Note: Descriptions are shown in the official language in which they were submitted.
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
1
INFORMATION ACCESS
The present invention relates to information access and finds particular
application in sharing of information among a group of users.
The Internet is an example of a distributed file storage and retrieval
system, being a multimedia computer communications network built on world-wide
telephone and data networks. Over 100,000 servers of various types are
presently
connected to the Internet providing a publicly accessible distributed data
store.
Data may be stored on a server in a form accessible using an Internet
communication protocol called the "HyperText Transfer Protocol" (HTTP). A
server
storing and making data available in this form is known as an "HTTP server".
Data
files stored on HTTP servers and accessible by means of HTTP are known as "web
pages" which together form the "World-Wide Web", or simply the "Web". Web
pages are written using a special Web language called HyperText Mark-up
Language (HTML) that includes a facility to create links to other pages on the
Web,
as appropriate, and enables a user to navigate through information on the Web
by
means of such links. Information held on the Web is accessible to anyone
having a
computer connected to the Internet and with an interest in accessing it.
An HTTP Uniform Resource Locator (URL) has been adopted as a Web
standard to provide a consistent international naming convention to uniquely
identify the location of any Web resource, including for instance documents,
programs, sound and video clips. The HTTP enables URL-identified files (web
pages) to be located and transferred for reproduction at user equipment
connected
to the Internet. Underlying transport protocols, primarily TCP/IP, enable
connections to be established, between an Internet user and a Web server for
example, for the intercommunication of data.
Internet users may access information on the Web using proprietary Web
browser products running on personal computers (PCs) or workstations linked to
the Internet. Web browsers communicate with Web resources using standard
Internet protocols to download selected web pages, interpret embedded HTML
commands inserted at the time of mark-up by web page authors and, if
appropriate, display those pages graphically. In particular, an HTML command
may
be embedded within a web page to create a link, by means of a URL, to another
web page. Browsers provide a convenient mechanism for displaying the existence
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
2
of such links in a displayed web page and for enabling a user to select a
particular
displayed link and automatically retrieve the referenced page.
There are now estimated to be more than 60 million documents on Web
servers world-wide. Every day millions of people trawl the Internet for
information
using any one of a dozen or more different search tools.
Accompanying this, there has been, increasing amounts of information
stored in private networks whose operation is based on Internet Protocols
tlPl.
These private networks are commonly referred to as intranet networks.
One of the main advantages of an intranet for a corporation is the ease
with which information, that was previously localised to individual
departments,
can be accessed from across the organisation as a whole.
Information management tools such as the JASPER agent detailed in
British Telecommunications plc's co-pending application W096/23265 provide a
knowledge sharing facility whereby users may be alerted to the storage or
identification of documents or other information sets according to whether
those
users' interests, defined by user profiles, match the content of the stored
document. However, while users may be alerted to the existence of documents of
potential interest, there is no facility for interaction between users of the
information system.
According to a first aspect of the present invention there is provided an
apparatus, for accessing sets of information stored in an information system
and
accessible by means of a communications network, the apparatus having:
i) an input for receiving a set of information selected by a first user;
ii) data storage, or means to access data storage, for storing one or more
user profiles, each said user profile comprising at least one predetermined
keyword;
iii) generation means, triggerable to generate at least one set of meta-
information from the set of information received at the input, said meta-
information including at least a pointer for the set of information, and to
store said at least one set of meta-information in said data storage;
iv) comparison means for comparing at least one of said one or more user
profiles with said at least one set of meta-information and for identifying,
in dependence upon the results of said comparison, a user having a profile
similar to said at least o.ne set of meta-information; and
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
3
v) alerting means to alert said first user to the identity of said user
identified
by said comparison means.
According to this first aspect, a user submitting an information set may be
alerted to the identity of other users interested in that information set and
potentially sharing an interest with the user submitting that information set.
Preferably, the comparison means are operable to compare a user profile
associated with said first user with at least one further user profile and
thereby to
identify a user having a similar user profile to that of said first user. In
this way, a
user may carry out a search for other users having certain interests in common
on
the basis of their users profiles.
In a further embodiment, the apparatus may include selecting means to
enable said first user to select one or more of said identified users and to
generate
an alert message for sending to said one or more selected users. Users
identified to
the first user may include those having an interest in an information set
submitted
by the first user or those identified as a result of comparing user profiles.
