Note: Descriptions are shown in the official language in which they were submitted.
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Methods, Systems, and Software for
Automated Growth of Intelligent On-Line Communities
Cross-Reference to Related Applications
The present application is a continuation of U.S. Provisional Application
60/294,070
which was filed on May 29, 2001, and a continuation-in-part of International
Application No.
PCT/LTS00/25886 filed September 20, 1999 and published in English on March 29,
2001 as
Publication No. WO 01/22292. The international application is a continuation
of U.S.
Provisional Application 60/154,885 filed on September 20, 1999. All these
applications are
incorporated herein by reference.
Copyright Notice and Permission
One or more portions of this patent document may contain material subject to
copyright
protection. The copyright owner has no objection to the facsimile reproduction
by anyone of the
patent document, as it appears in the Patent and Trademark Office patent files
or records, but
otherwise reserves all copyrights whatsoever. The following notice applies to
this document:
Copyright ~ 1999-2002, Bodyl, Inc.
Background
Recent years have seen the explosive growth of the Internet as not only a
communication
system of global scope, but also a global digital library. Indeed, the
Internet has become a key
source of news and information on a full spectrum of topics for users the
world over.
Fueling this growth, millions of users have submitted, or posted, information
to services,
such as public newsgroups, message boards, and subscription mailing lists,
that provide a
relatively simple and inexpensive mechanism for users to share information
electronically with
many people. However, despite success in rapidly increasing the quantity of
information on the
Internet, the present inventors have recognized that these services suffer
from at least two
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
problems that ultimately limit the utility and value of their vast stores of
user-generated
information.
The first problem concerns the accessibility of the information stored by
these services.
Specifically, many services use their own unique form of user interface for
posting and finding
relevant information. This diversity, or lack of standardization, requires
users of multiple
services not only to learn different interfaces, but also to post and search
messages on each
service separately, even when the services focus on related topics. Although
some interfaces
allow the searching of several newsgroups for current and past articles, most,
if not all, of them
require users to repeatedly define and initiate searches to check for new
relevant postings.
Moreover, some services delete the oldest submissions or content on a weekly
basis, reducing
chances of users finding the content most relevant to their needs.
The second problem is the lack of an effective mechanism for collecting
related
information in a central database. In particular, many newsgroups and message
boards focus on
similar, highly related topics, yet they are not otherwise integrated or
associated with each other
or to other online sources of similar information, such as theme-oriented
websites. For example,
there are numerous independent health-oriented newsgroups, forums, message
boards, and
health-oriented websites in direct competition. This disunity forces users
seeking the best user-
generated information and potentially more authoritative, professionally
authored information on
a given topic to search multiple services and websites. Additionally, because
many participants
behave as spectators and do not actively contribute information, the direct
competition not only
leads to fragmented collection of information, but also divides the
communities of active users
across multiple services, frustrating the development of larger user
communities.
Accordingly, there is a need not only to improve the value and utility of the
vast stores of
information on the Internet through new modes of access and collection, but
also to encourage
the growth of on-line communities.
Summary
To address this and other needs, the present inventors devised exemplary
systems,
methods, and software for automating the posting and retrieval of content from
multiple data
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
services, such as newsgroups, message boards, forums, as well as encouraging
active
participation and growth of on-line communities.
For example, one system includes a posting module, a retrieval module, and a
web server.
The posting module, accessible through a webpage on the server, allows users
to create and
initiate data postings that are sent automatically to several newsgroups,
message boards, and/or
other on-line information sources. The retrieval module automatically
retrieves replies to the
postings at each of the on-line sources and presents them through the
webserver for user review
and further reply, eliminating the need for users to repeated visit posting
sites in search of reply
messages. In addition to the retrieved replies, the retrieval module
automatically finds and
retrieves content of any file type based on stored search or fitness criteria,
ultimately enabling its
integration into the content of one or more websites or databases for
convenient user access.
Other aspects of this exemplary system include a feedback module for adapting
fitness
criteria over time to regulate the content of the databases according to user
or administrative
desires. Specifically, the exemplary system adapts the fitness criteria based
on user interaction
with retrieved data and/or the content of user-generated postings, thus
promoting evolution of
website toward user interests. (specific user queries?)
Another system deploys the automated posting-and-retrieval capabilities within
a
webserver system that grants users data-access rights based on their
contributions or submissions
to one or more databases. This system scores the contributions based on
quantity, quality,
and/or relevance, granting access rights based on the scores and thus
promoting further
contributions. Additionally, it records the queries of users of the database
and facilitates
communications between users having similar queries as well as users making
similar
contributions, further promoting growth of a communities of users around
shared information.
Brief Description of Drawings
Figure 1 is a hybrid block-flow diagram of an exemplary data-retrieval-and-
posting
system 100 incorporating teachings of the present invention.
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Figure 2 is a facsimile of a graphical user interface (or webpage) 200
suitable for use with
system 100.
Figure 3 is a facsimile of a graphical user interface (or webpage) 300
suitable for use with
system 100.
Figure 4 is a facsimile of a graphical user interface (or webpage) 400
suitable for use with
system 100.
Figure 5 is a hybrid block-flow diagram of an exemplary system S00 based on
system
100.
Figure 6 is a flow chart illustrating an exemplary method of operating system
100.
Figure 7 is a conceptual flow diagram illustrating various online discussion
modes
applicable to systems 100 and 500.
Detailed Description of Exemplary Embodiments)
The following description, which references and incorporates one or more
figures,
describes and illustrates one or more exemplary embodiments of the invention.
These
embodiments, offered not to limit but only to exemplify and teach, are shown
and described in
sufficient detail to enable those skilled in the art to make and use the
invention. Thus, where
appropriate to avoid obscuring the invention, the description may omit certain
information
known to those of skill in the art.
Exemplary System Architecture and Method of Operation
Figure 1 shows an exemplary automatic data-posting-and-retrieval system 100.
