JPH09288556A - Visualized system for hyper media - Google Patents

Abstract

PROBLEM TO BE SOLVED: To minutely observe a specified node by arranging nodes being information units on a two-dimensional plane, linking them, raising the two-dimensional plane and developing it on a three-dimensional space. SOLUTION: The nodes being the information units are arranged on the two-dimensional plane and they are linked. The specified node in the nodes on the two-dimensional lane are raised and it is set to be a three-dimensional model. When a fermented soybeans view node-link connection 1 is started, for example, the default values of the user are selected as noticed nodes 3 and 4. When the noticed nodes 3 and 4 are raised, nodes 5-8 connected to them are raised. FIFO for controlling the view 1 from an external program is opened. The view 1 requests the acquirement of link information near the noticed nodes to a mapped system server link manager. Link information is analyzed and 'fermented soybeans view' is displayed based on a response from the link manager.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention links information that is managed independently for each site, and lifts a hyper network consisting of linked organic information networks for each specific node. The present invention relates to a hypermedia search system for visualizing a three-dimensional model.

[0002]

2. Description of the Related Art A conventionally known hypermedia system on the Internet (Hypermedia system)
The system (hereinafter referred to as WWW) has succeeded in realizing a world-wide organic information network by linking information managed independently for each site two-dimensionally.

[0003]

Since the information network has become a vast information space in which the links of each node are intricate, there is a problem that the relation between the information cannot be grasped.

For example, when a specific node in a link is expanded, peripheral nodes other than the specific node become unclear, which makes it difficult to understand the global information structure. 2 more
The information developed three-dimensionally has a uniform visual field and recognition, and it is difficult to observe from different directions by changing the viewpoint, so there is a problem that the information cannot be standardized even when visualized.

[0005]

Therefore, the present invention is based on the above W
They succeeded in visualizing the WW in a three-dimensional space so that they could be observed, and solved the above problems.

That is, the present invention is characterized in that nodes, which are information units, are arranged in a two-dimensional plane and linked to each other, and then a specific node among the nodes in the two-dimensional plane is lifted to form a three-dimensional model. It is a media visualization system. According to another invention, nodes, which are information units, are arranged in a two-dimensional plane and linked to each other, and then a specific node among the nodes in the two-dimensional plane is lifted to make a three-dimensional model, and the lifted node is lowered and the like. It is a visualization system of hypermedia, which is characterized by making it possible to take a bird's eye view of the information structure from different viewpoints by an interactive operation of lifting the node. In another invention, nodes, which are information units, are arranged in a two-dimensional plane and linked to each other, and then a specific node selected from the nodes in the two-dimensional plane is lifted to make a three-dimensional model, and this three-dimensional model is rotated. Or make a partial enlargement to overlook the specific node,
It is a hypermedia visualization system characterized by making it easy to see.

The present invention is intended to solve the problems in the conventional WWW, but in order to understand its characteristics, the problems will be described.

The WWW is an extremely flexible system based on human associative memory. It can be said that hypermedia has become a core application of the Internet because of its high degree of freedom.

However, users are still "lost" in hypermedia. I don't know where to look to get the information I want, and I don't know where or how to go to register new information. Such deficiencies are becoming more and more serious with the diversification of information.

Most of the conventional hypermedia are closed in one application or one group. Therefore, it was possible to construct an orderly hypermedia by using a unified creation system and guidelines based on the data model. In addition, even in the case of wide and open hypermedia that has been released to many people, there is no change in that it needs to be constructed by a professional writer with technical skills based on the unified guidelines.

The hypermedia has become widespread in the form of WWW, and with it, anyone can write the hypermedia and transmit information. And individual users had a completely free environment without having to be caught by the guidelines at all.

As described above, anyone can write "hypermedia""freely" and a wide open environment has been set up, so that the problem of getting lost in hypermedia becomes more serious. For example, even if one group adopts a uniform data model, other groups on the WWW build hypermedia under completely different guidelines, so they no longer define a uniform approach. This alone cannot solve this problem.

[0013] Furthermore, what has changed with the WWW is that hypermedia is no longer a finite world. It is still evolving day by day into an information space that is almost infinite.