An information system operating over the Internet for example, may be
arranged to operate only with HTML documents. According to a further
embodiment of the present invention, where stored sets of information conform
to
a first predetermined format, the apparatus includes conversion means to
enable a
set of information received at the input in a format other than said first
predetermined format to be converted into said first predetermined format and
stored in said data storage. In this way, information such as e-mails and
other
electronic documents, can be shared.
When storing an e-mail or an electronic document, on a Web site or a Web
page, the document or e-mail needs to have some measure of relevance to the
pre-
existing content, and possibly the structure, of the Web page or Web site on
which
it is stored. If it only has marginal relevance, or no relevance at all, to
the content
or structure of the Web page or Web site on which it is stored, then it may be
difficult for parties interested in the information to locate it.
Hence, when a person decides to share some information by means of an
intranet, it usually requires that they either design a new Web page or that
they re-
design an existing Web site. This ensures that the information is stored in a
logical
manner on the intranet.
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
4
However, such designing, or re-designing of a Web page represents a
significant amount of work. This usually results in the information in
question
remaining stored on a local server or an e-mail system.
However, most e-mail systems and local servers only permit access to the
5 information they store by a closed group of users. Also, many e-mail systems
and
local servers do not provide access to the information that they store by
means of
Internet protocols.
Embodiments of the present invention provide an interface between
systems that store information in a distributed manner, such as the Internet
and
corporate intranets, and information that does not fit logically within the
pre-
existing pages of such systems.
An example of the type of information that may not fit logically within a
pre-existing Web page or Web site is an e-mail or a memo. It is known to store
e-
mails on the Web via news groups. However, news groups have a logical
structure whereby e-mails are grouped into related topics.
However, where an e-mail relates to, say, a particular project and where
that project has its own Web site, it may not be sufficient to simply store
the e-
mail somewhere on the Web site. It may be necessary to create a separate Web
page for the e-mail so that it can be viewed within a particular context.
20 Embodiments of the present invention provide functionality for interfacing
e-mails and other electronic documents with the Web. This ability represents
an
improvement to the tool detailed in W096/23265, referred to above.
The use of profiling can provide a pro-active information sharing scheme
whereby users who have previously had an interest in a particular topic are
alerted
when a new item of information matching that topic is stored within the
information access system.
Thus e-mails and other electronic documents can be pro-actively
distributed to users whose interests are referenced by a tool or system that
embodies the present invention.
30 In particular the person storing the e-mail need not know every person to
whom the information needs to be distributed. Thus an e-mail relating to a
particular project can be stored and then shared with future members of the
team.
According to a second aspect of the present invention there is provided a
method of monitoring stored information sets accessible by means of a
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
communications network, for the purpose of alerting a first user to the
existence
of a second user having a shared interest in an information set selected by
said
first user, the method comprising the steps of:
a) storing a user profile for each user, which profile comprises at least
5 one keyword and an identifier for the user;
b) receiving a set of information selected by said first user;
c) generating a set of meta-information dependent on said received
information set;
d) comparing the generated set of meta-information with a stored user
profile other than that for said first user and, in dependence upon the
result from the comparison, identifying a second user having a user
profile similar to said meta-information;
e1 transmitting an alert message addressed to the first user comprising
at least the identity of said second user.
Embodiments of the present invention will now be described, by way of
example only and with reference to the accompanying drawings, of which:
Figure 1 is a schematic representation of an information access system
incorporating embodiments of the present invention;
20 Figure 2 is a flow chart of operation performed by the information access
system of Figure 1;
Figure 3 is a flow chart of step 205 of Figure 2 in greater detail;
Figure 4 is a flow chart of step 335 of Figure 3 in greater detail;
Figure 5 is a flow chart showing the steps by which a user may access
meta-information;
Figure 6 is a flow chart showing the steps in operation of the system
when a user selects a hype-text link;
Figure 7 is a flow chart showing the steps involved in searching for other
users having similar user profiles;
30 Figure 8 is a flow chart showing the steps in identifying other users
having
an interest in a particular information set.
Figure 1 depicts an information access system, referred to as a Knowledge
Sharing Environment tool (KSE tool) 100, that is connected to an information
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
6
system that stores information in a distributed manner, such as a corporate
intranet or the Internet 105.
