleaned 110 by looking up the directory title 5 hierarchy.
Title-derived 110 keyWords 8 may be given the highest

directory Without contemporaneous user input comprising:

regard.
The ?nal step in keyWord collation 11 is ranking 112 the
gleaned sets of keyWords 8 from directory 3 documents 2 by
cumulating and collating keyWords 11. This is, in essence, a
Way of comparing documents via their derived keyWords 8.
If a document’s 2 keyWords 8 vary markedly from other
documents 2 in its directory 3, that document’s 2 keyWords

accessing at least a ?rst document via a ?rst directory
Without contemporaneous user selection of said ?rst

document, said ?rst document comprising at least in
10

deriving at least one keyWord indicative of at least one

topical content from said ?rst document;
searching as a background operation a plurality of docu

8 may be disregarded. The outcome is a set of directory

ments in storage in at least one computer Without
contemporaneous user input of a search location, such

keyWords 88 Which may retained, along With attendant data
or intermediate results, to avoid unnecessary repetition of

that said search comprises searching for documents

the directory keyWord collation process 11.

related by said at least one keyWord to said ?rst

A Boolean logic search 12 for relevant documents 2
throughout all or part of a computer’s or netWork storage

(52, 62) proceeds based upon directory keyWords 88. Can

document, thereby accessing a second document;
20

didate documents 2 may be found using cursory search 120

directory.
2. The method according to claim 1, Wherein at least part
of said storage is on a different computer than the computer
25

played). Culling discarded links 86, though optional, is
highly recommended, as not doing so degrades utility:

30

making a user discard the same links 3 repeatedly Would
annoy the user.

compared 121 to directory keyWords 88. Unlike keyWord
collation 11, Where keyWords 8 may be incorporated (albeit

7. The method according to claim 1, With the additional
step of comparing the relevance of said second document to

evaluation Which provides the basis for ranking candidate
documents 122 for directory supplementation 6. A variety of

a preset threshold.
40

step of displaying said results directory.

a single link 1. This may be done after analyZing the pages
2 to determine the page 2 most closely related 33 to the
desired information. As a result, the selected link 1 for
supplementation 6 may be the site’s home page 2, the
top-most page 2 for that topical aspect of the site, or the
particular page 2 With the most relevant information. A

10. The method according to claim 1, further comprising

recognizing a precondition for autonomously augmenting
said results directory, prior to accessing said ?rst document.
11. A computer-implemented method for augmenting a

directory comprising:
autonomously initiating operation based upon a stored

standout page 2 should not be hidden: in the instance of a
50

precondition;
accessing at least a ?rst document Without contempora

is to use both.

neous user selection, Wherein said ?rst document com

Finally, in the preferred embodiment, the target directory

prises at least in part topical textual content;

3 is supplemented 6 With links 1, concomitant to breadth

deriving at least one keyWord indicative of at least one

level setting 101 if employed. Optionally, visibly signify
55

ment, the target directory 3 is supplemented 6 With neWly

topical content Within said ?rst document;
as a background operation, searching in storage in at least
one computer for documents related by said at least one
keyWord to said ?rst document, Wherein at least some
of said searched documents are independent and not
organiZed in relation to one another;
determining relevance of a search-accessed second docu

found documents 2 in a manner similar to the preferred

embodiment.

FIG. 6 depicts an example directory 3K of links relating to
the musical group King Crimson. The top section of the
directory 3K shoWs existing links 1K. During the process of
checking knoWn linked documents 2 to derive 9 keyWords 8,
the “Krusty King Crimson” link is found obsolete 13, and
visibly signi?ed as such. The bottom section of the directory

3K illustrates directory supplementation 6K. In the depicted
example, three neWly discovered links 1F are displayed,
along With indication 66 of their respective relevance 33. If

8. The method according to claim 1, Wherein said results

directory is said ?rst directory.
9. The method according to claim 1, With the additional

Links 1 to pages 2 on the same site may be collated into

links 66 to indicate relevance 33. In an alternate embodi

3. The method according to claim 1, further comprising
deriving a plurality of keyWords.
4. The method according to claim 3, further comprising
ranking at least tWo of said plurality of keyWords.
5. The method according to claim 1, further comprising
accessing a plurality of documents in said ?rst directory.

directory.

on a prioritized basis), candidate document keyWord com
parison 121 to directory keyWords 88 is a critical ?tness

fairly relevant site With a spot-on page 2, the smart choice

storing said ?rst directory.

6. The method according to claim 1, With the additional
step of signifying the relevance of said second document to
documents in the ?rst directory When displaying said results

Candidate document 2 keyWords 8 are derived 9, then

methods for rating found documents 122 for relevance 33 to
target keyWords 88 are knoWn to those skilled in the art.

determining relevance of said second document to said at
least one keyWord; and
adding a reference to said second document in a results

techniques, as WinnoWing may occur after documents 2 are

found.
Once candidate documents 2 are found 120, links 1 to
pages 2 or sites previously eliminated from the target
directory 3 may be culled 86. The obvious implication is that
to perform this function, previously deleted links 1 from a
directory 3 must be remembered (though no longer dis

part topical textual content;

ment to said at least one keyWord; and
adding a reference to said second document in a results

directory.
65

12. The method according to claim 11, Wherein said
storage is on a plurality of computers connected to at least
one netWork.

US 7,047,482 B1
8

7
13. The method according to claim 11, further compris

searching as a background operation storage in at least
one computer for documents related to said plurality of

1ng:

stored keyWords;

deriving a plurality of keywords; and
determining relevance of said second document to said

determining relevance of a found second document to said

plurality of keyWords.

plurality of stored keyWords;

14. The method according to claim 11, further comprising

conditionally adding a reference to said second document
in a results directory.
17. The method according to claim 16, With the additional
step of comparing the relevance of said second document to

comparing the relevance of said second document to a preset
threshold.

15. The method according to claim 11, further comprising
conditionally adding said reference to said second document
depending upon Whether said reference to said second
document already exists in said results directory.
16. A computer-implemented method for augmenting a

a preset threshold.

18. The method according to claim 16, Wherein said
storage is on a plurality of computers connected to at least
one netWork.

directory comprising:
accessing a plurality of grouped documents Without con
temporaneous user selection initiating said access;
deriving a plurality of keyWords indicative of an aggre

19. The method according to claim 16, Wherein adding a
duplicate reference in said results directory is avoided.
20. The method according to claim 16, Wherein adding a

gate content of said grouped documents;
prioritizing a relative relevance of said keyWords;

reference that Was previously deleted from said results

storing said plurality of keyWords With regard to said

relevance;

directory is avoided.
20