JPH03149634A - Hypertext system - Google Patents

Abstract

PURPOSE:To make it possible for a reader to efficiently read out a hypertext by automatically selecting and displaying a node to be read out next in matching with the reader's request or knowledge. CONSTITUTION:When the reader inputs an instruction for allowing a system to determine a node to be observed next without specifying the node required to be observed by the reader, a hypertext control part 5 determines the node to be displayed next based upon default bus information stored in a node access storage part 6 and displays the determined result on a display part 3. In the case of reading out the hypertext along link information or a guide line applied by an author, the node to be read out next is automatically determined by the system side in matching with the reader's request or knowledge even if the reader forcedly selects the link. Consequently, a node close to a reader's request can be effectively and preferentially traced.

Description

[Detailed Description of the Invention] [Object of the Invention] (Industrial Icheon Field) This invention relates to a hypertext system. (Prior art) For example, when searching an encyclopedia or language dictionary stored on an optical disk or magnetic disk by displaying it on the display,
When you are searching for a specific item, if you see a word or phrase you don't know appear in the description of the specific item displayed on the display, you may want to search further for that word or phrase. In such cases, when you specify a new word or item that you want to search using a click method such as a mouse, an explanatory text about the newly specified word or item will be displayed in a multi-window format on the display. , it becomes easier to search for various items, and it becomes possible to search for various items and phrases in the same manner as when reading an actual book. In addition, even in the case of a single document, such as a manual for OA equipment that is displayed on the display, the chapters that should be read from one specific chapter to the next are related to each other based on the characteristics of the equipment, and the operator If has finished reading that particular chapter, then rN
By predetermining the next chapter to be displayed using the EXTJ command, it is possible to easily handle the relevant OA equipment. Also, when you come across an unknown word while reading a particular chapter, you can click on the word and
It will be convenient for the operator if the explanatory text is displayed in a window format. A system that meets these demands is a hypertext system. In a hypertext system, a document is divided into several partial documents, or blocks, using paragraphs as units, and blocks are divided by connecting δ blocks with pointers. Build a document network structure with nodes and pointers as links. Then, a display window is assigned to each node. By reaching the node, the user displays the node in a window and reads the contents of the corresponding document block. Since there are links between related nodes, by following the links, you can quickly move to and read the document block containing related information. In general, links between nodes are from words or phrases to blocks,
Sentences are connected to blocks, paragraphs to blocks, and so on. By moving the cursor to the linked word, phrase, sentence, or paragraph and clicking a button such as the mouse, you can follow the link and click the link destination. Move to the next node where . Conventionally, in such hypertext systems, a document is read by following links from node to node, but the system also provides path information (a list of nodes) in advance that indicates the order of nodes to be followed as a guideline. It is also possible to eliminate the need for the operator to select and follow a link by assigning a link to a link. If you do this, when there are many nodes in a large document, you will not know where you are in the document and forget which node to go back to, or you may end up getting lost, or you may end up needlessly accessing information that is just a branch or a terminal node. can be prevented. (Problem to be Solved by the Invention) However, in conventional hypertext systems, creating links between such nodes and creating path information as guidelines are mainly performed by the creator of the hypertext, that is, the author. Because of this, it is difficult to reflect the different needs and knowledge of each reader.
There were problems in that readers had to read the same nodes over and over again, or they accessed nodes that they didn't need to read because they were about information they were familiar with, resulting in a lot of unnecessary access. This invention was made in view of such conventional points of understanding, and it refers to information on the reader's past access history and current context, and provides information that is more responsive to each reader's different needs and knowledge. The purpose is to provide a hypertext system that can preferentially follow nodes close to the reader's requirements without waste. [Structure of the Invention] (Means for Solving the Problems) The hypertext system of the present invention includes a database for storing text information, and the text information stored in this database is divided into nodes and managed by networking. A hypertext management section, a display section that displays the nodes processed by this hypertext management section, an input section that accepts input from the user for the nodes displayed in this display section, and an input section that indicates the order of nodes to be followed. Default path information specified as a guideline, access history information that shows when in the past each node was accessed and how long the node was read, and context information that shows the context in which the node is currently being read. It is equipped with a node access storage section for storing information, and an inference section that refers to the contents of the node access storage section and determines the next node to be followed. (Operation) In the hypertext system of the present invention, when a reader issues a command from the human resources section to specify a node that he/she wants to read, the hypertext management section interprets this, determines the node to be presented to the reader, and displays the node. Text information corresponding to the node is searched from the database and displayed on the display unit. Generally, the reader manually clicks on the link of the node he or she wishes to view next, and in this case, the designated node can be displayed according to the reader's instructions. On the other hand, if the reader does not specify the node he or she wants to see next, but manually instructs the system to decide which node to see next, the hypertext management section uses the default path stored in the node access storage section. Based on the information, the next node to be displayed is determined and displayed on the display section. At this time, the inference section refers to the access history information in the node access storage section and the context information of the node that the reader is currently reading, and determines the next node to follow according to each reader's needs and knowledge. decide. In this way, when reading the hypertext according to the link information and guidelines provided by the author, the reader does not have to choose a link (but the system automatically selects the next node to read according to the reader's needs and knowledge). You can decide to access the same node repeatedly,