The KSE tool 100 has a KSE data store 115 that stores sets of KSE
documents 120 (detailed further below), sets of meta-data 125 and sets of user
profiles 130.
The KSE documents are generated by the text to HTML conversion
application 155 and the meta data 125 is generated by the meta data generation
application 150. A message generation application 160 is used for
automatically
generating e-mail messages that are sent to users 140 of the KSE tool. A
10 comparison application program 145 is used for comparing items of meta data
125
against user profiles 130 and search queries. It may also be used for
comparing
user profiles 130 against other user profiles 130.
The KSE tool 100 may typically be hosted by a communications terminal
that is connected to a communications network and that is accessible by use of
software clients, such as Internet and Web browsers 135.
The KSE tool 100 also has access to a Web server 1 10 by means of a CGI
program. The KSE tool 100 uses the Web server 110 to communicate KSE
documents 120 and other information to the Web browsers 135. The Web server
1 10 preferably has the capability to transmit electronic mail to its users
140 using
the Simple Mail Transfer Protocol (SMTPI.
Accordingly, the KSE tool 100 can be said to exist in a distributed system
or a distributed environment, often called a heterogeneous distributed system
or
environment. Such systems are typically based on a client server architecture.
They may have numerous clients with the capability of accessing numerous
servers by means of one or more communications networks. The Internet is an
example of such a distributed environment.
The present invention however, is not limited to the Internet. It may be
implemented in other distributed systems based on other protocols and
languages.
For examples, Object (as in Object Oriented) based systems that use CORBA
(Common Object Request Broker Architecture).
Referring now to Figure 2, an overview of the operation of the KSE tool
100 is provided. In operation, a user 140 may have at step 200 generated or
located some information that they determine should be shared between the
users
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
7
of the KSE tool 100. At step 205, in order to share this information via the
KSE
tool 100, the user 140 stores it as a KSE document 120.
At step 210 the KSE tool generates meta data in respect of the newly
stored KSE document 120. - And at step 215, the KSE tool 100 compares this
meta-data 125 against each of the user profiles 130 that the KSE tool 100 has
stored in the KSE data store 115.
Meta-data is a term used to refer to a data set that describes information,
i.e. information about information. A simple set of meta-data would be a
summary
of a web page combined with the URL of the web page. Some other types of
information that may also form meta-data include, details on when a Web page
was last updated, who it was that updated it etc and any previous URLs that
the
page may have been stored under.
At step 220, those users whose profile 130 matches the meta data 125
of the KSE document 120 are notified that it has been stored within the KSE
tool
100. This notification may take place, for example, by e-mail or on the
opening
screen of the KSE tool at their next KSE session.
In this way, use of user profiles 130 allows the information to be easily
shared amongst users of the KSE tool without the need either to generate a new
Web page or to re-design an existing Web page. An additional advantage is that
the user profiles 130 enable the KSE document 120 to be shared amongst users
140 of the KSE tool 100 without the user 140 who is storing the KSE document
120 needing to know who might have an interest in it. This also enables the
information to be shared with future users of the KSE tool 100 who may not
have
an account when the information is stored. Similarly, it allows the
information to
be shared with users whose profile may change at some future point in time so
as
to match the meta-data 125 now being stored.
Typically the sort of information that will be stored as a KSE document
120 is e-mail messages and electronic documents such as memos, reports,
letters
of advice and other information that may not fit logically within the pre-
existing
contents or structure of a Web page or a Web site.
Referring now to Figure 3, further detail of generating KSE documents as
detailed at step 205 of Figure 2 is provided. As detailed above, at step 205,
a
user 140 submits an item of information to the KSE tool 100. At step 305 a new
text file is created within the KSE data store 115. This text file may be
generated
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
8
through the use of a text box on the Web browser 135. The browser 135 may
then submit the text file to the KSE tool 100.
At step 310, the KSE tool 100 stores the text file in the KSE store 1 15 by
assigning it a number, which is one greater than the number of the previous
text
file stored. Alternatively the text file may be stored within an area of the
KSE
store 1 15 that is dedicated to the user 140 storing the text file. In this
case, the
number assigned to the text file is one greater than the number assigned to
previous KSE document 120 stored in that user's area.
