Background of the Invention An expert system refers to artificial intelligence software capable of solving practical problems like a human expert, having the expertise of site experts stored in it and programmed to apply such expert knowledge to solve practical problems. The method for expressing human experts'knowledge in an expert system is called a"rule", which has the form of : If... (conditional sentence) Then... (action sentence) The method of using knowledge of the human experts as represented by the rules in an expert system for solving practical problems is called"inference", in which the action sentence is performed if the conditional sentence is true. However, in practice, since there exists not just one, but a plurality of rules, the act of checking the conditional sentence and performing the action sentence shall be done in repeated manner.

In addition, a conditional sentence or an action sentence of a rule comprises also variables, as e. g. in a set of rules for animal classification stated below:

IF"The young is fed with milk"= TRUE, THEN kind 1 = a mammal.

IF kind 1 = a mammal, AND age<5, THEN kind 2 = a baby mammal.

In the above example, "The young is fed with milk", kind 1, kind 2, and age are variables, while TRUE, a mammal, 5, and a baby mammal are constants. There are three types of variables: true/false variable, selective variable, and numerical variable."The <BR> <BR> young is fed with milk. "in the above example is a true/false variable, which can have as its value"TRUE"or"FALSE", while"kind 1"is a selective variable, which can have a selective value, such as"a mammal", "a bird", etc. , and"age"is a numerical variable, which can have a numeral as its value.

Software capable of comprehending the relationships between these rules and of conducting an inference by producing appropriate questions in correct sequence and in compliance with the types of the variables contained in the rules in order to derive a conclusion (an answer), is called an inference engine.

For instance, an inference in respect to the above two rules would yield a dialogue between the inference engine and a user as follows: Is"The young is fed with milk. "true ? Yes.

What is its"age" ? 3.

Conclusion: It is"a baby mammal".

A rule diagram is used for representing such rules in an easily understandable manner. A rule diagram is a diagram which represents knowledge of the human experts in the form of graphs, allowing not only to easily represent the expert knowledge, but also to easily maintain the knowledge by correcting the corresponding parts in the graphs when the knowledge has been changed. In addition, the knowledge as expressed in such graphs can automatically be converted into rules and can be used in inferences by an inference engine.

An example of rule diagram capable of generating the above rules is shown in Fig. l.

In case the expert knowledge has been changed, it can easily be maintained by changing the value of the node of the graph that corresponds to the rule (e. g. in case a <BR> mammal younger than three years of age is a baby mammal, "age < 5"is changed to"age < 3"), or by adding a node in case a new rule has been added, or by changing the arc between the nodes if necessary.

Further, software capable of generating a rule as exemplified above from a rule diagram is not only possible, but also has actually been developed (e. g. UNIK-LIFT). A rule thus generated is inputted in an inference engine (e. g. UNIK-BWD) and then used for inferences.

A web site is a set of numerous web pages significantly linked to each other, wherein hyperlink between the web pages occurs in a process similar to that of an inference utilizing the rules of expert system.

Detailed Description of the Invention The present invention, conceived to solve the above problems, aims to provide a

method for easy management of web site, by expressing relationships between web pages in graph form based on the inference theory in an artificial intelligence expert system and tools therefore.

The present invention aims to provide a web site management system and method, which is capable of automatically producing web sites based on the relationships between web pages that constitute a web site.

Another objective of the present invention is to provide a web site management system and method capable of expressing visually the relationships between web pages that constitute a web site in an easily understandable manner.

Still another objective of the present invention is to provide a web site management system capable of extracting information pertinent to the connections between web pages in a web site and of expressing the same visually in an easily understandable manner.

Another objective of the present invention is to provide a web site management system, which allows easily modify the contents of web pages constituting a web site, and the connections between these web pages.

In order to achieve the above objectives, the present invention provides a web site management system comprising an authoring interface, which allows a user to generate connections between web pages of a web site the user desires to construct in form of graphs consisting of a plurality of connection lines, that connect a plurality of nodes and two random nodes among them; a memory device for storing information relating to a plurality of nodes constituting the graphs and a plurality of connecting lines; and a code generation part for automatically generating source codes of the web pages using the information in graph form, wherein the above plural nodes comprise true/false nodes, selective nodes, numerical nodes corresponding, respectively, to true/false variables,