There is no need to access well-known nodes over and over again, and nodes that are close to the reader's needs can be preferentially followed without waste. (Example) Hereinafter, an example of the present invention will be explained in detail based on the drawings. FIG. 1 shows the system configuration of an embodiment of the present invention, which includes a database 1, a human power device F2 including a keyboard 2a and a pointing device 2b such as a mouse, a display section W3, and a hypertext processing section 4. The hypertext processing unit 4 is further comprised of a hypertext management unit 5, a node access storage unit 6, and an inference unit 7.Next, regarding the operation of the hypertext system having the above configuration, explain. Database l is divided text. is stored in node units. Information regarding this node includes the author name of the node, reader access information, and the like. Link information between nodes is also stored in this database 1. The display device 3 is a CRT or the like with a multi-window function, and is also capable of displaying networks between nodes.The reader can view the text information of each node displayed on the display device 3. The pointing device 2b of the human-powered device 2 points to the nodes and links of the device i13.
, specify the next node to read, and display that node on the display device 3. The hypertext processing unit 4 determines the next node to be displayed (the node the reader wants to read) according to the reader's request, searches the database 1 for text information regarding that node, and displays it on the display device 3. The main processing of this hypertext processing unit 4 is performed by the hypertext management unit 5, and the display is performed so that the reader can proceed to the next node he or she wants to read while looking at the nodes already displayed on the display unit W3. Specify from the input device 2 a link that appears in the node. This input device 2
The link is specified by clicking a button on the pointing device 2b. The specified link is a link to the next node to be read, and the hypertext management unit 5 searches the database 1 for nodes connected to the displayed node by the specified link. Then, the newly obtained node is sent to the display section W3 and displayed. Regarding the nodes displayed on the display device 3 in this way, the displayed date, time, read time, etc. are stored in the node access storage section 6 as access history information. The above node search process is carried out by the hypertext management unit 5.
is as simple as directly following the links between adjacent nodes, but the reader may manually issue a command that causes the system to determine the next node to read.In this case, the reader may manually issue a command from the keyboard 2a. or by selecting a menu using the pointing device 2b. In this way, when the system side is entrusted with determining the next node to be displayed, the inference section 7 refers to the default path information stored in the node access storage section 6.
determines the next node to follow. In this case, the access history information and context information stored in the node access storage unit 6 are referred to (if necessary, information regarding candidates for the next node is searched from the database l), and the next node to be traced is determined. The information is determined according to the needs and knowledge of each reader, and the inference section 7 transmits the result to the hypertext management section 5. The hypertext management section 5 searches for the node determined by the inference section 7 and displays it on the display device 3. FIG. 2 shows an example of a display screen of the display device 3, in which three nodes, a process 85, an interprocess communication 103, and a system call 31, are displayed using a multi-window. A portion of the network NW created by links between nodes is displayed in the lower window. It is also possible to browse the display of this network NW and find a node of interest. Figure 3 (a) shows a logical network created by links between nodes, and Figure 3 (b)
shows an example of a display on the window of one node "Inter-process communication". This window is divided into three parts A, B, and C. The top A shows the title "Interprocess Communication" and its node number (103). The middle part B below it is a window consisting of icons representing control buttons, and commands can be issued by moving the cursor to the button icon and clicking the button on the pointing device 2b. For example, clicking on the rNextJ icon issues a command that causes the system to determine the next node to view. The bottom part C is text information possessed by the node. In this text information, you can see words and phrases surrounded by boxes, such as "pipe" and "socket," which indicate the existence of a link to the next node. If you move the cursor to this word or phrase and click the button on pointing device 2b, the nodes connected by this link will be displayed and you can move there.
Here, as can be seen from the network shown in Figure (a), "Pipe (107) J and [Socket (121)
)”. FIG. 4 shows a data structure for managing nodes. In other words, every node is uniquely numbered, and the table NSD consists of pointing from the node number NN to data NS indicating the entity of the node.
It is equipped with T. The data NS indicating the entity of the node includes a title for each node, a list of link source nodes, a list of link destination nodes, the file name in which text information is stored, and the words and phrases that are links. Lists etc. are written. Therefore, what kind of operation processing is executed when % rN ex t J is clicked in the sub-window display state of FIG. 3(b) will now be explained. When the reader clicks rNextJ, the system side uses the inference unit 7 to determine the next node to view. At this time, the default path information DPI shown in FIG. 5 stored in the node access storage section 6 is referred to. The default path information DPI is a guideline recommended by the author. It is an ordered list of node numbers, and the author proposes that nodes be read in the order written in this list. The next node is determined according to this guideline. Therefore, as an example of the operation process, the operation of determining the next node to be displayed according to the default path information of guideline number 15 will be described below. Guideline No. 15 recommends reading 85, 103, 107, 108, 151, etc. as the next node number to read, and when compared with the data in Figure 4, this indicates that "process", ""InterprocessCommunication", [PipeJ, rpipeJ, rdupJ,
It corresponds to reading on. Here, since we are currently reading [Interprocess Communication (103) J], when we click "NextJ", "Pipe (107)" is recommended as the next node to be read in the default path information DPI. However, with this method, when you click "NextJ, [Pipe (107)J] is always selected, and the difference between the reader is a beginner or an expert, and the reader's current context (this time ``pipe'' may be accessed and read in a different way.'')
For example, it is impossible to dynamically change the node to be selected based on the reader's knowledge, such as someone who has read Pipe J many times in the past and is familiar with the content. In the embodiment, access history information AH shown in FIG. 6 stored in the node access storage unit 6