From this point of view, it can be said that almost all existing hypermedia theories have been established on the premise that hypermedia is a finite world. The existing theories are "design theory" and "search theory," but "navigation theory" represented by graphical maps
It was completely dependent on its finiteness.

In the conventional navigation systems, it is an implicit understanding that almost all nodes and links are known (can be revealed). Therefore, it has been assumed that the user can always grasp the entire structure of the system even if he / she gets lost in hypermedia.

However, such a premise does not apply in WWW. Not only the user, but even the system
It is no longer possible to get a complete picture of hypermedia. Hypermedia has become a completely "unknown land."

As described above, the stray problem of hypermedia is becoming more and more serious. As a solution to this problem, the present invention proposes two approaches that interpolate each other.

(1) Information Visualization: Hypermedia, which deals with the relationship between information and information, has a complicated graph structure, so it is necessary to devise a way to make the overall image look as simple as possible. (2) Collaborative search: Multiple users who deal with human relationships form a group and search for information, thus sharing the search work and preventing information from being overlooked and efficiently collecting information. To do.

In other words, the technique for visualizing the information space can be said to be an easy-to-understand expression of the relationship between information. Regarding information visualization, as a typical method, it is possible to improve search efficiency by an information visualization method of displaying a huge amount of information that exceeds the processing capability of the human in an easy-to-understand manner.

If the visualization focuses on the relationship between the information items, the technique for facilitating the collaborative search is:
It can be said that it focuses on the relationship between humans.

Regarding the need and approach method of collaborative search, it is also possible to increase the efficiency of the search by the help of multiple colleagues of the group with each other.

However, since the structure of hypermedia is not so simple that it is represented in a two-dimensional space,
He also said that there was an opinion that it would be almost impossible. Certainly, there is a certain point in that way of thinking.
Hypermedia is an extremely complex graph (network) structure, and there is no standard for its production. In the end, I wonder if there is a visualization method that respects that hypermedia has a graph structure.

In fact, in many existing systems, hypertext is forcibly layered or linearized. However, these methods should not overlook that hypermedia is downgraded to a general non-hypermedia database system. Such a method diminishes the viewpoint of the relationship between free information in hypermedia, and is not the form that the user really wants.

A general problem with the layering and clustering method is that the entire structure of the hypermedia, or almost the entire text, must be known.

Unless all the link structures are known, it is not possible to proceed with the division from the connected state,
Unless all the text is known, it is not possible to determine the characteristics of the text set and split it.

In other words, the hierarchical method can only be used in a limited way in a certain range in open hypermedia such as WWW that spreads over a computer network. It is impossible to get all the information in a computer network in terms of the amount of access to the network and its response time.

Now, it is examined whether or not the graphic fisheye lens view, which is an excellent view of the graph structure, can be applied to a hypermedia system spread over a computer network such as WWW.

The fish-eye lens view can be seen as a structure suitable for hypermedia because it can know the global structure and details of details at the same time. However, it is difficult to use this graphic fisheye lens view for hypermedia. Because the biggest problem with the graphic fisheye view is that it targets a single "image". Furthermore, it is premised that the graph structure drawn in the image is arranged in a regular manner on a plane and rarely changes. On the other hand, under access via a computer network like WWW, it is almost impossible to examine the entire structure of hypermedia in advance. Therefore, it cannot be expected to arrange nodes and links on a plane in an orderly manner. Moreover, even if the overall structure of the Yoshiba hypermedia is known, it is not a simple substitute that can be placed on a two-dimensional plane like a highway network. Moreover, the structure of hypermedia in an open environment is constantly changing. If the optimal node arrangement is calculated every time the hypermedia structure changes, the map will change each time. Maps that change frequently do not serve as maps.

Accessed node (information) in the above
From the beginning (incrementally), that you can create a map without knowing the link status of all nodes, and in computer networks,
Access must be minimized from the two viewpoints of communication volume and response time.