The
connecting lines in the figure indicate that various components of the system
are interconnected
permanently or temporarily via a wired or wireless wide-area-network, such as
the Internet, or a
secure wired or wireless local-area networks, such as a corporate intranet.
They further indicate
through use of arrowheads, whether communications over the connections are
generally one-
way or two-way. System 100 includes servers 110 (at top of figure), user
access devices 150 (at
bottom), and a web server system 120 (in the middle).
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
More particularly, servers 110 include one or more types of servers, such as
one or more
Usenet (NNTP or network-news-transfer protocol) servers 112, one or more
bulletin or
message-board servers 114, one or more web servers 116, and one or more other
types of
database servers 118, such as mail servers, messaging servers, and file
servers. The
invention, however, is not believed to~be limited to any particular collection
of servers or
server types. Indeed, the invention is not limited to specific file type,
since for example,
various embodiments can retrieve graphics, video, audio, text, olfactory, and
biofeedback
files.
User access devices 150 include access devices 152, 154, 156, and 158 are
representative.
The term "access device," as used herein, encompasses browser-equipped
personal computers,
network appliances, personal digital assistants, telephones, cell phones, web
phones, televisions,
web television, etc. Thus, the present invention is not limited to any
particular class or form of
access device.
Selectively coupled between servers 110 and access devices 150 is webserver
system 120
which not only automatically retrieves and post data from and to servers 110,
but also provides a
common interface structure between these servers and access devices 150.
Specifically, webserver system 120 includes, in addition to conventional
processors 120.1
and memory 120.2 (shown in lower left and right corners), various software and
data
components which can take a variety of forms, such as coded instructions or
data on an
electrical, magnetic, optical, and/or magnetic Garner medium. These
components, arranged to
show an exemplary software architecture and methodology, include a crawler
module 122 (upper
left corner of box), inbox 124, decision block 126, acceptance webpage 128,
client interfaces
130, fitness database 132, feedback module 134, and forum database 136.
Additional
components include outbox 138, decision block 140, approval page 142, and post
module 144.
Crawler module 122 generally searches one or more external databases,
accessible via
servers 110, for data items or content based on search or fitness criteria
defined in fitness
database 132. The databases searched are defined in a control database 122.1.
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
The control database includes one or more site-, group-, or user-specific
master
subscription list of newsgroups and other servers, which are maintained by a
system
administrator, individual user, or automated software. Some embodiments can
also
automatically assess the fitness of new newsgroups and other network-
accessible information
sources based on a sampling of their content and then automatically add or
recommend addition
of newsgroups and sources that meet the given fitness or relevancy criteria to
the control
database or specific portion of the database. Newsgroup suggestions can also
be triggered by
looking at messages that others posted on multiple newgroups, since these
messages may suggest
some similarity of newsgroup focus. (In addition, control database 122.1
stores control and state
about a target information space. This is used to memorize state of the
information space from
the crawler module's point of view should the information space be dynamic or
temporal.)
In the exemplary embodiment, crawler module 122 is implemented as a scheduled
service (or daemon) that connects to a networked data servers via sockets,
automatically sending
appropriate access information (for example, usernames, passwords or other
security credentials.
Additionally, fitness database 132, which is initially supplied with at least
one keyword or more
generally content attribute 132.1 from an administrator and modified by
feedback module 134,
with each attribute associated with an indicator of whether its is desirable
or undesirable, and a
magnitude indicator of the relative strength or relative priority the crawler
module should apply
to the attribute in evaluating content for retrieval. The present invention is
not limited to any
particular type or class of content attributes. For example, some embodiments
use attributes
related to source code, object code, applets, particular binary data
sequences, digital watermarks,
file names, file sizes, file types, headers, tags, and metadata , markup
languages, and semantic
concepts.
More precisely, the crawler module uses a two-part screening process. The
first
evaluates the content prior to retrieval based solely on inclusion or
exclusion of specific content
attributes, such as keywords 132.1, defined in fitness database 132. (Some
embodiments
generate and retrieve summaries or facsimiles of found content rather than the
found content in
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
its entirety.) The second filters the initially retrieved content data based
on full-text keywords
search using an objective fitness function having the form
N
TotalFitness = ~ weight(k; ) freq(k; ), (1)
a=t
where N denotes the total number of keywords from the fitness database or
relevant subportion
of the fitness database that are in the text (or more generally content) being
evaluated; freq(kl )
denotes the frequency of the i-th keyword within the text being evaluated; and
weight(k; )denotes
the current weight assigned to the i-th keyword This can be normalized or
scaled any numerous
ways, for example, by standard decimal scaling into a particular range, such
as
zero to one.
Some embodiments use non-linear fitness functions. For example, one embodiment
uses
the following fitness function:
N freq(k; )
TotalFitness'= ~ weight(k; ) ~ j ,
i=t ~=t
where the weight for each keyword is the same as in equation (1), but the
effect of each keyword
on TotalFitness' increases non-linearly with each additional occurrence of the
keyword. To
illustrate, the following table compares the fitness contributions for various
frequencies of a
keyword that has an associated weight of 4, computed using equations (1) and
(2):
Frequency Linear Non-Linear
Total FitnessTotatFitness'
1 1 (4)=4 1 (4)=4
2 2(4)=8 (1+2)(4)=12
3 3(4) =12 (1+2+3)(4)=24
In some embodiments, a configuration file or graphical interface feature
allows
administrators, users, or groups to select between use of equations (1) and
(2) for fitness
measurements. (Selection and adjustment of the fitness function, for example,
via an
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
administrative interface, facilitates granular control over the search
results.) Still other
embodiments use other fitness functions and/or expand the range of selectable
fitness-
measurement options with other fitness functions based for example on
combinations or
variations of equations (1) and (2) andlor on weights that are functions of
frequency andlor other
document parameters. Some embodiments apply different fitness functions to
different sources,
source types, or file types.
