 |
Description  |
|
|
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to information retrieval. More specifically,
the present invention relates to a client/server model for information
retrieval based upon a user-defined profile, for example, for the
generation of an "electronic" newspaper which contains information of
interest to a particular user.
2. Background Information
The development of computerized information resources, such as the
Internet, and various on-line services, such as Compuserve, America
On-line, Prodigy, and other services, has led to a proliferation of
electronically-available information. In fact, this electronic information
is increasingly displacing more conventional means of information
transmission, such as newspapers, magazines, and even, television. The
primary issue in all of these resources is filtering the vast amount of
information which is available in order that a user obtain that
information which is of interest to him.
Currently, a user who subscribes to one of the above-services, or uses the
Internet, must manually scan through the various information resources in
order to obtain articles, postings, or other files which are of interest.
Typically, the user manually scans in areas or sources of interest
(publications, USENet Newsgroups, fora, or other areas classified by
topic) to find topic areas which may be of interest to the user. The user
then retrieves articles or files which have subject headings, for example,
matching those which the user wishes to read. For example, certain USENet
newsgroups on the Internet have an established etiquette (known as
"netiquette"), wherein postings must have relevant subject headings to
permit this type of manual scanning. Article headings in
newspapers/magazines serve similar functions. Thus, if the user does not
wish to view those stories with products for sale, then he simply ignores
those postings having the term "For Sale" in the subject heading. Finally,
as a last level of filtering, the user can read the stories which have
been filtered by topic, and subject heading, and if anytime during the
viewing of the story the user wishes to ignore the rest of the article, he
can stop reading it and simply discard the story. If desired, the user can
download the remaining stories in any number of prior art ways, for
off-line viewing.
As is clearly evident, this manual scanning process which a user must
engage in is very time-consuming. To some extent, news is already filtered
for readers. Those that subscribe to certain magazines/newspapers, or view
certain television networks, already have the vast base of
electronically-available information (e.g. wire services) filtered for
them. However, this "filtering" is performed on a large-scale basis, for a
wide audience. The tailoring of specific stories of interest to particular
users has not been performed. Thus, a need has arisen to automatically
sort through the large variety of electronic sources in order to generate
a subset of the stories available in electronic form which is tailored to
a user's specific interests is desired.
Thus, the prior art of obtaining user-relevant information from electronic
sources suffers from several shortcomings.
SUMMARY
An automatic method and system for retrieving information based on a
user-defined profile (e.g. a personalized newspaper). A user-controlled
client establishes communication with a stateless server, the server
presenting a list of options to the client (e.g. via Hypertext Transfer
Protocol (HTTP) exchanges) between the server and the client. The client
provides an identification of the user-defined profile. The server engages
a first application program (e.g. via a Common Gateway Interface (CGI)),
the first application program retrieving the user-defined profile wherein
the user-defined profile identifies information which is of interest to
the user. The first application program examines a database of information
and automatically retrieves a subset of the information from the database
based upon which information is of interest to the user as specified in
the user-defined profile. The server presents the subset of the
information from the database as generated by the first application
program to the client.
In one implementation the client comprises an HTTP browser application
program and the server comprises an HTTP server application program
operative on a remote computer system. The first application program
stores a file containing the user-defined profile in order to retain a
state of the user-profile, and cause the stateless HTTP server to emulate
a server which retains its state from session to session.
In one implementation, the first application program examines the database
of information and automatically retrieves a subset of the information
from the database based upon the user-defined profile at periodic
intervals, in order, for example, to keep the newspaper-up-to-date.
The user-defined profile can include source identifications and associated
search terms wherein the first application scans in the information (e.g.
a raw news source, USENet newsgroup or other resource) for sources
identified by the source identifications. A first set of files in the
sources containing the associated search terms may then be identified and
the first application program places the first set of files into the
subset of the information, for creation of the personalized
information--the newspaper. The source identifications and associated
search terms from the user-defined profile can be stored by topic wherein
the subset is displayed to the user arranged by topic.
The first application program can also cause the server to present options
to the user to create or modify the profile, including, a first option to
allow the user to specify source identifications in the information and
associated search terms to search for in the source identifications to the
user-defined profile; and a second option to specify delete and/or change
the source identifications and/or the associated search terms in the
profile.