(1) Preparation of map (arrangement of nodes, etc.)
Can be done by simple processing. Relevant to being incremental. (2) New additional data such as node attributes should be kept to a minimum, fully functioning on existing existing systems, and systems that require new attributes should It will not be available on computer networks until it spreads. (3) It is a model that respects the concept of nodes and links in hypermedia, and some existing graphical maps are very intuitive to humans in order to ensure theoretical consistency as an algorithm. Some present a non-view. But,
In the present invention, it is important that the node, which is the basic information of hypermedia, and the link, which is its relation, be visually recognizable. (4) It is desirable that the system has an effect similar to that of a fisheye lens view, and has a global structure like that of a fisheye lens view and a system that allows detailed details to be understood at the same time. For that purpose, it is necessary to support the concept of “attention” and to display the vicinity of the “attention” node conspicuously. (5) Consistency that the same node is always in the same position, and if the display of the node position etc. changes each time the navigation application is executed, the user will be You will not be able to use your memory. Then, since the user gets lost, in order to utilize the memory of past access,
Positional consistency must be ensured. (6) Even if the connected state of the nodes changes, the change of the whole image (scenery) of the map is minimal. Since the connected state of one node has changed, the positions of all the nodes are recalculated. If it is necessary, the response time will be long and the positional consistency described above will not be secured. (7) Supports multi-users, has a means to easily display where other users are on the map, and this is related to collaborative search. To proceed, it is necessary to be able to easily display in the map where other users (group members) are. (8) Not only the system's preparation, but also the ability to interactively take a bird's-eye view of hypermedia, and in a hierarchical system, enforce the viewpoint according to the hierarchy (classification method) presented by the system. Some are done. In navigating the information space, it is desirable to be able to take a bird's-eye view of the entire structure from various viewpoints, and it is also desirable to be able to do it interactively.

In the present invention, it has become clear that there is a limit to the attempt to create a hypermedia graphical map in a two-dimensional plane. Therefore, I came up with the idea of using a three-dimensional space instead of a two-dimensional plane.

In fact, in recent researches on information visualization, the use of a three-dimensional space has been enthusiastically attracted attention. He invented the "natto view" based on the idea that the arrangement of nodes and links, which would be messed up in a two-dimensional plane, would work in a three-dimensional space.

In the natto view, hypermedia is modeled just like natto. What does that mean?
You can compare "nodes to natto beans" and "links to natto thread". If you focus on a certain node and try to extract it, the size is such that the node connected to it by a link and associated with it slips along like natto.

The basic items of the natto view creation algorithm for realizing the natto model will be described below.

(1) Utilizing a three-dimensional space. (2) Nodes are represented by spheres, and links are represented by lines. (3) First, the nodes are arranged on the xy plane. At this time, the uniform representation method of information location (Uniform Resou
rce Locator (hereinafter referred to as URL). (4) When the node of interest is pulled up (in the z direction), connected nodes will follow. (5) The coordinate system can be viewed from a position that is easy to see by enlarging and rotating the coordinate system.

There are various other seasoning parts, but this is the basic of the natto view. The point of this simple natto view is that the user can operate interactively.

The natto view supports the "attention" of the user. When the user decides which node to "focus on" and selects it, that node and the links that exit it are displayed in a distinctly different color.

Further, by lifting the node, the node of interest is taken out from the bottom random cluster of nodes, and the link going out from that node becomes clearer.

Then, if the node is further lifted, the node connected to the link that goes out from the node of interest is slid out and displayed so that the relation of the link can be clearly understood.

As described above, in the natto view, by selecting and lifting a node of interest and rotating the coordinate space, the user can take a bird's eye view of the overall structure of the entire hypermedia, and can see the vicinity of the node of interest. It is possible to take out and enlarge the state of.

A special point of the natto view is that the hypermedia space can be viewed from various angles by the user's interactive operation.

In many existing layered systems, the system is automatically determined in consideration of the user's convenience in terms of which node is used as the starting point for layering. Even in a system specified by the user, there is a long waiting time from the specification of the starting node to the end of the layering process.

However, in this natto view, the user can proceed with the hierarchization process little by little during the interaction with the computer. Furthermore, the process is
Since the operation of pulling up the node is extremely easy to understand, the user is not burdened with unnecessary cognitive load.

In the natto view, the user interactively selects and determines an arbitrary node of his or her interest, and interactively examines how the other nodes are related to each other starting from the node. be able to. In other words, the natto view allows one hypertext to be viewed in various ways from various viewpoints.