Other embodiments associate one or more sets of keywords and/or weights in the
fitness
database with specific.users or groups of users. (Users may be identified by
their use of specific
access credentials, such as username and password.) Additionally, some
embodiments store user-
defined Boolean and natural-language queries and/or necessary credentials and
payment
authorizations to access commercial or other restricted;access databases.
If the fitness for a particular document is greater than a threshold (a number
provided
during configuration, (most often set to 0, but can be made more restrictive
for low memory
environments) then the entire information file (text, source, time and other
information
associated with the content) is copied to inbox 124.
Inbox 124 simply aggregates all data retrieved from servers 110. However, some
embodiments assign the data to one or more specific users or groups of users
based on the user-
or group-specific fitness criteria that led to its retrieval. Still other
embodiments provide user- or
group-specific inboxes.
Decision block 126 determines whether to automatically forward the contents of
inbox
124 to forum database 136 or to queue it for further automated or manual
screening. In the
exemplary embodiment, this determination entails checking the status of an
auto-acceptance flag.
If the flag is true, the data is transferred to forum database 136, and if the
flag is false, indicating
no automatic acceptance, the data is incorporated into webpage 128 for manual
screening.
Other embodiments use user-, group-, topic- or source-specific auto-acceptance
flags.
Still others combine user-,group-, topic- or source-specific auto-acceptance
flags with temporal-
window definitions to allow for automatic acceptance of certain defined
content for certain users
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
during certain windows of time. And still other embodiments define the auto-
acceptance flag as
a function of not only user, group, topic, source, and time, but also
additional relevance or legal
criteria. (See discussion of auto-approval flag relative to decision block 140
below.)
Webpage 128 shows contents of inbox 124 and enables a website administrator or
other
authorized entity, to individually accept or reject each retrieved item. This
information can be
viewed as a complete scrollable listing or a sublisting based on a fitness
threshold or range of
fitness scores. Webpage 128 also allows an administrator to alter fitness
values of some
information files, with any alterations ultimately translated to the keyword
scores in fitness
database 132.
Figure 2 shows an exemplary webpage or graphical user interface 200 used in an
alternative embodiment of webpage 128. Interface 200 includes an inbox display
region 210, a
a selected-content display region 220, administrative action region 230, and a
fitness update
region 230. Inbox display region 210 includes sort-command inputs 211-214,
item markers Z I5,
item date indicators 216, item identifiers 217, previous input,218, and next
input 219. Sort-
command inputs 211-214 invoke sorted display of inbox items by date, by source
(or origin), by
fitness score, and by thread. Item markers 215 are selectable to mark or
designate one or more of
the displayed items for further action. Item marker 215.1, representative of a
selected marker, is
filled in to denote its selected status. Item-date indicators 216 indicate
respective dates, such as
dates of retrieval or dates of original posting at a particular source,
associated with the displayed
items. Item identifiers 217 which are selectable independent of item markers
216, present indicia
of the content of the retrieved items current displayed in inbox display
region 210. Item identifier
217.1, representative of a selected marker, is shaded to denotes its
selection. Selected content
display region 220 displays content and other data associated with selected
data item in the
inbox. Administrative action region 230 includes command inputs 231-237, which
provide
options to accept all data items marked in the inbox or accept an entire
thread associated with a
selected data item.
Figure 3 shows another alternative webpage or graphical user interface 300
useful as an
alternative or addition to webpage 128. Notably, webpage 300 includes an input
310 for
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
inputting a fitness level that causes display of all items having a fitness
greater than the input
fitness level, an input 320 for causing acceptance of a retrieved discussion
thread into a specific
designated category. Input 330 shows the computed fitness of the displayed
item; however, this
fitness may be changed by an administrator. (Such a change will ultimately
result in
modification will be counted as a positive or negative event for relevent
content attributes in the
fitness database.) Region 340 shows markup language associated with or
embedded within the
displayed content.
As Figure 1 indicates, content that is accepted via an interface such as
webpage 12~, 200,
or 300, is stored (or referenced via hyperlink or other pointer) in forum
database 136 for
subsequent access via one or more clients (access devices) through client
interfaces 130, which
include a graphical-user interface, such as webpage 131 and other interfaces
132.
Webpage 131 presents the contents of forum database 136 in combination with
other
non-forum content to one or more specific users or groups of users.
Specifically, webpage 131,
which is associated with one or more theme-oriented or general-interest
websites, includes a
forum region 131.1, a non-forum region 131.2, a search region 131.3, and a
login region 131.4.
Forum region 131.1 provides access to all or a user- or group-specific portion
of forum database
136, and non-forum portion 131.2 provides access to other website content.
(Figure 4 shows an
exemplary webpage 400 , which is displayed in response to forum region 130.1.
and Figure 5
shows a webpage that is displayed in response to selection of an identifier
for a retrieved item on
page 400)
Forum region 131.1 includes a reply region 131.11 and a post region 131.12.
Selection
of reply region 130.11 allows a user to reply to one or more messages in forum
database without
having to navigate to the source of the messages. And, selection of post
region 131.12 allows a
user to post original messages to one or more newsgroups, bulletin boards, and
mailing lists.
In some embodiments, forum database 136 organizes the retrieved content into
categories based
on total fitness and/or particular keyword scores to assist users in locating
specific content.
Search region 131.3 invokes a search interface on webpage 131 to allow users
to search
against already retrieved content or to command specific natural-language,
Boolean, or
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
structured-query languages (SQL) searches against all the newsgroups or
particular newsgroups
or data sources using the crawler module or one or more separate search
engine. In some
embodiments, the search interface provides users the option to define target
data and target date
for receiving search results, such as overnight, next two days, or within a
week, or within a
S month. During the specified time, the crawler module (or one, or more other
search engines) will
search multiple sources across the Internet for the target data. For some
searches, the system will
report search results via email, instant message, page, etc. Still other
embodiments allow users
to navigate through the forum database, select one or more particular items,
and then invoke
display of other retrieved items that are most similar to the selected items
or command a search
based on the contents of the selected items.