At step 315, the KSE tool 100 places a Hyper Text Mark-up Language
(HTMU framework around the newly stored text file. Typically, the HTML
framework is HTML tags indicating the title of the page and tags indicating
the
body of the text. This framework may also include under an HTML meta tag, a
comment detailing the time and date that the text file was stored in the KSE
tool
100 and an identifier for the user who submitted the information. The HTML
15 framework preferably includes information on the appearance of the KSE
document
120 on a Web browser 135. This HTML framework enables the text file to be
passed from the KSE tool 100 to the Web server 110 and then on to a Web
browser 135, using Internet protocols.
At step 320, once the KSE document 120 has been generated and
stored, the KSE tool 100 then generates meta data 125 about the KSE document
120 and then at step 325 stores this meta data 125 in the KSE data store 115.
This meta data 125 typically includes a URL pointing to a location on the
Web server 110 where the KSE document 120 can be accessed.
The meta data may also contain a summary of the information in the KSE
25 document 120. Such a summary may be generated automatically by a text
processing tool such as the ConTextT"" tool from Oracle. Alternatively the
summarisation tool called Prosum detailed .in "Davies, NJ, Weeks, R ( 1998)
ProSum Profile based text summarisation, First Automatic Text Summarisation
Conference ISUMMAC-1 ), May 1998, Virginia USA" may be used.
The meta data may also contain a list of key words and key phrases
appearing in the document. Such a list of key words and key phrases may be
extracted by text processing tools such as those listed above in relation to
tools
for generating summaries of text documents.
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
9
The meta data may also include details such as the person who requested
that the information be stored as a KSE document 120 and the date and time
that
it was stored in the KSE tool 100. The meta data may also contain an
annotation
about the original document. This annotation may be entered by a user when the
document is stored or by subsequent readers of the document. Such annotations
usually provide a clue as to why the document was selected for storage. For
example, where the document is relevant to a particular project, the
annotation
may simply be the name of the project concerned.
The meta data, having an embedded link to the relevant KSE document
120 is the means by which a user 140 of a Web browser gains information about
the KSE document 120 concerned and also the URL of the KSE document 120
concerned.
In this sense the meta data generated by the KSE tool 100 can be seen as
an index of the KSE documents 120 that are referenced by the KSE tool 100.
Once the text file has been stored and an HTML framework placed around
it, the text file may be referred to as a KSE document 120. The file name of
the
KSE document 120 may be used as a parameter that can be passed to a KSE-CGI
access program stored in the CGI bin of the Web server 1 10.
The KSE-CGI access program may be a standard form of CGI program
used for accessing data bases, such as the KSE data store 115, via the
Internet.
With such an arrangement, a standard URL will designate the KSE-CGI access
program. The file name of the KSE document will be passed as a parameter URL.
The KSE-CGl access program may operate in accordance with standard
CGI techniques for retrieving data from a data base into a Web server.
The KSE document 120 may be retrieved into the Web server 110, by
passing a KSE document 120 file name, to the KSE-CGI access program, as a
parameter within a URL. The URL containing the file name parameters will be
the
URL of the KSE-CGI program. The parameters may be passed within the URL in a
manner similar to passing query strings to search engines.
Once a KSE document 120 has been loaded by the KSE-CGI program into
the Web server, it may then be loaded to the tirowser that requested it.
The KSE-CGI access program provides a common interface facility
between the KSE tool 100 and the Web 105. When the KSE tool 100 is initially
accessed, it may present information to a browser 135 as a list of meta data
125.
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
Preferably the mesa data 125 is presented in chronological order from the most
recently generated item of meta data 125 down to the first item of meta data
that
was created. This enables the KSE tool to present the most recently stored
information to the user when they next log on to the KSE tool 100.
5 The KSE-CGI access program and the meta-data 125 operate together to
form a Web site where the content of the information presented and the
structure
with which it is presented can change dynamically.
For example, the meta data allows KSE documents with similar
information content to be referenced against each other. This, for example,
allows
10 e-mails and other electronic documents to be grouped together according to
their
project name. Hence a user need only do a search on the project name to
retrieve
the relevant information stored in the KSE tool 100.
This referencing of new KSE documents 120 against other KSE documents
120 allows new items of information to be added to the KSE tool 100 without
the
need for creating a new Web page or for amending an existing Web page.