The natto view can easily deal with multi-users, who are important in promoting collaborative search. In order to display where the group members are, you can change the color of the node.

Next, as a final explanation of the natto model and the natto view which is its display form, it will be described that it satisfies the requirements for a navigation map.

The natto view has the following characteristics in terms of compatibility with computer networks.

(1) The nodes can be created sequentially (incrementally) from the accessed node. (2) The map can be prepared by a very simple process. (3) There is no new additional data such as the attribute of the node.

In terms of user interface, there are the following characteristics.

(1) The concept of nodes and links in hypermedia is respected. The display shows the nodes and links as they are. (2) It supports the concept of attention like a fisheye lens view. (3) There is positional consistency in the xy coordinates. Therefore, when the natto view is viewed from directly above (z +), the same information is always in the same position (FIG. 3). (4) After the connection state of the nodes changes, only the nodes related thereto are affected. (5) A relatively large amount of information can be displayed in the three-dimensional space. (6) Multi-user support. (7) It is possible to overlook the hypermedia interactively from various viewpoints.

Many of the above-mentioned searches mainly focus on improving the individual's search ability. However, it cannot be denied that there is a limit in individual ability. For example, when researching something in daily life, it is clear that it is more efficient to perform a search in cooperation with several people instead of one person. Therefore, we tried to incorporate this concept of cooperation into information retrieval in hypertext. Generally, the larger the amount of information and the wider the search range, the more difficult the search becomes. In such a case, it is more efficient to search by several people rather than one person.

Since there is a limit to the search ability of an individual, in addition to the idea that multiple searches are efficient, this collaborative search has the advantage of more fundamental things. The search methods so far inevitably have some limitations because they are computers. It is a human being who searches, and there are all kinds of images and ways of understanding even one word, and it is exactly ten different colors. It is extremely difficult to analyze it using only numerical data.

Of course, there are some servers that have a searchable meta index as a means for searching information on the WWW. However, since this is also a keyword search, it is common that unnecessary information comes in or necessary information cannot be obtained.

When some of the users work together to do something, “cooperative” is more familiar than “cooperative”. For example, when searching for a book in the library, it is often the case that several people sort out the desired book. In addition, most of the systems that have been considered up to now have involved sharing and searching information within a very limited group with the same purpose.

However, for the future information society,
It will be necessary to share information and search in a wider range, not only within those groups. It also shows that human interests and interests are diverse. Even in the same group, the information that you want to retrieve is usually different depending on the role within the group. In other words, it is necessary to support people with different search purposes.

Therefore, it is important to gather people with the same purpose to perform “cooperative” search, but it becomes necessary to gather people with various purposes to perform “cooperative” search. . In order to carry out a joint search, it is necessary to first gather people with similar goals.
However, it is not always the case that there is a fellow with the same purpose, and even if there is one, it may not be found immediately. That doesn't mean you can always search when you want.

Therefore, "cooperative search" is required so that people with different purposes can search with their strengths at any time.

The following necessary conditions can be considered for the collaborative search.

(1) Mutual understanding of search objectives: A mechanism for teaching each other what kind of information the group members want. (2) Mutual understanding of search history: A mechanism that allows you to list which information is being accessed by fellow group members. (3) Exchange of evaluation information: A mechanism for teaching the existence of useful information to fellow group members. And in addition to these, this time, the following points are focused on. (4) Consideration of relationships among searchers: A mechanism that supports communication according to the content of information being searched and the human relationships between searchers.

The collaborative search is an embodiment of the present invention in order to satisfy the above-mentioned requirements, but in order to do so, an interface that can display each member information of the group as a list and perform necessary communication is provided. To do. The interface design takes the following into consideration.

In the present invention, the efficiency of collaborative search is considered to be improved by making communication of intentions smooth. The most important part of communication is to tell exactly what you want to search for. In other words, it is essential for collaborative search to indicate what kind of information you are looking for. Conversely, it is so important that it can be said that there is no way to cooperate without understanding this.

Therefore, in this system, a short comment for information to be searched is displayed.
This is also an important point, and it is clear that sentences are more effective than words such as keywords to convey things accurately.