In addition to webpage 131, client interfaces includes other interfaces 132.
This set of
interfaces includes any other conventional or emerging interface structure or
protocol recognition
and conversion facilities necessary to communicate with any available or
emerging form of
access device. Thus, for example, other interfaces 132 includes an email
interface 132.1 for
communicating via email with one or more accesses devices, a wireless-access-
protocol (WAP)
132.2 for communicating webpages to mobile devices, such as cell phones and
personal-digital
assistants; a pager interface 132.3 for communicating with pagers, and an
instant-messaging
interface for communicating using one or more instant-message protocols.
Outbox 138 serves as a first-in-first-out queue for outgoing reply and post
messages from
all users of the website(s) originating from client interfaces 130. However,
other embodiments
provide user- or group-specific outboxes. Outbox 138 outputs to decision block
140.
Decision block 140 determines whether to automatically forward the contents of
outbox
138 to post module 144 or to queue it for further automated or manual
screening. In the
exemplary embodiment, this determination entails checking the status of an
auto-approval flag, a
preset true-false flag. However, other embodiments employ a more complex auto-
approval
process which entails measuring the fitness of the user original or supply
post in a similar
manner as used to evaluate incoming data within crawler module 122 and
automatically
approving postings that satisfy a threshold fitness level or fitness range.
11
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Some of these embodiments relax the fitness criteria for internally generated
postings
(submissions) relative to the criteria for importing externally generated
postings. Still others
define the auto-approval flag as a function of user, group, topic, source,
time, legality, and/or
relevance criteria, including fitness data 132. Thus, the present invention is
not limited in terms
of the type of auto-approval decision process. In any event, if the flag is
true, the data is
transferred to post module 144, and if the flag is false, indicating no
automatic acceptance, the
data is incorporated into webpage 142 for further screening.
Webpage 142 provides access to contents of outbox 13~ and enables a website
administrator or other authorized entity to individually screen reply and post
items. In the
exemplary embodiment, webpage 142 follows a similar layout as webpage 200 or
300, as shown
in Figures 2 and 3 respectively. In some embodiments, webpage 14.2 includes
one or more of
the following additional control features, beyond what is shown in Figures 2
and 3: query-author
feature for sending a email to the author (assuming extraction of email
address from message or
public address book) or for querying forum content (or broader Internet for
other content from
this author; an evict-author feature for deleting all content from a
particular author from the
forum database and preventing retrieval of his or her items; a reject -similar
feature for rejecting
all items that are very similar to a marked or selected item based on a
conventional semantic
analysis, a check-legality feature for comparing the item to a listing of
pirated or likely-to-be
pirated content.
If an item is disapproved for posting, a system message is generated and
transmitted to
the author of the message. For disapproved postings that meet or appear likely
to meet fitness
criteria of other affiliated forums, websites, or databases (based on a list
and other criteria held
by webserver 120), some embodiments provide messages that refer users, via a
hyperlink, to
another website with its embodiment of system 100. Disapproved items, in some
embodiments,
are posted externally as desired, but shunted to a separate axea of forum
database reserved for
non-conforming content. Approved items are forwarded to post module 144.
Post module 144 distributes items, based on explicit or implicit approval, to
one or more
of servers 110 and forum database 136. In distributing the items, post module
144 performs any
12
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
necessary data and/or protocol conversions as necessary to ensure acceptance
at each of the
destinations. (Post module 144 also includes one or more hyperlinks,
designating webpage 131,
an associated recruitment page, affiliate websites, forums, etc. )
The exemplary embodiment implements post module 144 as a service (daemon) and
that
operates according to a schedule defined during configuration. For example,
post module 144
can be scheduled to post hourly, daily, etc. Other embodiments drive post
module on an event
basis, such as immediately on receipt of a new item for posting or after
accumulating a
predetermined number of new items. In some embodiments, posted items
originating from
webpage 131 are posted directly to forum database 136, bypassing outbox 138
and the
subsequent approval process.
Post module 144 also identifies keywords and/or concepts in the posted items
and
forwards this information to feedback module 134.
Feedback module 134 modifies or adapts fitness criteria to reflect greater or
lesser
desirability of content having specific attributes. In the exemplary
embodiment, adaptation
occurs in response to user or administrative actions on content containing
specific keywords.
Specifically, the exemplary embodiment treats user responses that reflect
approval or disapproval
of crawler-retrieved content as triggers to modify fitness criteria. Modifying
the fitness criteria
entails increasing or decreasing specific keyword weights or attribute
weights, adding or deleting
certain keywords, or more generally changing a parameter in a given fitness
function or changing
fitness functions. For example, a user reply to a crawler-retrieved message or
click on a link of a
crawler-retrieved webpage signal content approval, whereas deletion or
inattention to retrieved
items signal disapproval.
Other embodiments collect more direct user feedback through polling or
surveying users
about regarding how useful a particular retrieved document is and then
adjusting the fitness
criteria accordingly. For instance, one embodiment includes "more like this
one," "less like this
one," or "no more like this one" buttons in the frame of a content-display
page to receive direct
user input regarding desirability of particular content. Another embodiment
uses a multi-point
rating system. Common examples of such systems include four, five, and ten
points.
13
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Administrative actions, such as changing the fitness score of a retrieved item
or rejecting data in
the inbox can also signal an opportunity for adjustment of fitness criteria.