BRIEF DESCRIPTION OF THE DRAWINGS
The present invention is illustrated by way of example and not limitation
in the figures of the accompanying in which like references indicate like
elements and in which:
FIG. 1 shows a client/server system in which embodiments of the present
invention may be implemented.
FIG. 2 shows a more detailed view of the client/server illustrated in FIG.
1.
FIG. 3 shows the structure of a workstation in which the client or server
may be operative.
FIG. 4 shows a more detailed view of the processes operative within the
server.
FIGS. 5a and 5b shows a sequence of steps performed in the server for
generation of a profile and/or creation of a personal newspaper.
FIG. 6 shows the display screen for user control of the process illustrated
in FIGS. 5a and 5b.
FIG. 7 shows a process of adding a topic to a profile.
FIG. 8 shows a display used for controlling the operation of the process in
FIG. 7.
FIG. 9 shows a process of editing a user profile.
FIG. 10 shows a display used for controlling the editing process.
FIG. 11 shows the results of execution of the personal newspaper generation
process according to the user-defined profile.
FIG. 12 shows the viewing of a single article contained in the personal
newspaper.
DETAILED DESCRIPTION
The present invention is a method and apparatus for automatically scanning
information using a user-defined profile, and providing relevant stories
from that information to a user based upon that profile. Although the
following will be described with reference to certain particular
embodiments, including data structures, flow of steps, hardware
configurations, etc. . . , it will be apparent to one skilled in the art
that implementations of the present invention can be practiced without
these specific details.
Implementations of the present invention use a client/server architecture,
as illustrated in FIG. 1, wherein user requests 110 for news are sent by a
client application program 100 to a server 150 (typically, a remote
computer system accessible over the Internet or other communication
medium). The server 150, as will be described in more detail below,
performs scanning and searching of raw (e.g. unprocessed) information
sources (e.g. newswires or newsgroups), based upon these user requests,
presents the filtered electronic information as server responses 160 to
the client process. The client process may be active in a first computer
system, and the server process may be active in a second computer system,
and communicate with one another over a communication medium, thus
providing distributed functionality and allowing multiple clients to take
advantage of the information gathering capabilities of the server.
A more detailed view of the client and server are shown in FIG. 2. Although
the client and server are processes which are operative within two
computer systems, these processes being generated from a high level
programming language (e.g. C or C++), compiled and executed in a computer
system (e.g. a workstation), it can be appreciated by one skilled in the
art that they may be implemented in a variety of hardware devices, either
programmed or dedicated.
Client 100 and server 150 communicate using the functionally provided by
the World-Wide Web (WWW). Clients and servers of the WWW communicate over
a communication medium 250 using a standard known as the Hypertext
Transfer Protocol (HTTP). In some embodiments, the client and server may
be coupled via Serial Line Internet Protocol (SLIP) or TCP/IP connections
for high-capacity communication. Active within the client is a first
process, known as a "browser" 200, which establishes the connection with
server 150, and presents information to the user. Any number of
commercially or publicly-available browsers may be used, in various
implementations, however in this implementation, browser 200 is the Mosaic
brand browser (version 2.0 or greater) available from the National Center
for Supercomputing Applications (NCSA) in Urbana-Champaign, Ill. Other
browsers such as the Netscape, Netcruiser, or the Lynx brand browsers, or
others, which are available and provide the functionality specified under
HTTP and the Mosaic version 2.0 brand browser or above may be used.
The server 150 executes the corresponding server software which presents
information to the client in the form of HTTP responses. The HTTP
responses correspond with the web "pages" represented using Hypertext
Markup Language (HTML), or other data which is generated by the server, as
will now be discussed.
A shortcoming of HTTP is that it is a stateless protocol. The Web "page"
from which user transactions may be performed in the server under control
of the client is not recalled from client-session to client-session. In
this implementation of the present invention the user is able to connect
to the remote server and specify a user profile, setting forth his
interests. The user is able to specify the context for the information to
be searched (e.g. the date). The user is able to save the profile on the
remote machine. Finally the user is able to retrieve the personal profile
(with any access control, if desired) and edit (add or delete entries) and
save it for future operations.