This is because in a very huge space such as WWW where oneself knows what kind of information is being searched for and where it is, it is not possible to know where there is powerful information. In such a situation, it is inefficient to search for the same place as or around other members, so it is better to search from where no one else has access, that is, from all angles. There is a high possibility. Even in this respect, cooperation seems to have great merit.
Next, within the group of members, the person with a higher position (equivalent to a position in the company, etc.) is given priority. People with high status, that is, they are in a position to control and manage the initiative of the group, and by giving information to those people first, the direction of the group will be clarified and the work will proceed smoothly. is there.

Give priority to a person who frequently provides powerful information to himself. Since this contributes to that much to myself, if you do not give back to the person who got the information, it is possible that the information provision from that person will be cut off thereafter. .

If information is not exchanged in this way, only a specific person will help other members in addition to his / her own search work, or conversely, he / she will not care about the members themselves. There is an imbalance in the amount of work, such as performing only the search work of and only receiving information sent from other members. It is important to consider the degree of contribution to oneself in order to prevent it and keep the human relationship smooth.

First, when there are a plurality of people, it is difficult to pay attention to each of them, and when the number of people increases, it becomes even more difficult. Therefore, in the system of the present invention, it is impossible to perform a search while considering all the displayed members. Therefore, we tried to solve this point by giving priority to the display of members.

In promoting collaborative search, communication among searchers (group members) must be supported as a system. Therefore, in order to measure integration with the following basic communication tools on the Internet, asynchronous communication by e-mail and synchronous communication by conversation or communication are required.

Needless to say, the searcher himself / herself must be able to communicate by exchanging information while actually accessing the Internet and performing the search work, but if the searcher himself / herself is not actually involved in the search work, In other words, it should support asynchronous communication.

By doing so, it can be expected that the actual work time is shortened and the efficiency is improved. Further, since it is only necessary to search freely according to the progress of the work of each member, it is not necessary to adjust each schedule in order to work as a group.

In order to realize this function, when the searchers are both engaged in collaborative search (logged on), they work synchronously, and when they are not, they temporarily store the data asynchronously. Dedicated messaging system is required.

[0071]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention raises and lowers a two-dimensional plane hypermedia in a three-dimensional space, rotates it as necessary, or partially enlarges it to operate interactively to clearly grasp the desired information. It is a hypermedia visualization system that is designed to do.

[0072]

Embodiment 1 An embodiment of the present invention will be described with reference to the drawings.

A complete view of the system used to implement the invention is shown in FIG.

The implementation is roughly divided into two parts.
That is, NattoView 1 that presents a "natto view", which is a bird's-eye view of the hypermedia space by information visualization, to the user, and a collaboration river that provides a communication tool for collaborative search used as an example ( CollaboRiver).

The implemented program is a Sun Spark Workstation (SunSparc Wor) manufactured by Sun Microsystems.
kstation) and compatibles developed under UNIX and X window system (X-Window)
It is implemented using the function of System). In the implementation of Nattoview, SGI's OpenGL (Op
The free clone of the enGL library, the Mesa 3D graphics library, is used for 3D processing. Also, in the implementation of the collaboration river O
SF / Motif is used.

The processing associated with the activation of Natto View is as follows.

(1) When Natto View is activated, first, the default values of the user are selected as the attention nodes 3 and 4. When the node of interest 3, 4 is lifted, the nodes 5, 6, 7, 8 connected to it are lifted (FIG. 2). (2) FIFO (hereinafter control FIFO) for controlling Nattoview from an external program
open. (3) Nattoview requests the mapping system server link manager (LinkManager) to acquire the link information near the target node (performed by the SCAN command of the link manager). (4) Based on the response from the link manager, the link information is analyzed and the "natto view" is displayed. (5) Enter the GUI event loop that waits for input from the user.

The control FIFO is used for controlling the display of the natto view by the control panel of the natto view and a collaborative river which will be described later.
You can perform almost all operations on Natto View.

To access the link manager,
It is done according to the protocol by TCP connection. FIG. 2 is an overall view of the browser Natto view for realizing the Natto view.

In Natto view, the node is a sphere,
The link is displayed as a line segment connecting the spheres. The display method will be described below.