In response to a detected opportunity for adjustment, the exemplary embodiment
generates and sends an information file containing text of the relevant
document and any
associated user-created documents to the fitness module. In turn, the feedback
module not only
searches the associated information file for keywords that are present in
fitness database 132 ,
but also identif es new keywords or other semantic attributes for inclusion in
the database. The
weights for preexistent keywords are then increased or decreased by some
marginal amount or
percentage based on a desired sensitivity or control loop dynamic. Weights for
new keywords
axe initially set at the marginal weight and allowed to evolved to other
keyword weights in
response to user interactions with content containing the keywords or other
weighted attributes.
One embodiment adjusts the weights of keywords based on positive or negative
interactions with retrieved items. Specifically, after a predetermined number
of positive or
negative interactions with a specific retrieved (or original) item, the
feedback module adjusts
weights of all relevant keywords (or other attributes, such as source, author,
header tags, size) by
respective predetermined amount for positive and negative interactions. For
example, if the
threshold number of interactions for triggering an adjustment is 100
interactions, the positive
adjustment is +1.1, and the negative adjustment is -0.9, every 100 positive or
negative
interactions would cause a 10% increase or decrease in the relevant attribute
weights and thus a
corresponding impact on the net growth or expansion of the forum or website
content.
In some embodiments, administrators may also skew the fitness functions or
keyword (or
databases in specific ways for specific periods of time to encourage or
discourage specific types
of content. Thus, for example, for a political website, an administrator .(or
administrative
program incorporating semantic analysis) might assess whether the forum
database had a balance
of content on both sides of an issue, such as increasing taxes, and adjust the
fitness function or
keyword weighting to establish or maintain balance. Alternatively, if balance
was not desired in
all or a portion of the forum database, the fitness function or attribute
weighting could be
14
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
adjusted, again by an administrator or administrative program, to achieve the
desired mix of
content along one or more semantic dimensions.
Other embodiments also update the fitness data or criteria to include new
keywords,
concepts, or attributes of original and/or reply postings. The feedback module
has the ability to
S modify keyword weights based on the utility (fitness, number of replies,
acceptance rate of
similar posts) of a given post. In addition, the feedback module could also
suggest new
keywords by identifying similarities between posts deemed to be useful.
Exemplary Applications
Exemplary system 100 and its various components have countless applications.
For
example, system 100 can be used as an intelligence gathering mechanism by
corporations, non-
profits, individuals, and government agencies. For examples, corporations
interested in
assessing their public images or product appeal can set of a website using the
exemplary
software or a suitable variant to collect relevant postings from multiple
forums, message boards,
etc. Moreover, the corporations may post specific content, such as advertising
or product
testimonials, to incite a "public" response. Additionally, the exemplary
software can be used to
post classified advertising for goods and services to multiple forums, to
collect survey data, to
conduct a multi-bid-point auction.
System 100 can also treat emails from one or more associated email servers as
postings
or content for retrieval. From the perspective of posting, these embodiments
monitor content
from all or specific outgoing or incoming emails and modify the fitness
functions to reflect a
desirability of the content in these emails. Additionally, some embodiments
include a further
option to find new, previously undesignated recipients for the email (or
posting) from a
continually updated list of website members or newsgroups, mail lists, etc.
Notably, this email
feature can also be deployed within a group or organization to identify
relevant content from
emails, such as job skills found on resumes or detect transfer of sensitive
information.
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Further applications include: supplemental subscription-based search service
for web
portals, such as Google or AltaVista; aggregation and organization of data of
multiple file types
across all or a subset of nodes in a corporate intranet to facilitate use of
employee-generated data,
such as address books, expense reports, etc; finder for rare items, missing
persons, experts,
professionals; search across company-specific portals for relevant online
posts, such as resumes,
new customers, or technical competencies.
Exemplar~ystem 100 with Community-Building Software
Figure 5 shows an enhanced version of system 100 in the form of an exemplary
webserver system 500. In addition to the components of system 100, system 500
includes
membership database 510, theme database 520, and community-building software
530.
Membership database 510 includes information regarding members or potential
members
(users) of system 500. The figure shows this information as representative
records 512 and 514,
which are substantially identical in structure. In the exemplary embodiment,
record 512
includes member-identification data 512.1, member-profile data 512.2, member-
contact or -
linking data 512.3, and member-participation data 512.4. Member-identification
data 512.1
includes data for identifying or authenticating the identity of a user. Member-
profile data 512.2
includes data describing the professional biographies and credentials of the
member. Member-
contact data 512.3 includes data, such as one or more postal addresses,
telephone numbers, e-
mail addresses, or URLs for facilitating contact or communications with the
associated user.
And, member-system-participation data 512.4 includes quantitative and
qualitative information
regarding actual and permitted use of the system by each user. For instance,
the exemplary
embodiment maintains one or more access scores for each member, indicating
levels of access to
respective portions of theme-oriented database 520.
Theme-oriented database 520 includes theme data 522, site data 524, query data
526, and
user data 526. Theme data 522 includes one or more keywords, terms, concepts,
or website
16
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
addresses which define one or more aspects of the thematic or topical content
of database 520.
(In some embodiments, fitness data 132 and theme data 522 comprise the same
data.)
Exemplary themes or topics includes general healthcare and wellness
information for humans or
other animals, such as dogs, cats, or fish; specific healthcare information
for various parts of the
S human body, such as joints (knees, hips, elbows, spine, etc.) or organs
(heart, lungs, stomach,
kidney, liver, eyes, ears, skin, etc.); specific medical conditions, such as
allergies (food, plant,
etc.), cancer, arthritis, obesity, mental illness; auto-immune deficiency
(HIV). Other exemplary
topics include technology breakthroughs, health-technology breakthroughs,
children, cooking,
sports, entertainment, celebrities, politics, law, restaurants, consumer
products, motion pictures,
videos, music recordings, corporations, government officials, criminal
activity, schools, science,
wines, beers, foods, professional service providers (lawyers, doctors,
contractors, artisans, etc.)
colleges, alumni of educational institutions, genealogy, gossip, or sex. One
exemplary health-
oriented database includes user-generated health content, medical journal
content, and an archive
of health-oriented feature stories. Thus, the present invention is not limited
to any particular
theme or class of themes.