selective variables, numerical variables used in an expert inference system, frame nodes for expression of web page layouts, container nodes, and grouping nodes for simplification of the graphs. A true/false node is a node, that can be connected to another node based on the value of"yes"or"no"in respect to a predetermined question, a selective node is a node, that can be connected to another node based on a selected condition among a plurality of predetermined conditions given for a predetermined question, a numerical nodes is a node, that can have a numeral as its value, and can be connected to another node based on the conditional sentence corresponding to this value, a frame node is a node for expression of frames in a web page, a container node is a node for expression of menu-bundles in a web page, and a grouping node is a node for grouping of a plurality of nodes to simplify the graphs. Connections between the web pages are realized in the source codes generated by the code generating part, using script language, form input, or anchor; while the code generating part generates source codes, as two or more nodes are grouped by node grouping using the authoring interface, to ensure that the grouped nodes constitute one web page, and the code generating part generates source codes, as two or more nodes representing the same question are grouped by question variable grouping using the authoring interface, to ensure that the same question in respect to the nodes appears only once in the web page.

The above system can additionally comprise a template database storing template files inclusive of the source codes corresponding to the contents of the web page as exempted of the reserved words required for connecting the nodes among the source codes of the web page to be automatically generated. In such case, the code generating part generates automatically source codes of the web page using the graph information and the template file stored in the template database.

Further, the above system can additionally comprise a code analyzing part, which

receives the source code of the web page inputted, obtains the connections between the web pages by analyzing the script language, form, or anchor of the above source code, and converts the above connections obtained into nodes and connection line information, whereby the code analyzing part stores the source code corresponding to the contents of the web pages except for the source code for the connection information, which represents the connections between the web pages, in the template database.

The web site management method in accordance with the present invention comprises the steps of establishing connections between web pages of a web site desired for structure by a user, expressed in graph form composed of a plurality of nodes, including one or more of : a true/false node, which, corresponding to the true/false variables in an expert inference system, can be connected to another node based on the value of"yes"or "no"in respect to a predetermined question; a selective node, which, corresponding to selective variables used in an expert inference system, can be connected to another node based on a selected condition among a plurality of predetermined conditions given for a predetermined question; a numerical node, which, corresponding to numerical variables used in an expert inference system, can have a numeral as its value and can be connected to another node based on the conditional sentence corresponding to this value ; a frame node, which is a node for expression of frames in a web page; a container node, which is a node for expression of menu-bundles in a web page; a grouping node, which is a node for grouping of a plurality of nodes to simplify the graphs; and a plurality of connecting lines connecting any two of the above nodes; and automatically generating source codes of the web pages using the graph form information, wherein connections between the web pages are realized at the above step of automatically generating the source codes using script language, form input, or anchor; source codes are generated as two or more nodes are grouped together at the above step of establishing connections between the web pages so

that the grouped nodes constitute one web page; and source codes are generated as two or more nodes representing the same question are grouped together into node groups as classified by the question variables at the above step of establishing connections between the web pages so that the same question in respect to a node group as per the question variables appears only once in the web page.

Further, at the above step of automatically generating source codes, source codes of the web pages can automatically be generated using a template file inclusive of the source codes corresponding to the contents of the web page as exempted of the reserved words required for connecting the nodes among the source codes of the web page to be automatically generated.

In addition, at the above step of automatically generating source codes, source codes of the web pages can automatically be generated so that a node connected to a database is allowed to fetch the corresponding value in the database with the value inputted in the node. Here, the process can also be so constructed that the value fetched from the database is allowed for use as an input value of the corresponding node after a user authentication.

Brief Description of the Drawings Fig. 1 shows an example of a rule diagram in an artificial intelligence expert system.

Fig. 2 is a schematic diagram of a general system for generating and managing web site, using the web site management system in accordance with the present invention.

Fig. 3 is a block diagram showing the internal structure of an embodiment example of the web site management system in accordance with the present invention.

Fig. 4 is a graph showing a true/false type connection used in the web site

management system in accordance with the present invention.

Fig. 5 illustrates an example of web page generated from the true/false type connection graph in Fig. 4.

Figs. 6a through 6d show the source codes of each web page in Fig. 5 as generated by the web site management system in accordance with the present invention.

Fig. 7 is a graph showing a selection type connection used in the web site management system in accordance with the present invention.

Fig. 8 illustrates an example of the web page generated from the selection type connection graph in Fig. 7.

Fig. 9 shows source codes of the selection type connection web page generated by the web site management system in accordance with the present invention.

Fig. 10 is a graph showing a numerical type connection used in the web site management system in accordance with the present invention.

Fig. 11 illustrates an example of the web page generated from the numerical type connection graph in Fig. 10.

Fig. 12 shows source codes of the numerical type connection web page generated by the web site management system in accordance with the present invention.

Fig. 13 shows the rule diagram in Fig. 1 as expressed in graph form by the web site management system in accordance with the present invention.