All of this functionality requires an underlying mechanism which is unmet
by current versions of HTTP. Under the Mosaic brand browser 2.0 and
greater, in addition to HTML functionality 210 provided by the server
(display and retrieval of certain textual and other data based upon
Hypertext views and selection of item(s)), a Common Gateway Interface
(CGI) 220 is provided which allows the client program to direct the server
to commence execution of a specified program contained within the server.
Using this interface, and HTTP, the server may notify the client of the
results of that execution upon completion. The server's application
program, the personal newspaper generator, maintains a record of the state
of each user's profile, and thus, provides state functionality from
session to session to an otherwise stateless protocol.
In order to control the parameters of the execution of this server-resident
process, the client may direct the filling out of certain "forms" from his
browser. This is also provided by the "fill-in forms" functionality 230
available under Mosaic version 2.0 and greater, which allows the user via
his client application program 100, to specify a "profile" in which the
server will cause an application program to function (e.g. the types of
stories/articles which are of interest to the user). The details of a user
profile will be discussed below.
A computer system, such as a workstation, personal computer or other
processing apparatus in which the client 100 or server 150 may be
operative is illustrated in FIG. 3. A workstation in which one
implementation of the present invention may be practiced includes system
300. 300 comprises a bus or other communication means 301 for
communicating information, and a processing means 302 coupled with bus 301
for processing information. System 300 further comprises a random access
memory (RAM) or other volatile storage device 304 (referred to as main
memory), coupled to bus 301 for storing information and instructions to be
executed by processor 302. Main memory 304 also may be used for storing
temporary variables or other intermediate information during execution of
instructions by processor 302. System 300 also comprises a read only
memory (ROM) and/or other static storage device 306 coupled to bus 301 for
storing static information and instructions for processor 302, and a data
storage device 307 such as a magnetic disk or optical disk and its
corresponding disk drive. Data storage device 307 is coupled to bus 301
for storing information and instructions. This may be used for storage of
the databases to be described here which maintain information about
currently defined problem descriptions using commercially available
software products.
System 300 may further be coupled to a display device 321, such as a
cathode ray tube (CRT) or liquid crystal display (LCD) coupled to bus 301
for displaying information to a computer user. Such a display 321 may
further be coupled to bus 301 via a frame buffer 310, which information
such as a single or multiple frames or images for display upon display
device 321. An alphanumeric input device 322, including alphanumeric and
other keys, may also be coupled to bus 301 for communicating information
and command selections to processor 302. An additional user input device
is cursor control 323, such as a mouse, a trackball, stylus, or cursor
direction keys, coupled to bus 301 for communicating direction information
and command selections to processor 302, and for controlling cursor
movement on display 321.
Note, also, that any or all of the components of system 300 and associated
hardware may be used in various embodiments, however, it can be
appreciated that any configuration of the system may be used for various
purposes according to the particular implementation.
In one embodiment, system 300 is one of the Sun Microsystems.RTM. brand
family of workstations such as the SPARCstation brand workstation
manufactured by Sun Microsystems.RTM. of Mountain View, Calif. Processor
302 may be one of the SPARC brand microprocessors manufactured by Sun
Microsystems.RTM., Inc. of Mountain View, Calif.
Note that the following discussion of various embodiments discussed herein
will refer specifically to a series of routines which are generated in a
high-level programming language (e.g., the C or C++ programming language)
and compiled, linked, and then run as object code in system 300 during
run-time, for example by the SPARCompiler available from SunSoft of
Mountain View, Calif. (SPARC and SPARCstation are trademarks of SPARC
International, Inc. and are licensed exclusively to Sun Microsystems).
These further are used in conjunction with the browser and server software
available from NCSA, as described above, including the specification of
the appearance of displays in HTML. It can be appreciated by one skilled
in the art, however, that the following methods and apparatus may be
implemented in special purpose hardware devices, such as discrete logic
devices, large scale integrated circuits (LSI's), application-specific
integrated circuits (ASIC's), or other specialized hardware. The
description here has equal application to apparatus having similar
function.
FIG. 4 illustrates the interaction between the server 150 and newspaper
generator application program 400, which is operative under control of the
CGI in the server. Depending upon user input, requests are sent to the
active application in the server, in this case the personal newspaper
generator 400, which causes the application to perform certain functions.