This referencing also assists with reducing problems of information
overload as the new KSE document 120 can viewed in the context of other
relevant KSE documents 120.
This structure of the KSE tool 100, with a KSE-CGI access program
separating KSE documents 120 from HTML files on the Web server files, has a
further advantage of increasing the ease with which the KSE tool 100 can be
ported from one server to another.
The KSE tool also provides a mechanism whereby users 140 of the KSE
tool 100 may be automatically notified when the KSE tool 100 generates an item
of meta data that matches their profile. This is achieved by comparing newly
generated items of meta data 125 against the user profiles 130 held in the KSE
store 115.
Referring now to Figure 4, the following details how the user profiles 130
are compared against newly created KSE documents 120.
At step 400, an item of meta data 125 is generated by the KSE tool 100.
At step 405 the KSE tool 100 accesses the first user profile 130 stored in the
user's KSE data store 1 15. At step 410, the KSE tool 100 compares the item of
meta-data 125 against the user profile using a vector space algorithm. More
detail
on the vector space algorithm is provided below. If at step 415, a positive
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
11
comparison is detected, then at step 420 an alert message, which may be an e-
mail message, is generated by the KSE tool 100 and at step 425 the e-mail
address of the user concerned is retrieved by the KSE tool 100 from the users
profile information 130 that is stored in the KSE store 1 15.
At step 430 the alert message along with the user's 140 e-mail address is
sent to the Web server 110 which then combines the message and the e-mail
address into an e-mail. At step 440, the Web server 110 transmits the alert
message to each user concerned.
After this, at step 445, the KSE tool 100 tests whether all of the user
profiles stored in the KSE data store 115 have been compared against the newly
stored item of meta data 125. If they have, then the process ends. If they
have
not, then the KSE tool 100 loops around to commence from step 405 above,
where it accesses the next user profile 130 stored in the KSE store 1 15.
Alternatively, when a user amends their user profile, by adding one or
more key words or key phrases, the KSE tool 100 may compare the new profile
against all of the meta-data 125 stored in the KSE store 115. The user is then
alerted to any items of meta-data 125 that match the changed profile 130.
Similarly, when a new user 140 is provided with access to the KSE tool 100
their
profile 130 is compared against the meta-data 125 in a manner similar to that
of
Figure 4
A user profile 130 is typically a list of key words and key phrases that
define topics that a user 140 is interested in. Further detail on user
profiles may
be found in British Telecommunications plc's co-pending application W096/23265
referred to above.
The above description refers to the KSE tool 100 referencing documents
generated by its users. However, the KSE tool 100 need not be limited to this.
The KSE tool 100 may also reference Web pages and other Internet and intranet
sites.
This may be achieved by passing to the KSE tool 100 a URL (Internet
address) of a Web page that a user 140 deems worthy of being shared with other
users 140 of the KSE tool 100.
In response to receiving the URL, the KSE tool 100 causes the Web server
110 to access the Web page in question and to generate meta data in respect of
it. This meta data is generated in the same or a similar manner as discussed
above
CA 02339217 2001-O1-31
WO OOI08570 PCT/GB99/02495
12
in relation to the KSE tool and typically comprises the URL in question along
with a
summary of the Web page held by the URL and a set of key words for the Web
page in question.
As stated above, the meta data for each document referenced by the KSE
tool 100 contains a list of the key words and key phrases that have been
extracted
from the document in question. Each key word or key phase may have a number
associated with it that corresponds to the number of times that the key word
or
key phrase in question occurs in the document from which it was extracted.
These numbers are then copied into their corresponding entry in a T x D
term document matrix M. The matrix M represents a set T of all the terms t;
appearing in all the sets of meta data held by the KSE tool 100. The matrix M
also
represents the set D of al) the documents d, referenced by the KSE tool 100.
The
sets T and D are set against each other to form the matrix M. The value of
each
entry in the matrix M is obtained from the, list of key terms and key phrases
generated for the rneta data, i.e. the term document matrix identifies the
number
of times each term t, appears in a document d,.
The term document matrix M may be searched against both user profiles
and key word search queries, so as to locate documents referenced by the KSE
tool 100 that match either the user profile or the search query. This is
because
user profiles and search queries are similar in that they both consist of key
words
and key phrases.