Figs. 14 and 16 are examples of web pages generated from the graph in Fig. 13.

Figs. 15 and 17 illustrate the source codes of the web pages in Fig. 14 and Fig. 16, respectively.

Fig. 18 shows another example of web page generated from the graph in Fig. 13.

Fig. 19 shows an example of web page, which can be generated in case an identical question is repeated twice.

Fig. 20 shows a part of the source codes of the web page illustrated in Fig. 18.

Fig. 21 is an example source code obtained from a conventional web page in the web site management system in accordance with the present invention.

Fig. 22 shows an example of a command to obtain source code of a web page after having analyzed the source code in Fig. 21.

Fig. 23 shows an example of a template file used in the web site management system in accordance with the present invention.

Fig. 24a shows an example of an upper/lower division of a screen using a horizontal frame node, Fig. 24b shows an example of left/right division of a screen using a vertical frame node, and Fig. 24c shows an example of dividing a screen using a combination of horizontal and vertical frame nodes.

Figs. 25a through 25c show examples of web pages formed from Figs. 24a through 24c.

Figs. 26a and 26b show examples of expressing Fig. 25a in HTML, wherein a frame is used in Fig. 26a, while a table is used in Fig. 26b.

Fig. 27 shows an example graph using a container node.

Fig. 28 shows an example of web page generated from the graph in Fig. 27 using a container node.

Figs. 29a and 29b show examples of expressing the web page in Fig. 28 in HTML, wherein Fig. 29a is example of a template file defining the menu bundle layout of the container node, while Fig. 29b shows an example of. HTML file generated from the template file in Fig. 29a and the graph in fig. 27.

Fig. 30 shows an example of graph prior to its simplification through a grouping node.

Fig. 31 shows an example of graph after having been simplified through a

grouping node.

Fig. 32 shows an example of detail graph of a grouping node.

Fig. 33 shows an example of inference process by an ASP program after the database has been connected to the node, and the connection is confirmed.

Fig. 34 illustrates an example of the confirm screen generated by the program in Fig. 33.

Fig. 35 shows an example of inference process by an ASP program after the database has been connected to the node, and the connection is not confirmed.

Fig. 36 illustrates a structure example of storage for information on the nodes and the connection lines in the web site management system in accordance with the present invention.

Fig. 37 illustrates a structure example of storage for branching information relating to the branching of a web page from another web page in the web site management system in accordance with the present invention.

Description of the Preferred Embodiments The preferred embodiments of the present invention are explained below making reference to the appended drawings.

Fig. 2 is a schematic diagram of a general system for generating and managing web site, using the web site management system in accordance with the present invention.

The web site management system in accordance with the present invention (210) utilizes web site information directly inputted externally, or extracted from the web sites already established. The initial method of generating a web site is to create a plurality of web pages 220 based on the web site information inputted externally. The web pages 220 thus generated are transmitted to the web server 230, and then, provided via the Internet

240 to the users. The web site can be modified by changing web site information stored in the web site management system 210. On the other hand, management of an existing web site starts with accessing the corresponding web site via the Internet 240, and fetching the web site information by the web site management system 210. The information thus fetched is stored in the web site management system 210, with which information the web site is managed as in case of the newly generated web site.

Fig. 3 is a block diagram showing the internal structure of an embodiment example of the web site management system in accordance with the present invention.

The authoring interface 320 allows a user desiring to generate and manage a web site to design and edit a web site structure in form of graph such as a rule diagram enabling easy visual understanding of the web site structure. The authoring interface 320 is linked with a web site information database 340 storing node information and link information of the web site generated by the user, as well as with a template database 350 storing template files including other information such as constitution of the web page. The HTML code converter part 330, while linked with the authoring interface 320, converts the web site constitution generated in graph form by the user using the authoring interface 320 into a corresponding HTML source code. The HTML code converter part 330, being also linked with the web site information database and the template database 350, generates HTML source codes using information stored in these databases. The web site management system 210 further comprises an HTML code analyzing part 310, which receives information inputted from the existing web sites and shows in graph form to the users after having analyzed the same. The HTML code analyzing part 310, while linked with the web site information database 340 and the template database 350, stores the analyzed information in the web site information database 340 and the template database 350. The web site information thus stored is inputted in the authoring interface 320 to be provided to

the users for the users'view or modification, and can further be converted to a new HTML code by the HTML code converter part 330.