For example, in the case of a profile which does not already exist in the
server, newspaper generator 400 allows the creation and editing of search
specifications, known as a "profile", on the server. This allows the
server to generate the personal newspaper for the client when requested,
or at regular intervals, according to implementation.
The user profile or requests which are received from the CGI 220 in the
server 150 cause the newspaper generator to perform certain actions. For
example, in the case of a profile creation, the profile is created and
stored in local files, such as 410, which may be accessed at a later time
by the server during editing operations or during creation of the
newspaper. In this way, the stateless protocol of HTTP is transformed into
a state-protocol, wherein the client or user profile is recalled from
session to session. This also allows periodic automatic generation of the
personal newspaper, for example, at regular intervals, such as several
times a day or week, according to implementation.
As illustrated, the newspaper generator 400 uses a plurality of raw news
sources 420, 430, and 440, which may be any of a number of available raw
news feeds. For example, each of these may be a separate news source or
wire service (e.g. Reuters, Associated Press), while another may be an
electronic discussion group (e.g. a USENet newsgroup). Any number or types
of sources may be used, according to implementation.
FIGS. 5a and 5b illustrates the logic of the operation of the personal
newspaper generator. First, at step 502, the newspaper's front page, or
welcome screen is displayed. This is the page which is displayed when the
use first accesses the server via the specification of a URL (Uniform
Resource Locator). A display screen, such as 600 illustrated in FIG. 6, is
displayed upon the user's console. Options and other data entered on the
form displayed on the console are committed, causing invocation of the
newspaper generation program, upon selection by the user of the "doit"
icon 610 on display 600. At step 504, the user can then enter, via the
forms support in the browser, a profile or user name in field 602. This is
enabled by selecting using a selection device or other means, and filling
in the selected blank with the profile name (typically, an e-mail address
of the user). If desired in the given implementation, access control may
be performed at step 506, wherein the user is queried for a password.
If the user wishes to edit the profile as detected at step 508, which is
indicated by the selection of the user interface object 604, then process
500 proceeds to FIG. 5b. If not (user interface object 606 is selected),
then the user may specify a date for which the personal newspaper will be
generated, such as by filling in the field 608 in the user interface form,
by a pull-down, pop-up menu, or other means. Then, at step 512 it is
determined whether the profile for the user exists. If so, then the
newspaper is generated using the stored profile. According to the
topics/subject/keywords selected by the user in his profile, the raw news
source(s) are scanned, and that information matching the profile is
gathered at step 514. If the profile does not exist, then at step 516, the
user is given the options for creating and editing the profile.
If the user desired to edit and/or lookup the profile (icon 604 is
selected) then, it is determined at step 518 on FIG. 5b whether the
profile exists. If not, then the user is given the option to create a new
profile. Then, the user can edit the profile, if created or it already
exists, at step 522. Upon completion of any editing operations, the
process is complete and returns to FIG. 5b.
A personal news profile is stored in a ASCII text file by topic by the
newspaper generator. Each topic contains a number of sources (e.g.
publications or newsfeeds) to search and corresponding search term(s). As
in certain full-text retrieval systems, search terms may be delimited by
disjunctions (the ".vertline." character). An example of a news profile is
as follows:
______________________________________
Sun News.about..*.about.Sun Microsystems.vertline.SMCC.vertline.SunSoft.ab
out.insensitive
Microsoft.about..*.about.Microsoft.about.insensitive
Internet.about.usa.headlines.vertline.usa.national.vertline.opinions-edito
rials.vertline.
businesswire.vertline.prnewswire.vertline.business.finance.vertline.
international.france.vertline.international.germany.vertline.international
.japan
international.other.about.Internet.vertline.Information
SuperHighway.vertline.Mosaic.vertline.WWW.about.insensitive
Colorado.about.usa.headlines.vertline.usa.national.vertline.opinionseditor
ials.vertline.
businesswire.vertline.pr-newswire.vertline.business.finance.vertline.inter
national.france
international.germany.vertline.international.japan.vertline.international.
other.about.