Site data 524 includes feature articles, journal articles and other content
added to database
524 manually by its creators, sponsors, or other parties governing or
maintaining the database or
automatically by the system itself. Query data 526 includes a listing of one
or more queries (or
query summaries) made by registered users or members of the community, against
the database,
with each query associated with one or more portions of the membership data
for its submitting
member. User data 526 includes user contributions to the database, with each
contribution
logically associated with or appended to one or more portions of the
membership data for its
submitting member. Site data 524 includes forum database 136.1 and user data
526 includes a
user-specific forum database 136.2.
In its exemplary operation, system 500 not only facilitates posting,
retrieval, and
searching of content across multiple on-line data services, but also promotes
the intelligent
growth and development of a theme-oriented database or website by awarding
users data-access
17
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
rights based on their data contributions and new-member referrals, and
promoting further
interaction of system users.
Exemplary System Operation
More specifically, Figure 6, which shows an exemplary flowchart 600,
illustrates an
exemplary method of operating system 500. Flow chart 600 includes blocks 602-
638, which are
executed serially in the exemplary embodiment. However, other embodiments of
the invention
may change the order of execution and/or execute two or more blocks in
parallel using multiple
processors or a single processor organized as two or more virtual machines or
subprocessors.
Moreover, still other embodiments implement the blocks as two or more specific
interconnected
hardware modules with related control and data signals communicated between
and through the
modules. Thus, the exemplary process flow is applicable to software, firmware,
and hardware
implementations.
The exemplary method begins at block 602, with automatically or manually
establishing
an initial version of theme-oriented database 520. To this end, the exemplary
method executes
blocks 602.1-602.4.
Block 602.1 entails determining a theme or topic by storing one or more
keywords,
semantic concepts, or other content attributes in fitness data 132 and/or
theme data 642. The data
in the fitness data and theme data are initially the same data or at least
related to the same topic.
However, some embodiments may use dissimilar data. The theme or fitness data,
in some
embodiments, identifies one or more key experts or authorities in a particular
field. This allows
for retrieving content authored by these experts or content that references
them and thus
potentially determining by association others experts and so on. Once the
theme or topic of the
database is determined, execution continues at block 602.2.
Block 602.2 entails identifying a set of one or more candidate members.
Exemplary
candidate members include existing websites or portions of websites related to
the theme and
persons or firms with expertise or indicated interest in the theme. The
publishers of identified
publications or websites are then invited to register as members of the
system. The exemplary
embodiments identifies them by conducting searches based on the content
attributes and/or
18
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
conducting searches based on relevant journals, research citations, articles,
posts, periodical
subscription lists, or expressed interest in electronic newsletters, requests
in online or offline
forums or bulletin boards. Identification can also include polling existing
experts, contributors,
and/or community members.
After identification of a set of candidates, the exemplary method evaluates
them in block
602.3. This entails sorting them into three classes: experts, contributors,
and community to
facilitate prioritization of recruiting efforts. Although some embodiments
employ subjective
manual sorting techniques, the exemplary embodiment applies an quantitative
algorithm that
assesses the degree of overlap of candidate recommendations using a "6 degrees
of separation"
model" which assumes that more widely known expert, contributors, or community
members
should have greater relative stature (access rights) with the website. An
alternative algorithm
determines a ranking based on the research citations, number, and quality of
research papers, and
number and quality of posts, number of press citations regarding the theme or
fitness data,
recency of relevant posts, informational requests, publications, or press
citations.
Some embodiments use crawler module 122 capability to parse metadata attached
to
retrieved content together with its fitness algorithms to conduct quantitative
andlor statistical
analysis of content. In these embodiments, crawler module 122 targets content
from one or more
candidate members using its metadata, and measures and stores the fitness of
the data, ultimately
enabling one to readily determine statistics, such as the average fitness of
content posted by a
given author within one or more specific on-line data services. More
generally, the method
could be used to assess the average fitness of all content for a particular
online data service or a
content author. The availability of such statistics facilitates comparison or
ranking of content
providers or authors in any number of desirable ways. Indeed, the highest
ranking set of
providers may become targets of high-priority recruitment efforts, and/or
included with the
content attributes driving crawler-module retrievals to ensure a steady input
of high-quality
content to the website.
Additionally, some embodiments use this methodology to cull poor-perfornZing
on-line
data services from the search space of the crawler module and to suggest
addition of new ones.
19
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
Still other embodiments may cluster other types of metadata, such as
copyright, publisher,
geographic origin of data, date of origin, around measured fitness levels,
enabling use of
conventional database querying to calculate a wide range of statistics, for
example, average
fitness of data posted from a specific geographic region. (Other embodiments
with appropriate
S semantic analyzers may be able to infer meta data from the content itself.)
After identification, execution then continues at block 602.4. This block
entails
recruiting or inviting one or more of the expert, contributor, or community
candidates to join the
website. In the exemplary embodiment, the initial recruiting efforts include
manual or automated
transmission of e-mail, letter, or telephonic solicitations to join.
Incentives, such as data- access
rights, but a financial stake in the website community, reciprocal web-traffic
opportunities,
publicity or other promotion, and priority access to events or new information
(information
access rights)-- or combinations thereof. Others include direct financial
compensation based on
the quality, quantity, and relevance of contribution provided by the expert,
contributor, or
community member.
In other embodiments, recruiting entails crawler module 122 posting messages
encouraging participants of newgroups, message boards, etc., from which it
retrieves content, to
visit or request information about one or more of its affiliated websites. In
some cases, the
messages includes hyperlinks or URLs to the site and some incentive for
visiting the website,
such as an electronic coupon or gift certificate or bonus access rights to
otherwise inaccessible
content on the site. Some embodiments even reply directly to specific
retrieved postings with
such messages. An exemplary message might state that based on the content of
your post, you
would be interested in the one or more of the following websites, on-line
forums, newsgroups,
etc.