On the other hand, since the concept of variable in the above expert inference system is capable of expressing every possible connection, as described above, all possible connections between web pages can be expressed using this concept. The variables used for expression of such connections are classified into three types, i. e. a true/false variable, a selective variable, and a numerical variable. Connections between web pages can be expressed using these three variables, e. g. a move to another web page through click on a menu appearing frequently on a web site, can be understood as a connection via a numerical or a true/false variable. Further, a connection to another web page via a form input and push on the confirmation button can be understood as a connection via a numerical variable.

Now, the process of expressing in graph of connections via each variable, and the process of generating a corresponding web page are described below.

First, the true/false type connections are explained making reference to Fig. 4 through Fig. 6d, wherein Fig. 4 is a graph showing a true/false type connection used in the web site management system in accordance with the present invention, Fig. 5 illustrates an example of web page generated from the true/false type connection graph in Fig. 4, and Figs. 6a through 6d show the source codes of each web page in Fig. 5.

As shown in Fig. 4, a true/false connection is one to be established dependent on the value of"Yes" (or"True") or"No" (or"False") in respect to a given question (variable).

A connection is made to different conclusions (Conclusion 1, Conclusion 2) depending on the value to the question. In order to create a web page comprising such connections, a user can generate a graph as in Fig. 4 using the authoring interface of the web site management system as per the present invention.

And then, four web pages as shown in Fig. 5, which correspond to the four graphs in Fig. 4, respectively, are automatically generated using the web page generating function of the web site management system of the present invention. The source code file name of the web pages can be, e. g. WBIO. HTML, WBI1. HTML, WBI2. HTML, and WBI3. HTML.

Further, as illustrated in Fig. 5, it can be known that each node in the graphs in Fig. 4 corresponds to the respective web page in Fig. 5. The HTML source codes corresponding to each web page in Fig. 5 are as exemplified in Figs. 6a through 6d.

As shown in the source code of the file WBIO. HTML (Fig. 6a) corresponding to the title page 510 in Fig. 5, title of the text is expressed as"Decision FT","True/False Type", and a move to the next page WBI1. HTML can be provided by a click on the "Confirm"button.

The source code of the question page 520, which is the second page of the true/false connection, is shown in Fig. 6b.

Here, the conclusion pages WBI2. HTML or WBI3. HTML are connected, respectively, by the corresponding answer"Yes"or"No"to the question, using script language, whereby the input is made through radio buttons. However, other input means, such as hyper linking to a corresponding conclusion page via an <A> tag of HTML language, using a checkbox, etc. may naturally be used as well.

Upon selection of"Yes"on the question page, the third page 530 WBI2. HTML is connected, while the fourth page 540 WBI3. HTML is connected upon selection of"No"on the question page. Each of the above conclusion page 530,540 shows a conclusion (Conclusion 1, Conclusion 2) corresponding to the value of the question (variable) on the second page, whereby a click on the"Confirm"button effects a return connection to the first page 510.

The second type of connection is selection type connection, of which the examples

of graphs and the web pages corresponding thereto are illustrated in Figs. 7 and 8.

As shown in Fig. 7, a selection type connection takes a value, such as"Condition 1","Condition 2", etc. in respect to a certain question (variable), whereby a connection to a divergent conclusion (Conclusion 1, Conclusion 2, etc. ) is established dependent on the values a question takes. For manufacture of a web page comprising such connections, a user can produce a graph as in Fig. 7 using the authoring interface of the web site management system in accordance with the present invention, whereby a web page such as one illustrated in Fig. 8 is produced automatically. The title page 810 and conclusion pages 830,840, 850 are similar to their respective pages in the true/false connection described above in reference to Figs. 4 through 6d. Source codes of the second page 820, corresponding to the question page, are shown in Fig. 9.

As shown in Fig. 9, the page is described in script language, wherein a user is allowed to select one among the given conditions using radio buttons, so that a connection to corresponding conclusion page is established in accordance with the selected condition.

The third type of connection is numerical type connection, of which the examples of graphs and the web pages corresponding thereto are illustrated in Figs. 10 and 11.

As shown in Fig. 10, a numerical type connection takes a numeral as its value, in respect to a certain question (variable), whereby a connection to a divergent conclusion (Conclusion 1, Conclusion 2, etc. ) is established dependent on the values a question takes.

If necessary, numerical connections in specific forms such as time or date may separately be adopted.

For manufacture of a web page comprising such connections, a user can produce a graph as in Fig. 10 using the authoring interface of the web site management system in accordance with the present invention, whereby a web page such as one illustrated in Fig.

11 is produced automatically. The title page 1110 and conclusion pages 1130,1140 are

similar to their respective pages in the true/false connection or in the selective connection.

Source codes of the second page 1120, corresponding to the question page, are shown in Fig. 12.