To compare a user profile against the matrix M, a vector space matching
and scoring algorithm may be employed. An example of such an algorithm is
detailed in Salton, G "Automatic Text Processing", Addison - Wesley, Reading,
Mass., USA, 1989. It measures the similarity between a user profile p and a
document d according to:
sim( p, d ) t. * t. ~l t. * ~ t?
~i=l,n ~ ~p id i y,n ~P i=l,n ed
Where tp is the l~h term or phrase of user profile p;
where td is the I~h key term or key phrase of document d; and
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
13
where I? is the number of unique terms in the combined profiles of
document d and the profile p of the user in question.
The matrix M is of indefinite size as it cannot be determined how many
documents or terms it will reference. For this reason, the matrix M may be
created
using a dynamic array, linked lists or similar types of data structures.
This vector space matching and scoring algorithm may also be employed
in the identification of users of the KSE tool 100 who have similar profiles.
In this
case, a term user profile matrix U, rather than the term document matrix M, is
created. The term user profile matrix U is a matrix of the set of all the
terms T in
all of the user profiles held by the KSE tool 100. This set of terms T is set
against
the set of all of the user profiles P of the KSE tool 100. The entries in the
matrix
U may be either a 1 or a 0. Alternatively, the value of the entry may indicate
the
user's level of interest in the term.
Using the vector space scoring and matching algorithm against the term
user matrix U generates a measure of the similarity between the profiles of
two
users U ~ and UZ. In this instance, the algorithm is expressed as:
sim(u~ ~ Zl2 ~ = L.Ji-l,n ~t~~ * t~2 ~~ ~i=l,n t 1 * L...ii=1,n t 2
Where t;~ is the I~h term or phrase of user Ui profile;
where t;2 is the I~h term or phrase of user U2 profile; and
where I~ is the number of unique terms in the combined profiles of users
U ~ and U2.
Referring to Figure 7, a flow diagram is provided showing the steps in using
the term user matrix U to identify other users of the KSE 100 having a similar
profile of interests to a first user U, for example. Having generated the term
user
matrix U at step 700, at step 710 the above algorithm is run against the
matrix U
to determine, in turn, a measure of the similarity between the profile of user
U, and
that of each of the other users. At step 720, those users having the most
similar
profiles to that of user U, are identified and user U, is notified of the
identity if
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
14
each such user. At step 730, user U, is able to select one or more users from
those notified such that the KSE 100 is triggered to generate a message for
sending to each selected user, for example as an introduction from user U, and
to
inform the selected user of their shared interests.
Referring to Figure 8, a flow diagram is presented showing the steps in
operation
of the KSE 100 in enabling a user U, to identify any other users whose user
profile
matches meta data generated from a set of information submitted or selected by
the user U,. Following receipt by the KSE 100~of the information set at step
800,
at step 810 the KSE 100 generates a set of meta-information for the received
information set. At step 820 a matching algorithm, such as the vector space
scoring and matching algorithm described above, is applied to compare the
generated meta-information with the user profiles of each of the users of the
KSE
100. At step 830, on the basis of the matching algorithm results, those users
having user profiles most similar to the generated meta-information, and hence
with a likely interest in the information set submitted by user U,, are
identified and
their identities notified to user U,. At step 840, user U, is able to select
one or
more users from those notified such that the KSE 100 is triggered to generate
a
message for sending to each selected user, for example as an introduction from
user U, and to inform the selected user of their shared interests.
The KSE tool 100 however is not limited to simply referencing KSE
documents 120. The KSE tool 100 may also reference other Web pages. To do
this, a user 140 submits a URL to the KSE tool. On receipt of such a URL, the
KSE
tool 100 instructs the Web server to access the Web page concerned. Meta-data
for the page is then constructed in a manner similar to that for KSE documents
120. i.e. a title is stored, a summary is generated, key words are extracted,
date
and time of storage in the KSE tool is recorded as is the name of the user
storing
the information, etc.
The meta data for Web pages is preferably indistinguishable from the meta
data stored for KSE documents 120. This is to enable e-mails and other
electronic
documents stored as KSE documents 120 to be shared with other users of the
KSE tool indistinguishably from Web pages.
By enabling KSE documents 120 to be shared indistinguishably from other
Web pages, the KSE tool 100 provides a seamless Web based interface between
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
traditional Web documents and other electronic documents and e-mails that
traditionally have not been shared by use of the Web.