In FIG. 2, a plane displayed in a checkerboard-like pattern has an x-coordinate of 0-
It is the y-plane 2. First of all, the node is this xy
The position in plane 2 can be determined. The procedure is as follows.

(1) Obtain the URL of the node. (2) Obtain a file name that does not include the directory name connected by / from the URL. That is, in the URL /
Find the last string that does not contain. (3) However, when the URL ends with a directory name that ends with /, the last character string that does not include / is similarly obtained for the directory name. Similarly, if the URL ends with a host name, the host name is temporarily used as the file name. (4) The first character and the second character of the character string obtained as described above are regarded as the coordinate values (x, y) of the checkerboard and moved to the cell there. Links-www → (“L”, “I”) (5) Next, the third and fourth characters of the character string are similarly regarded as (x, y) coordinates, Move the same way. That is, the cells are arranged in one square as if they were arranged on the checkerboard. Links-www → (“L”: “N”, “I”:
"K") (6) Similarly, the position is recursively determined. Anything other than the alphabet is treated as "?". Links-www → (“L”: “N”: “S”:
"W": "W", "I": "K": "?": "W")

By using such a procedure, the (x, y) coordinates of the nodes can be uniquely determined and the nodes can be distributed so as not to overlap each other as much as possible. Further, since the URL is used instead of the determination of the position of the node at random, it is convenient at the time of searching.

In fact, when the natto view is viewed from directly above (FIG. 3), the position of the node does not move no matter what operation is performed, so the memory in the past access can be effectively utilized.

Nodes involved in many links are likely to be important nodes, such as a menu for branching to other nodes. Such nodes often serve as landmarks and should be displayed more prominently than regular nodes.

Therefore, in the nut view, depending on the number of links related to the node, its size and initial z
The coordinate value (bias) is determined. Specifically, it is determined as follows using logarithm.

(1) Let n _out be the number of links (outgoing links) starting from the node. (2) Similarly, if the number of links (incoming links) with that node as the end point is n _in , (3) the radius of the node is R = R ₀ log (n _out + n
_in +1). (4) The bias of the node is z _b = z _b0 log
(N _out ).

Since the total number of links related to a node has a meaning of how many nodes the hypermedia will appear in, it is considered to be the presence of the node and corresponds to the radius of the node. Let

The number of links coming out from a node is considered to be higher in hierarchy than other nodes like a menu, so it is decided to correspond to the bias.

Of course, when the access to the network is repeated, the number of nodes increases and the number of related nodes also changes accordingly. Therefore, the bias and the size of the nodes may gradually change (increase). is there.

The color of a normal node is yellow, but when it is selected in order to pay attention to the node, it becomes orange. Along with that, the color of links going out of the node is usually changed from white to green.

It should be noted that when drawing links, many links do not overlap each other and the background behind them is invisible.
It uses a semi-transparent color. As a result, the presence of unfocused links was greatly diminished, and the nodes and links that were focused on could be highlighted.

Furthermore, for a node that has already been accessed and the state of the link from that node has been found, the color of the node is darker than that of the node that has not been accessed yet, and the user is still not sure where to look. I made it possible to know if it is a non-existent node.

Above the node (z coordinate), the HT of the node
The plate with the ML title is displayed. The letters written on the plate are 8 to avoid cluttering the screen display.
Although it is a character, the title is displayed in full for the node of interest (selected). Also, normally, the plate is dark blue and the characters are white, but when selected, the plate changes to maroon.

Operations such as enlargement and rotation of the entire coordinate system can be performed from the control panel (FIG. 4).

The control panel is a program different from Nattoview and is implemented using Tcl / Tk. This is, so to speak, an interface to the natto view control FIFO, and the natto view can be controlled by sending command data to the natto view control FIFO.

The three-dimensional operation may be performed by selecting the move / rotate menu in the system menu (FIG. 5), or by directly typing the keyboard according to the guide on the control panel. Perhaps the keyboard interface is the most practical in terms of responsiveness.

The bottom line of the Natto view is a status line, and normally, information (URL and title) on the currently selected (attention) node is displayed. This line displays a message from the system and informs the user of the status of access to the network.