In an event after initialization in block 602, execution of the exemplary
method proceeds
to block 604. Block 604, one or more of the candidate members establish a
communications link
with the system through webserver 120. This entails each of the candidate
members using an
access station, such as access device (or station) 152, to invoke the URL to
the system. For
example, the user at access station 152 would invoke "www.domain-name.com" to
connect her
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
computer system (or other network appliance) to webserver system 120. After
establishing the
link to webserver 120 execution proceeds to block 604.
Block 606 entails receiving registration information from the candidate
member. The
registration information includes member-identification data, member-profile
data, member-
s contact or -linking data, and member-system-participation data. Member-
identification data
includes data for identifying or authenticating the identity of a user, such
as a username and
password. Member-profile information includes professional biographical
information, such as
present employment, professional achievements, educational or other
promotional type material
indicating or suggesting the authority or credibility of the registering
member in the topic.
Member-contact data includes data, such as one or more postal addresses,
telephone numbers, e
mail addresses, or URLs for facilitating contact or communications with the
associated user.
Member-system-participation data includes an access score that governs the
level of access that
the associated member has over theme-oriented database 520. The exemplary
system
determines an initial access score based on whether the user was referred by
an existing member,
or whether the user was given a special invitation based on his or her
expertise in the theme. If
an existing member referred the user, the access score for the existing member
is increased upon
registration of the new user. After alI registration information has been
received for a particular
member, exemplary execution proceeds to block 608
In block 608, the system records the received registration information in
membership
database 510. Although the exemplary embodiment maintains membership database
510
separate from theme-oriented database 520 for heightened security, some
embodiments combine
the databases. With recording of the registration information, the exemplary
method advances to
block 610.
Block 610 entails the new member logging into the system to access theme-
oriented
database 520. Specifically, this entails the new (or an existing) member
manually or
automatically entering a username and password, for example at login region
131.4 (shown as
part of webpage 131 in Figure 1) in client interfaces 130. (Existing members
bypass blocks 602-
608 to reach block 610.) The username and password are then verified against
those in
21
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
membership database 510. Affirmative verification advances the exemplary
method to block
612.
In block 612, the system presents the member a home page, such as webpage 131
in
Figure 1, for theme-oriented database 520. From the home page, the member
decides to query
database 520 or to contribute data to database 520 as indicated by decision
block 652.
A member decision to make a contribution to the database branches execution to
block
616, which entails receiving a contribution from the member. In this exemplary
embodiment,
contributions include posting or responding to a post at one or more on-line
database services,
such as forum data 136.1 within site data 644, or one or more other affiliated
on-line services
using post module 144. It also includes referring or inviting one or more new
members to the
website. In some embodiments, contributions also include responses to one or
more survey or
polling features on the website, or even clicking through to affiliate or
sponsor websites.
Execution then continues to block 618.
In block 618, the system evaluates or scores the contribution based on its
quantity, quality
1 S and/or relevance to the theme-oriented content of database 520. To
evaluate data contributions,
the exemplary embodiment uses the evaluation logic of crawler module 122 and
the fitness data
in theme database 542 and/or fitness database 132. Another embodiment converts
the
contribution to a natural-language query and executes this query against all
or part of database
520. The natural-language searching algorithm produces quantitative measures
of the relevance
of the contribution. Other embodiments produce the relevance or quality
measures using
inverse-document-frequencies factors that favor rare terms and/or frequency
factors which favor
terms that in the document to be scored. In some embodiments, the contribution
is summarized
using specialized software, such as that described in U.S. Patent 5,708,825,
entitled Automatic
Summary Page Creation and Hyperlink Generation, which is incorporated herein
by reference.
Other embodiments score the contribution based on additional factors,
including for example,
length (number of words), number of citations to leading journals, inclusion
of hyperlinks to
predetermined cites (such as sponsors of the system) and/or grammar. Still
other embodiments
score the contributions manually using an editorial board of experts in the
theme. And still other
22
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
embodiments factor in the credibility or stature of the author based on
previous contributions to
the website and/or
Execution then continues at decision block 620, where the system determines if
the
evaluation was good or not. That is, is the contribution of sufficient
relevance and quality to be
added to database 520. If the contribution has a good evaluation, based for
example on its score
exceeding some threshold value, the system proceeds to block 622.
In block 622, the system adds the contribution to database 520. In the
exemplary
embodiment, this entails appending the username along with one or more
portions of the
member's member-profile information and/or member-contact information to the
beginning
and/or end of the contribution. (Some embodiments use a side by side
presentation approach for
the contribution and its attribution.) Thus, other members accessing this
information can identify
its contributing member and assess credibility and authority of the
contribution. Moreover, if a
contributing member has elected to allow publication of its contact
information, such as its e-
mail address, future users of the information may establish communications
with the contributor.
Publication of a hyperlink or LTRL associated with the contributor will offer
opportunities for
reciprocal web traffic from webserver 120 to a webserver associated with the
contributing
member, creating an incentive for further contributions to the theme-oriented
database.
In block 624, after adding the contribution to database 520, the system
updates the access
score portion of the participation data for the member based on the score of
the contribution. In
the exemplary embodiment, this entails adding the score of the contribution to
the initial or
previous access score in membership database 510. However, other embodiments
adopt other
forms of update.
Tn block 626, the system offers the contributing member an opportunity to
connect with
one or more other members who have made similar contributions to database 520.
In the
exemplary embodiment, this entails defining a query based on the contribution
and executing this
query against the user data portion of database 520, more precisely user data
529. ~ Based on the
results of this query, the exemplary embodiment presents one or more usernames
and related
contact or biographical information for other members who have made similar
contributions to
23
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
the database, thereby facilitating communications between the contributing
member and other
contributing members.