Referring now to Figure 5 the following details how the KSE tool 100
retrieves and displays meta data when a user 140 first accesses the KSE tool
100.
5 At step 500 the KSE-CGI access program receives a HTTP Get command
from a user's 140 browser 135 requesting that the title page of the KSE tool
100
be down loaded to their Web browser 135.
At step 505 the KSE-CGI access program accesses and retrieves from the
KSE data store 1 15 the item of meta data that was last generated. At step 510
10 the KSE-CGI access program extracts the URL that is stored as part of the
meta
data and at step 515 it also extracts the title of the information from the
meta
data 125.
At step 520 KSE-CGI access program combines the title information and
the URL so as to form a HTML Hyper Text Link.
15 At step 525, the HTML Hyper Text Link is uploaded by the Web server 1 10
to the browser 135 that transmitted the HTTP Get request at step 500 above.
Select items of meta data may also be uploaded to the Web browser 135 at the
same time as the hyper Text Link. These may be, for example, the date that the
meta data was generated. The user 140 causing the meta data 125 to be
generated and any comment that that user 140 stored with the meta data.
At step 530, the KSE-CGI access program tests to see whether or not the
first item of meta data 125 that was generated has been uploaded to the
browser
135. If it has not, then at step 535, the KSE-CGI access program accesses the
item of meta data 125 that was generated immediately before the item of meta
25 data uploaded to the Web browser 135 at step 525 above. The KSE-CGI access
program then loops around to repeat steps 510, 515, 520 and 525 detailed
above.
Alternatively, step 530 may test to see whether or not a set number of
items of meta data have been uploaded to the browser 135. Where there are a
large number of items of meta data, this option may be preferred as it enables
a
small number of items of meta data to be displayed with out crowding the
screen
of the Web browser 135. An additional control loop may enable the succeeding
items of meta data to be uploaded, at a users request, at a later stage.
When step 530 returns a positive result, that all of the items of meta data
have been uploaded to the Web .browser, then at step 540, the CGI-KSE program
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
16
uploads to the browser 135 as KSE navigation tool bar. This KSE navigation
tool
bar is preferably up loaded as one of two HTML frames. The other HTML frame
displays the meta data.
The KSE tool bar preferably displays HTML buttons that enable a user to
navigate within the KSE tool 100. The use of two frames, one with a KSE tool
bar
enables a user to remain in the KSE environment when viewing a Web page stored
outside of the KSE tool 100.
The process detailed in relation to Figure 5 is not limited to when a user
first access the KSE tool 100. A similar process for displaying search
results,
subsequent to a user searching the meta data 125, may also be used.
Referring now to Figure 6, the following details the operation of a browser
and the KSE-CGI access program when a user selects a hyper text link that has
been uploaded to their browser 135 at step 525 of Figure 5.
At step 600, when a user 140 selects a hyper text link, the KSE-CGI access
program receives from the users browser 135 a HTTP Get request for a HTML
frame set. The HTML frame set contains two frames. The first displays a KSE
tool bar and the second displays a user 140 selected Web page or KSE document
120.
At step 605, the frame set instructs the browser 135 to upload the
requested Web page or KSE document into the second HTML frame of the frame
set, which is achieved by the browser issuing further HTTP get request for the
required web page or KSE document.
If the selected Web page is a KSE document 120 then a second HTTP Get
request is sent to KSE-CGI access program. The KSE-CGI access program will
access the KSE document 120 into the Web server 1 10 which then uploads it to
the second frame on the browser 135.
When the requested Web page is returned to the Web browser 135 it is
uploaded into the second HTML frame on the browser display.
At step 610, the frame set requests the KSE-CGI access program to upload
a KSE tool bar, referred to as a KSE feedback tool bar to the first HTML
frame.
This KSE feedback tool bar enables the selected Web page or KSE
document 120 to be viewed from within the KSE tool 100. It also allows KSE
functionality to be accessed whilst viewing a selected Web page or KSE
document
120.
CA 02339217 2001-O1-31
WO 00/08570 PCT/GB99/02495
17
The KSE feedback tool bar preferably contains HTML buttons that enable a
user to indicate their interest in the selected Web page or KSE document 120.
This feedback on interest provided may be used to update the users profile.
The
feedback tool bar also provides a means for gaining access back into the KSE
tool
100.