In addition, the control panel and the system menu can be used to select functions such as ending the program, returning to the default viewpoint position, and wire frame mode.

If a high-speed machine is not used, the wireframe mode (FIG. 6) will improve the speed to some extent.

In order to handle information that is a node in the natto view, it is necessary to select a node of interest. This section describes the implementation and operation method related to node selection.

To select a node, aim at the node you want to select and click. Then, a list of all the leads that are over the mouse cursor is displayed as a pop-up menu (FIG. 7) (it is possible to see information about links by selecting the Links submenu here).

When a node is selected, the node is displayed in orange and the links from the node are green. Also, the selected nodes are from top to bottom (in the z-axis direction)
A blue line is also displayed as if it were skewered.

This blue line is a "cursor" for pointing to the location of the selected node even when it disappears because it is hidden behind another node. Also, various operations can be performed from the pop-up menu (FIG. 8) displayed by clicking this cursor.

After selecting the node, browse (B) in the cursor menu (Fig. 8) or control panel (Fig. 4).
Netscape (Ne
tscape) can be controlled to browse the substance of the node as hypertext (HTML).

By using this function, the user can browse directly from the Natto view without following a link, and can immediately refer to the node being accessed by a group member.

At present, this is implemented using a function dedicated to Netscape.

Ideally, the "cursor" would not have been used in Natto View. Nodes should be able to be dragged and moved, or double-clicked to control the browser.

However, there are various problems in reality.

(1) In a normal computer which is not a dedicated graphic workstation, there is a "some" (this is a fairly conservative expression) from the detection of a mouse click to the determination of the selected object. There will be a time lag. Then, the user releases the mouse during this time (it is not a drag). (2) It becomes difficult to operate a node hidden behind many nodes. (3) If the node you want to select is hidden behind another node, you can rotate the coordinate system, but it is not practical considering the speed of calculation and drawing. Became the biggest factor and forced us to adopt a cursor.

When a node that has never been selected is selected, the network is accessed through the link manager (SCAN command), and the 1
Document information such as title and link information are acquired for the node at the tip. Then, the new (x, y, z) coordinates of the node are determined by the algorithm described above.

That is, each time a new node is selected,
The next node will be displayed. By repeating the selection in this way, the number of nodes increases and the map space expands.

Next, by raising or lowering the node, which is the maximum selling point of the “natto view”, the information that the user is paying attention to is examined in detail, and the function of being able to overlook the hypermedia from various viewpoints will be described.

After selecting a node, in order to raise or lower the node, lift up / down can be selected from the control panel (FIG. 4) or the popup (FIG. 8) when the cursor is selected. The currently selected node is raised (lowered) by one step.

When the number of steps is further increased to two steps and three steps, a node closer to the node (that is, a node can easily follow the link from the node) due to the link connection is pulled (FIG. 9). 1
1).

With this function, the user can see the entangled yarn near the node of interest as if it were unraveled. Also, instead of using a tailored hierarchical structure, it is possible to view various aspects of the hypermedia in a multi-faceted manner by using various nodes as the top level hierarchy.

When the Natto View is used in cooperation with the collaboration river, the positions of the group members are displayed on the map (FIG. 10). Nodes being accessed by group members are displayed in green, and their email address is added to the title.

With this function, it is possible to immediately know who is accessing which information, and it can be useful for exchanging evaluation information. Further, if the function of calling the browser is used, it is possible to browse the node (document) immediately.

In the implementation, this is realized as a general-purpose "mark" function, not as a special function for displaying the member name. Therefore, if there is a node that you want to temporarily store, you can mark it.

When the icon button at the top right of FIG. 12 is clicked, a window as shown in FIG. 13 is popped up. Here, from top to bottom, name, email address,
Comments on the content you want to search, phon (phon
e) Enter the address (address including the currently logged-in host name) and press the Ok button at the bottom left of the window.

Next, for the server user manager,
This is done by setting the attribute value of the user with a set command. Further, when the setting of an important attribute value is changed, the notification service of the user manager immediately notifies all users.

When the registration is completed in this way, these data can be referred to and communicated by the server in response to requests from other members (clients).