As shown in Fig. 12, the question page 1120 provides a window for input of a value in respect to the variable"a", and then, connects to a conclusion page corresponding to (the range of) the inputted value.

The process of generating a web site comprising connections similar to those in the rule diagram as exemplified in Fig. 1, using the above-described three connection types, is explained below in detail making reference to Figs. 13 through 20.

The authoring interface of the web site management system in accordance with the present invention provides a user interface as in Fig. 13. A user can produce graphs illustrating relationships between web sites consisted of starting nodes, true/false nodes, selective nodes, numerical nodes, conclusions, etc. using the user interface, and generate web pages using these graphs. The graph illustrated in Fig. 13, made for generating a web site having contents similar to the rule diagram in Fig. 1, describes connections between six web pages (Classification of animals, The young is fed with milk, Ages, Conclusions). <BR> <P>Here, the first node is a starting node; the second node, "The young is fed with milk"is a<BR> true/false node; and the third and fourth nodes, "Ages"are numerical nodes. A connection to the corresponding conclusion node is established based on the value of the"Age"node. <BR> <P>A screen for connecting the web page, "The young is fed with milk"to the web<BR> page, "Age"can be constructed as Fig. 14, i. e. after either one of the answers"Yes"or "No", to the question has been selected via a radio button, the corresponding web page, "Age"can be connected. The web page source codes for realizing such connection is as illustrated in Fig. 15. As shown in Fig. 15, in case"Yes"has been selected, the

corresponding"Age"web page,"WBI2. html" is connected, while in case"No"has been selected, the corresponding"Age"web page,"WBI3. html" is connected.

An example of the"Age"web page, and its source codes are shown in Fig. 16 and Fig. 17, respectively, in case of selection of"Yes"in the web page in Fig. 14. The source codes show that the conclusion page,"WBI4. html" is hyper connected if the value of age inputted via Java script is below 5. In case"No"is selected in the web page in Fig. 14, the screen generated is the same, but a source code is provided for to allow a connection to another corresponding conclusion page only when the age value is below 1.

However, there can be cases where connection to another web page is needed based on different values inputted simultaneously, in other words, an automatic generation <BR> <BR> of a web page containing two questions, "The young is fed with milk ?" and"Age ? " simultaneously, can be needed. The screen thus generated shall be one as shown in Fig. 18.

To enable expression in graphs of the above features, the concept of"node grouping"is adopted, wherein two or more nodes constituting a graph is grouped to <BR> <BR> indicate that they are to be displayed on one screen. If the web page, "The young is fed with milk"and the"Age"web page to the left, which is connected subsequently via"Yes" <BR> <BR> to the question, "The young is fed with milk ?" in the graphs shown in Fig. 13 are grouped<BR> together, i. e. "node grouped", a web page as shown in Fig. 18 can be obtained. In other<BR> words, if a user selects in an authoring interface as in Fig. 13 the web page, "The young is fed with milk"and the"Age"web page to the left, and then, selects the node grouping function of the menu, these two web pages are node grouped. As shown in Fig. 13, an <BR> <BR> indication 1: 1 is placed prior to the title of the web page, "The young is fed with milk", while an indication of 1: 2 is placed prior to the title of the"Age"web page, from which it can be seen that these two web pages belong to the same group, being the first figure"1" therein the group identifier.

However, since a connection to the"Age"web page is also required for cases <BR> where the question, "The young is fed with milk ?" is answered with"No", the three web pages inclusive of the"Age"web page to the right shall be grouped together.

A web page, which is automatically generated by grouping the above three web pages, can be one as shown in Fig. 19, wherein the"age"question appears twice. In order to solve this problem, the present invention adopts the (concept of)"question variable grouping", which allows to group the same questions to one so that repetition of the same question can be avoided, in addition to simply combining web pages by"node grouping".

In other words, if a user selects in an authoring interface as in Fig. 13 the"Age"web page to the left and the"Age"web page to the right, and then, selects the question variable grouping function of the menu, the questions in these two web pages are grouped to one.

As shown in Fig. 13, an indication 1: 2 is placed prior to the title of both"Age"web pages, from which it can be seen that questions in these two web pages belong to the same question group, being the second figure"2"therein the question identifier. Here, the script part of the source codes to be automatically generated can be one as shown in Fig. 20.

Java Script, VB Script, and the like may be used as the script language for the source codes to be generated automatically, and the environment for the web server may be any HTML or DHTML code generating environment such as ASP, PHP, Oracle PL/SQL, <BR> etc. , i. e. the present invention is not limited to specific input methods, script languages, server environments.