Colorado.vertline.Denver.vertline.Boulder.about.insensitive
______________________________________
Each of the lists of strings for topics, sources and keywords are delimited
in the file by the ".about." character. The initial string (e.g. "Sun
News") identifies the topic, the second set of strings (e.g. ".*")
identifies the files specifications for the sources (wherein ".*"
indicates searching in all raw sources), the next set of strings are the
search terms, and the final string indicates whether the search terms are
case sensitive or not. The creation and editing of a profile will now be
discussed.
The adding of topics to a news profile is shown in FIGS. 7 and 8. FIG. 7
shows the process steps performed by the profile generator portion of the
newspaper generator, and FIG. 8 shows the user interface which controls
the adding of topics to the profile. First, at step 702, the user labels
the topic of interest, such as by filling in the field 802 provided by the
fill-in forms feature of HTML or CGI. At step 704, the user specifies
search terms used in the full-text search. These are illustrated in field
804. Any number of search terms may be used and the ".vertline." character
is treated as a disjunction ("or"). Then, by selecting either of user
interface objects 806 or 808, the user specifies whether the search terms
are case sensitive or not. This is detected at step 706. At step 708,
using either a scrollable list containing selectable item(s), as
illustrated in field 810, or other means, the user specifies the search
context(s) (the publications, newsfeeds, etc. . . ) in which to search. By
the selection of icon 812 or other commit means, the data entered into the
screen 800 at step 710 is added to the user profile for generation of the
newspaper.
The process steps and user interface for editing a user profile is shown in
FIGS. 9 and 10. For example, for a particular profile as displayed on
screen 1000, any of the selection icons 1002-1010 may be selected for
modification of that topic. At step 902 it is determined whether the user
has selected the "delete" option for deleting topics from the profile
(e.g. the selection of icon 1012 on screen 1000). If so, then the selected
item(s) are deleted at step 904. Then it is determined at step 906 whether
the user has selected the "add" icon 1014. If so, then the add option
(process 700 of FIG. 7) is invoked at step 908. It is then determined at
step 910 whether the generation of the newspaper based upon the profile is
requested by the user. This is indicated by the user selecting icon 1018
(including, specifying a date for which the newspaper should be generated
in field 1016). The specification of the date limits the scope of the
search to the date(s) specified. If generation of the newspaper is
requested, then the newspaper is generated based upon the profile at step
912.
As previously discussed, the process for actually creating the newspaper
functions as a full-text retrieval system which is controlled by the
context(s) (e.g. source(s) and date(s)) and corresponding search term(s)
specified in each of the topics contained in the profile. This functions
much in the same way as other on-line full-text retrieval systems except
that it may be automatically performed by the server on demand or at
periodic intervals, and that it is under control of the stored profile.
The mechanics of this will be briefly discussed.
In implemented embodiments, raw news sources come in as large text files
with a predetermined format. The full-text searching process performed
during newspaper generation is executed upon these large files. Upon
location of certain search terms having the specified context(s) in the
file, the portion of the file containing the located article is parsed so
as to be converted into an individual ASCII file which is resident on the
server. The file names of the articles are unique, for unique reference
using HTML. In implemented embodiments, the article file names are in
numeric form and are the year, month, date, hour, minute and second the
article was created all concatenated in order to be unique (no two
articles are received exactly the same second). Then, via a second parsing
process, the title of the article is used to create an anchor in the HTML
page representation of the personal newspaper, which references the
article by its file specification. Upon selection by the user, the text of
the article can then be referenced at the server, and displayed as a
second HTML page.
FIG. 11 shows the results of the creation of a personal newspaper--the
personal newspaper main screen. This is displayed at the client as a
result of the full-text searching, the parsing and HTML page generation
process performed at the server as discussed above. As illustrated in
screen 1100, the profile name is displayed as 1102. Each of the topics
from the profile is listed in a first level heading, such as 1104, 1106,
1108, etc. . . Each of the first level topic headings are followed by
second level headings, such as 1106a-1106c, 1108a-1108l, etc. . . ,
listing the names of stories, if any, which match the search terms and
sources. Via the selection by the user of the headings for each of the
stories (the above-mentioned anchors), the text for the story may be
referenced and viewed. An example of the display of an article from the
personal newspaper is illustrated in screen 1200 of FIG. 12. Via HTML, a
referenc | | |