On the other hand, if the contribution had a poor evaluation based on the
threshold,
execution proceeds to block 628 which outputs a message to the member
indicating that the
contribution was rejected. Some embodiments may offer an explanation for the
rejection, and
others refer the member to an alternate system with an appropriate theme for
receiving the
information. Still others reroute the contribution to the alternate system,
automatically enrolling
the member in the alternate system by transferring a copy of the associated
membership
information. After rej ecting the data, execution returns to decision block
614, where the
member can choose to add new data or to submit a query for database 520.
From block 614, a decision to query database 520 branches execution to block
630,
which entails receiving a query from the member. Although the scope of the
invention
encompasses queries of any number of forms, the exemplary embodiment accepts
queries as a
textual string with Boolean connectors or as a natural-language query.
(Moreover, the
1 S exemplary embodiment permits the member to restrict the query to specific
portions of theme-
oriented database 520, such as to site data or user data.) Execution of the
exemplary method
then proceeds to block 632.
In block 632, the system executes the query against database 520. In the
exemplary
embodiment, this entails running the query against the entirety of database
520. However, some
embodiments restrict the query to one or more portions of database 520.
24
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
In block 634, the system presents query results to webserver 120 for viewing
by the member through an access station. The exemplary embodiment presents the
query results based not only on the query and the contents of database 520,
but also
on the access score for the member. Thus, for example, a low access score may
result in all or a portion of the results being withheld from the member. Some
embodiments advise the member quantitatively regarding the withheld portion of
the results. For example, the system presents a message that a certain
percentage of
the results is withheld. Other embodiments present citations for the portions
being
withheld to assist the member in determining the desirability of this withheld
information. Variations of this approach may present the profile of the
contributors
of the withheld results.
In presenting the results to the requesting member, the exemplary system
arranges or organizes the results based not only on relevance to the query
presented
but also on factors, such as the access rights of the respective contributors
of data to
the system. Thus, for example, data contributions from contributors that have
accrued relatively high access rights are generally presented before data
contributions from contributors with relatively lower access rights, assuming
the
contribution are of approximately equal relevance to the query. This
presentation
mechanism further encourages members, particularly those with related
websites, to
contribute content to the database.
Some embodiments present the results in a predetermined order based on the
portions of the database that contain them. For example, one embodiment
presents
found data in the order of feature articles, user contributions, and journal
articles,
with the items in each category arranged based on relevance and/or access
rights of
respective contributors.
In block 636, the system stores the query and associated member-profile
andlor member-contact information to the query portion of database 520. Once
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
stored in database 520, the query is searched like any other content within
the
database. When query results include one or more queries the queries are
presented
along with the usernames associated with the queries.
In block 63g, after storing the query, the system offers the contributing
member an
opportunity to connect with one or more other members who have made similar
queries of database 520. In the exemplary embodiment, this entails defining a
query
based on the query and executing this query against the query data portion of
database 520, more precisely user data 526. Based on the results of this
query, the
exemplary embodiment presents one or more usernames and related contact or
biographical information for other members who have made similar contributions
to
the database, thereby facilitating communications between the member and other
members with similar queries. Of course, the member then has the option to
contact
one or more of the other members. Other embodiments also presents the member
options to connect with members who have published information relating to the
query or to allow other members with similar questions to contact her in the
future.
Figure 7 shows a conceptual flow diagram 700 of an exemplary method of
fostering or developing online communities that are not limited to being
hosted at
one place at a particular time. In the diagram, the term "hosted community"
refers
to a community of users united in interest and by the central location of
their
discussion, such as at a data forum or website, but not united in time,
meaning users
or participants can join and leave a discussion at any time. The.term
"spontaneous
community" refers to a community of users united in interest and time, but not
by
any particular central location. With use of the data-posting-and-retrieval
system
(denoted crawler in the figure), virtual communities of people or
organizations with
similar interest can be built dynamically through an automated process that
identifies users of common interests, such as members of a theme-oriented
website,
automatically schedules an on-line meeting time, and sends invitations for a
26
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
spontaneous community discussions andlor hosted communities at one or more
theme-oriented websites. These hosted or spontaneous communities can bring
together users from public and/or private networks, such as users of a
corporate
intranet, members of a particular organization, or users from a particular
geographic
region. These electronic, discussion-based communities can unite users sharing
common interests, organization memberships, or geographical locations by
bringing
together contributors of valuable, relevant content.
An exemplary method procedure for creating a discussion community entails
using a crawler module to identify interests based on content and keyword
searches
across multiple discussion groups, compile contact information for the
interested
users (mobile phone, email, instant messaging, etc.), schedule a spontaneous
discussion and send invitations notifying users of the time and venue of the
spontaneous discussion and/or a hosted discussion. This provides users the
option
of participating in spontaneous discussion, or contributing to a hosted
discussion,
with contributors of valuable information.
In this manner, the posting and retrieval system can serve as an agent for
compiling interest profiles in addition to relevant content. These interest
profiles
can be used in combination with a scheduling module to dynamically organize
spontaneous electronic discussion-based communities, that can exist over fixed
or
wireless media via a variety of different platforms. These electronic,
discussion-
based communities can unite users sharing common interests, organization
memberships, or geographical locations by bringing together contributors of
valuable, relevant content.
Conclusion
Tii furtherance of the art, the present inventors have presented exemplary
systems, methods, and software automatically post and retrieve data from
multiple
27
CA 02455693 2003-11-21
WO 02/097678 PCT/US02/09626
online data services, such as newsgroups, message boards, etc, as well as
encouraging growth of online intelligence communities by rewarding data
contributions with access rights.
The embodiments described above are intended only to illustrate and teach
one or more ways of making and using the present invention, not to restrict
its
breadth or scope. The actual scope of the invention, which embraces all ways
of
practicing or implementing the teachings of the invention, is defined only by
one or
more of the following claims and their equivalents.
28