On the other hand, the present invention is also capable of expressing in graphs the structure of an existing web site through a reversed process of the above described one, and managing the same through the graphs. For this purpose, a HTML code analyzing part of a web site management system is used.

In a HTML file as exemplified in Fig. 21, the URL is http ://misnt. gsnu. ac. kr/-yusong/WBIl. html. It can be known from an analysis of this file that a file,"WBI2. html" is hyper linked at the part,"Yes". Namely, a complete graph can be extracted from the interconnected web pages through analysis of <A> tag or <FORM> tag of HTML file or DHTML file, and of the script language, which process is a reversed process of the above described process.

The problem here is to obtain the HTML file or DHTML file for source code of the web pages constituting the web site, which can be achieved by an agent system utilizing hyper text transfer protocol (hereinafter, "HTTP"). The web site management system in accordance with the present invention comprises an agent system for obtaining source codes for web pages constituting a web site in its HTML code analyzing part. Here, the processes of obtaining the source code of a web page using URL of the web page corresponding to the route node, and then, of obtaining the source codes of the connected web pages following the hyper links expressed in the above source code are repeated, so that the source codes of the corresponding URLs are obtained by transmitting HTTP command to the corresponding URL using the HTTP.

If a HTTP command as in Fig. 22 is transmitted to the web server, misnt. gsnu. ac. kr, this web server responds by transmitting a HTTP response including the WBI2. html source code. Accordingly, the agent in accordance with the present invention obtains the WBI2. html source code. Now, the next hyper link can be obtained by analyzing the source code of"WBI2. html" file anew. Likewise, in case of connections via a selection type variable or a numerical variable, the URLs of hyper linked sites can be obtained through analysis of the <FORM> tag and of the script language, and the source codes can be obtained through HTTP. The web site management system as per the present invention can obtain web site information from an existing web site in this manner, and then, express the

same in graphs.

Important elements of a web page include, in addition to the interconnection with other web pages, the visual design of the web page, the contents, etc. Thus, it is significant that the visual design, the contents, etc. are properly managed apart from the management of the interconnection with other web pages. The automatic web page generating system in accordance with the present invention adopts the concept of template for this purpose. That is, connection parts are indicated by predetermined reserved words that connections are required, while the other parts are expressed using a template file which contains the whole contents of the web page. The automatic web page generating system expresses connection of only those parts desired to be connected and keeps the others untouched so that the contents of the web page are kept unchanged. A user can generate or change the template file by double clicking the title part of the node.

The template file used for the above cases is exemplified in Fig. 23. In the template file in Fig. 23, <! WGBRANCH> and <! WGINPUTFORM"CONFIRM"3 ALIGN=CENTER> are parts indicating the need of connection expression. The automatic web page generating system in accordance with the present invention maintains the contents of the web page unchanged by changing only the above parts for connection expressions and maintaining the remaining parts of the template file unchanged.

Furthermore, the template file needs not to be one in number, but rather, each web page may have different template file, so that these divergent template file can be stored and managed in a template database. Thus, web pages having entirely different contents can be connected to each other and be managed by expressing these different web pages in template files. In particular, in case an existing web site as described above is managed, the contents of a web page can be maintained unchanged by changing only the connection

parts of HTML file or DHTML file obtained through HTTP to reserved words for connection, and then, by storing the remaining parts as template files.

However, the reserved words in Fig. 23, <! WGBRANCH> and <! WGINPUTFORM"CONFIRM"3 ALIGN=CENTER> are simply examples for convenience of the explanation, and the present invention does not specify any concrete name or form of a reserved word.

On the other hand, another important element in web page screen design is dividing of a screen. Most of the web pages are divided using frames or tables.

Accordingly, a web site management system shall be able to express division of a screen in graphs as well. For this purpose, the present invention provides frame node as a constituent of the graphs. A frame node comprises horizontal frame node for expression of upper/lower division of a screen and vertical frame node for expression of left side/right side division of a screen.

Figs. 24a through 24c show examples of graphs for dividing screens using frame nodes, wherein Fig. 24a shows an example of an upper/lower division of a screen using a horizontal frame node, Fig. 24b shows an example of left/right division of a screen using a vertical frame node, and Fig. 24c shows an example of dividing a screen using a combination of horizontal and vertical frame nodes.

Figs. 25a through 25c show examples of screens formed from Figs. 24a through 24c, wherein an upper/lower division as in Fig. 25a is obtained using a horizontal frame node as in Fig. 24a, a left side/right side division as in Fig. 25b is obtained using a vertical frame node as in Fig. 24b, and a division at will as in Fig. 25c is obtained using a frame node with a combination of a horizontal frame node and vertical frame node as in Fig. 24c.