The inference unit 7 refers to the RI and the context information CC shown in FIG. 7, and dynamically selects and accesses nodes according to the reader. First, FIG. 6 shows the correspondence between the node number table NSDT and the node access history information AIRI of each node, and the node access history information AHRI of each node includes the date of the last access (1ast-access) and the The access time (seconds, unit: seconds), total access time so far (total, unit: two hours), list of access dates (dates), and information on how many times it was accessed (access-count) are stored. There is. In addition, Figure 7 shows the context information CC■ indicating the current access status, including the number of the default path information currently being referenced (guideline number), the number of the currently accessed node (Current-node), and the current access status. A list of dot pairs of the accessed node number and the time the node was accessed (visited)
-nodes) are stored. Now, on the screen shown in Figure 3(b), select "NextJ".
When is clicked, the inference section 7 operates based on the flow shown in FIG. It is assumed that the current date is September 90, 1989. First, the context information CC stored in the node access storage unit 6
From ■, select the node currently being accessed.
-node, and the currently referenced default path information number is default-path-no (step S11). In the embodiment shown in FIG.
It is assumed that the rent-node-103, that is, the node "interprocess communication" is being accessed. Also, de
fault t-path-no, = 15, that is, the default path information is (85, 103, 107, 108, 1
51). Next, add the list of default path information specified by default-path-no to the table DPI in Figure 5.
Next-nod is the node that should be searched and displayed next to the node that appears next to current-node.
e (step S12); in this embodiment, nex
This is t-node-107. Next, a determination routine determines whether this next-node is appropriate as a node to be displayed next (step S13). Here, if it is determined that the node is not an appropriate node, the next element in the default path information list is next-n. (Step 512), and the determination is repeated until a suitable one is found (Step S13゜814).
If it is determined in step S13 that the node is suitable as a node to be displayed, information regarding this node is searched from the database 1 and displayed (step S1
5). Next, detailed processing of the determination routine in step S13 will be explained based on the flowchart of FIG. First, when next-node = 107, if we refer to AHRl in Figure 6 for the access history information of this node [pipe (107) Last accessed,
At that time, it can be seen that m1 has been read for 50 seconds, and that it has been read for a total of 500 seconds (step S21). So, next is [500≧TOTAL? J is determined (step S22), and since it is set to TOTAL-300 seconds, this determination step is YES, and it is determined that the node is inappropriate as the next node to be displayed. In other words, this node "pipe" has been read many times in the past, so the reader knows its contents well, and it is determined that it is a node that does not need to be read this time. Next, the next-node selected is D in Figure 5.
The next node in PI guideline number 15 is “pi
pe (108)J. And this node is the 6th node
Referring to the figure, the total past access time is 260 seconds, the determination in step S22 is No, and then it is determined whether it is the node read this time (step S2
3). This node rp i p eJ is the node read this time from the context information CCI shown in FIG. 7, and is read in step S23.
The determination is YES, and the process proceeds to the next determination step S24. In this step S24, the time currently read from the CCI regarding node rpipeJ is 40 seconds, and 40 seconds.
≧ SHORT? However, SHORT-30
If it is set to seconds, the determination result will be YES, and it will be determined that the node is unsuitable as the next node to be read. In other words, since this node rpipeJ has been reading for 40 seconds this time, it is determined that there is no need to read it again. However, if it is determined in step S24 that the current read time is shorter than the time set as SHORT, that node is determined to be displayed again (step S25). The next-node selected next is the next node rdup (1
51) J and the node access history information AH in FIG.
Referring to the RI, this node has a total access time of 15 seconds and has never been accessed this time either. Therefore, for this node rdupJ, step S
It reaches the determination step 26, but since it has been accessed in the past, it receives a YES determination result and goes to step S.
27. In this step S27, the determination is different depending on when it was read in the past, but this node [dupJ was last read in August IB, and FORGET-DA
If you set it as MS-30, rFORGET-DA
"before YS", the determination result is YES, and the process moves to step 528. In this step S28, it is determined whether the time last read was long or not, and if it is set to LONG - 180 seconds, the node rdupJ is last read for only 5 seconds, so "5≧LONG". The determination of ?J is NO, and this node [dupJ is determined as the next-node (step S25).However, in the determination of step 528, if the last read time is longer than LONG-180 seconds, Although the node was read more than 30 days ago, it has been read for a long time and the content is well known, so there is no need to read it again, and it is considered inappropriate as the next node to read. In addition, the node to be judged is FORGET-DAM
If the node has been read more recently than S-30th (step S27), if the node has been read for a MED time shorter than LONG, for example 120 seconds or more, the node still remembers what it read well and there is no need to read it again. is next
- judged to be unsuitable as a node (step S29)
. In the above example, FORGET-DAYS-30 days TOTAL - 300 seconds LoNG - 180 seconds MED - 120 seconds SHORT - 30 seconds, but these values are variable and can be set as appropriate depending on the reader and the content of the document. can be changed to something. however,
Generally, it should be TOTAL>LONG>MED>SHORT. As described above, according to the embodiment of the present invention, the reader can
By simply clicking rNextJ without selecting a link in minutes, the system will select and display the most appropriate node to read next. Moreover,
At this time, since the reader refers to the past access history and the current access context, the reader's needs and knowledge can be flexibly reflected in node selection. Therefore, users do not have to visit the same node over and over again or read familiar information over and over again, allowing for efficient access. Note that in the above embodiment, only one node to be recommended is selected, but it is also possible to weight the determination and present the nodes as display candidates in descending order of priority. Furthermore, the algorithm of the determination routine is not limited to that of the above-mentioned embodiments, and flexible and complex determinations can be made by creating rules for determination using a production system, for example. Furthermore, if node access history information is managed for each reader, nodes can be automatically selected for each reader even when the hypertext system is used by multiple people. [Effects of the Invention] As described above, according to the present invention, when reading hypertext according to the link information and guidelines provided by the author, the next node to be read is automatically determined according to the reader's needs and knowledge. Since nodes are selected and displayed, the reader can select and display nodes according to the reader's needs and knowledge without having to select the links that suit him/her, and the same nodes can be displayed selectively according to the reader's needs and knowledge. visit repeatedly,
Readers can read hypertext efficiently without having to read familiar information over and over again. It also remembers the context of what you're currently reading, so you don't have to worry about getting lost while following links, even when you're reading a large piece of hypertext.