Here, the ratio of division can be set in the registration information screen, which is

displayed by double clicking a branch of the frame node.

Figs. 26a and 26b show examples of expressing Fig. 25a in HTML, wherein a frame is used in Fig. 26a, while a table is used in Fig. 26b, assuming the screen height to be 600 pixels.

On the other hand, a web page frequently comprises a plurality of menu bundles.

The present invention provides container nodes for expression of such menu bundles in graphs.

A container node expresses menu bundles of a screen having the same design in branches, as exemplified in Fig. 27. The container node in Fig. 27 comprises three menu bundles, A, B, and C. This container node is designed to comprise menu A in its upper part, menu B in its lower left part, and menu C in its lower right part, as shown in Fig. 28, wherein detail menus under each menu bundle can be expressed by selection type connections. Additional menu items can be added to by adding desired menu item to a corresponding selection type node.

The user can change the location of a menu in the screen by double clicking the title of the container node. A template file as in Fig. 29a can be generated by the menu location as changed by the user. The location of a menu can also be changed by change of the template file. A screen as in Fig. 28 can be constituted based on the template file and the menu item inputted to the selection type node. Fig. 29b shows an example of HTML file generated from the menu items inputted to the selection type node in Fig. 27 and the template file in Fig. 29a, expressing the same screen as in Fig. 28.

The above explanation has been made with examples of relatively simple web pages. However, a graphic expression of a complex web page would require a greater

number of nodes, and thus, the graph shall be more complicated. The maximum quantity of material that can be processed at once by an individual is known generally to be 72.

Accordingly, the number of nodes to be expressed in a graph shall desirably be limited to these figures. The present invention adopts grouping nodes, which can group a plurality of nodes to one group to this end.

Fig. 30 shows an example of graph expressed in a plurality of nodes, wherein the nodes belonging to figure 3001 are formed to one group, while the nodes belonging to figure 3002 are formed to another group. The graph thus obtained is illustrated in Fig. 31.

Here, node 16 corresponds to figure 3001 in Fig. 30, while node 17 corresponds to figure 3002. As shown in Fig. 31, the general structure of the graphs can easily be understood using the grouping nodes.

In order to view the detail structure of the grouping nodes in Fig. 31, a user has to double click the title of the corresponding grouping node. Fig. 32 exemplifies a screen generated by double clicking the node 16 in Fig. 31.

As shown in Fig. 34, connector nodes (3201,3202) in pentagonal forms are used to indicate connections to the external graphs in detail screen of the grouping nodes. The connector nodes indicate where the present graph is connected to external graphs, whereby a connector node represented with a negative pentagon (3201) indicates input from an external graph, while a connector node represented with a positive pentagon (3202) indicates output to an external graph. A user can jump to connector node by double clicking a corresponding connector node (3201,3202).

A grouping node differs from node grouping or question variable grouping in that the former is used for simplifying the graphs and does not affect the form of an actual web page while the latter affect the form of an actual web page, but not the graphic expression.

In realizing a web based inference system using the web site management system, the factual values required for progress of the inference shall not necessarily be inputted by a user. In case such factual values were inputted in a database in advance, the stored values can also be used.

The connection information of the database can be obtained by input of the table name, the attribute name, and the query conditions of the database on the node registration information screen, which appears by double clicking of each node, i. e. for fetching such values from a member information table consisted of member ID, name, address, and age for a member, e. g. whose member ID is"yusong", the required input shall be table name = "member information", attribute ="age", query condition ="member ID = yusong".

A web based inference and a database can be connected to each other in two ways.

In the first way, the inference proceeds to the next step after a value fetched from the database has been displayed to a user and this value has been confirmed by the user, while in the other, the inference proceeds immediately to the next step without confirmation of a user. In the following, an explanation of each of these ways is given exemplifying a connection of a database to a numerical node inquiring whether the age is either at or over five, or below than five.

(1) Inference Progress After Confirmation For inference progress after confirmation, current value of the database is shown in the input window and the user confirms it by pushing the"Confirm"button. An ASP program as shown in Fig. 33, for instance, can be generated from the web site management program in accordance with the present invention.

The part <INPUT TYPE=TEXT NAME="age"VALUE=<% =age% > in Fig. 33 enables a user to view the value as in Fig. 34 in advance by allotting the initial value of the input window with"age"value fetched from the database. Here, the instructions to fetch an

"age"value from the database are contained in <%... % >, which also can be automatically generated.

In case the stored age is 32, a screen as Fig. 34 appears.

In the screen of Fig. 34, the user can either confirm the value 32 in the age input window and then correct it, or immediately proceed to the next step of inference by just pushing the"Confirm"button.

(2) Inference Progress Without Confirmation In case no confirmation by a user is necessary, the inference can proceed immediately to the next step without requiring a push on the"Confirm"button by a user.

For such a procedure, the OnLoad event of a BODY tag in the DHTML is used. An example of ASP program for this process is shown in Fig. 35.

The part OnLoad ="WBIBranch (document. WBIForml)" in the BODY tag of Fig. 35 enables the inference to progress automatically to the next step, whereby it shall be noticed that the name of the FORM with INPUT tag, i. e.

WBIForml is used here.

Although the above description is made with embodiment examples using ASP for connection with a database, the present invention is not limited to a specific platform, but rather allows generation of an inference site with divergent platforms including JSP, PHP, etc.

The web site management system should be capable of storing all information pertinent to the graphs in order to perform the above-described functions. A description on the method of storing the graphic information and on the data to be stored follows below.

Since the contents stored in the files and those stored in the main memory device are

almost the same, they are not separated in the following explanation.

As stated above, a graph is consisted of nodes and arcs that connect between the nodes. The nodes and arc are stored respectively in arrays as shown in Fig. 36.

The node information, constituting the elements of the node array, comprises information on the node itself as well as information on the branch. The node information comprises the followings: i) node type: start node, true/false node, selective node, numerical node, frame node, container node, grouping node, connector node, conclusion node, etc. ii) node location: coordinate on the screen iii) node grouping information: node group number (ID) iv) question variable grouping information: question variable group number (ID) v) template file name vi) output file name vii) connected database information viii) information by node type - start node and conclusion node: title - true/false node and selective node: question, question form (check box, radio button, hyper link, etc.) - numerical node: question, variable name - frame node: whether or not of horizontal/vertical division, ratio of the screen divisions - grouping node: ID of the lower level detail graphs - connector node: node ID of the connected external graphs

ix) output control information, etc.: line change, font and color of letters, background color, etc.

Since a node has several branches, the branch information is stored in form of arrays for each node. For instance, as the first node (The young is fed with milk) in Fig. 13 has two branches, the branch information is stored in a form as in Fig. 37. The branch information comprises the followings : i) the type of node to which the branch belongs: start node, true/false node, selective node, numerical node, frame node, container node, grouping node, conclusion node, etc. ii) information by branch type: - branch of the start node: description - branch of the true/false node: answer - branch of the selective node: answers (to enable selection from a multiple of answers) - branch of the numerical node: conditional sentence for branching - frame node: title of the screen for the corresponding frame - container node: title of the menu bundle - grouping node: the final branch of the grouped nodes - branch of the conclusion node : explanation iii) output control information, etc.: line change, font and color of letters, background color, etc.

The connecting line information comprises the followings : i) type of the line: rectilinear line, bended linear line, etc. ii) connection information:

- preceding node: address (for storage in the main device) or ID (for storage in a file) of the immediately preceding node - preceding branch: address (for storage in the main device) or ID (for storage in a file) of the immediately preceding branch - succeeding node: address (for storage in the main device) or ID (for storage in a file) of the immediately succeeding node iii) output control information, etc.: line color, etc.

However, the above-described storage methods serve only as examples for explanation of the present invention and shall not be understood to restrict the same. The present invention can store information in various other forms.

Industrial Applicability As described above, the web site management system in accordance with the present invention allows to easily and rapidly develop a web site by expressing in graphs of the connections between web pages that constitute the web site and by generating automatically web pages from these graphs, and to easily maintain the web site in case the connections have been changed, by just changing the graphs and then, automatically re- generating the web pages. In addition, the present invention allows a web site already generated to be expressed in graphs as well, so that these graphs can be used for the purpose of management thereof.

Further, the present invention allows embodiment of a web based expert inference system by reverse application of the inference system used in the present invention to a web. Since the connections in the present invention support the connections between the variables in the inference, a web based inference site is obtained by an automatic generation of a web site after the knowledge of human experts has been inputted in the web

site management system as per the present invention in form of graphs, and then, the generated web site is uploaded.

Although the present invention has been described above with reference to the preferred embodiments and accompanying drawings, it should be noted that the scope of the rights of the present invention is not limited thereto, but rather shall be determined by the appended claims, allowing various alterations, modifications and adjustments as those skilled in the art will understand. For example, although the above description exemplifies HTML as the language describing the web page, the present invention can use various <BR> other languages such as XML, WML, ASP, JSP, PHP, Java Servlet, etc. , as well, not restricted to a specific language.