[Brief explanation of the drawing]

Fig. 1 is a block diagram of an embodiment of the present invention, Fig. 2 is an explanatory diagram showing an example of a screen display of the display device of the above embodiment, and Fig. 3(a) shows creation of hypertext nodes of the above embodiment. An explanatory diagram showing an example of a network, FIG. 4B is an explanatory diagram showing an example of display of one node in the network, and FIG. 4 is a data structure of data managed by the hypertext management section of the above embodiment. 5 is a data structure diagram of default path information in the above embodiment, FIG. 6 is a data structure diagram of node access history information in the above embodiment, and FIG. 7 is a data structure diagram of context information in the above embodiment. FIG. 8 is a flowchart showing the inference operation of the inference section of the above embodiment;
FIG. 9 is a flow chart showing the operation of the next-node determination routine of the inference section of the above embodiment. 1-Database 2--Input device 3...Display device 4--Hypertext processing section 5...Hypertext management section Low--Node access storage section Fu inference section Patent attorney Hidefu Miyoshi table 1 praise ? l / 3 High 8° - 11th strike〒-RIBU'1J] Fig. 1 / 糾 35 Fig. 3 (a) 1 positive - = Tsuzuki 4 Fig. 3 (1)) μi yo saliva DT, NS No. Figure 4 Kite line @ Figure 5

Claims (1)

[Claims] A database that stores text information; a hypertext management unit that divides the text information stored in this database into nodes and networks them for management; and nodes that are processed by this hypertext management unit. A display section to display, an input section that accepts input from the user for the nodes displayed in this display section, default path information that specifies the order of nodes to be followed as a guideline, and information about when each node was accessed in the past. a node access storage unit that stores access history information indicating how long the node has been read and context information indicating the context in which the node is currently being read; A hypertext system comprising: an inference section that refers to and determines a node to follow next.