In the collaboration river, the URL currently referred to by the user is displayed next to the user's name. This feature allows the user to see where the members of the group are currently.

This display is performed in real time as the members of the group move, and the notification service of the user manager is used to realize it.

The collaboration river can execute electronic mail and horn communication, which are existing communication tools for the Internet.

On the member list window, the ordinary e-mail service provided on the Internet can be used. If you click the second and third icon buttons from the right in FIG. You can read or email someone.

In addition to the above electronic mail service, a horn can be used. This is for the case where the user wants to immediately notify by real-time communication, such as when he / she finds information that seems to be very influential for a certain person during the collaborative search.

When actually using, the display portion (NAME: column in FIG. 12) of the name of the other party who wants to put the horn 12
Press the middle button of the mouse on the top (drag), take it from the right side of Fig. 12 to the fourth icon button and release the button (drop). The horn command will be executed automatically after that. To be done.

In the implementation, the horn address is obtained by giving the one registered in the user manager.

In addition to the electronic mail service and the horn, the asynchronous and synchronous integrated information exchange unique to the present system can be performed. The usage is as follows.

First, to send a message to another member, drag the display part of the name of the other party and
Bring up to the fifth icon button 13 from the right of 2 and drop it, and a window as shown in FIG. 14 will pop up. All you have to do is enter a message and press the Post button at the bottom left of the window.

The message sending function is realized by the post command of the user manager. The message is notified immediately if the other party is logged on (user manager notification service), but is temporarily stored in the user manager server if the other party is not logged on.

If you want to see the message sent to yourself, click the sixth icon button 14 from the right in FIG. 12, a window as shown in FIG. 15 appears, and who sent what information. You can look up.

If a message is sent while I am logged on, the message will be received immediately through the notification service of the user manager. The message while the user is not logged on is acquired by requesting the RECV command to the user manager server.

In the collaboration river, the member information is 1
Display attribute data for each person in a horizontal row, name from left to right,
Comments about what you want to search for, the sites you most recently visited, your position in the group, and the importance of the member to you.

Of these, the last scale showing the degree of importance to oneself (scale: column in FIG. 12) is one of the greatest features of this application, and serves as a criterion for determining the priority order when displaying members. Is. Specifically, it is as follows.

(1) Frequent communication with oneself. (2) He provided me with powerful information. (3) High position in the group. When the conditions such as the above are satisfied, the scale value is increased (moved to the right in the figure), and persons with a high value are displayed in order from the top. That is, the display order of this member list constantly changes with the passage of time.

[0138]

[Second Embodiment] The server system is implemented by a UNIX (UNI)
X) mainly in the Perl language. In addition, an overall image of the system is shown in FIG. As is clear from this figure, the three servers that were implemented as servers were the “map system server”, the “cooperative search system server”, and the “modification of the proxy server”. In the following, "Mapping system server" and "Cooperative search system server" are named by their respective link managers (LinkManag).
er), user manager (UserManager)
r). The proxy server uses an existing delegate.

The outline of each role is shown below.

(1) Link manager (mapping system server): In response to a request from the natto view client, the WWW is accessed to scan the link information near the target node, and the result is returned to the client. This scan result is saved as a cache for a certain period of time, reducing access to the network,
We are trying to shorten the response time. (2) User manager (cooperative search system server): manages information about members of groups registered in the cooperative search environment. Specifically, W of each member
It manages the access status to the WW, whether or not the cooperative search is logged on, the message exchange system is maintained, and the attribute values that can be set for each member are managed. (3) Proxy by Delegate: Used to notify the user manager of access information from the WWW browser. Thanks to this delegate, you can manage information about user manager collaborative searches.

The link manager and the user manager have a lot in common regarding the mounting method and the format of the protocol, because the initial plan was to provide these two services by the same server.

Later, due to technical problems, these 2
Although it was concluded that the two functions should be separated, the format of the protocol is the same for clarity.

The request (inquiry) to the server and the response from the server use the TCP protocol (connection type) in TCP / IP, and are line-oriented and text-based similar to SMTP of e-mail and HTTP of WWW. Adopts a protocol. By doing this,
It is easy to set up with language pearl suitable for text processing, and debugging and operation test are also easy.

A request (inquiry) to the server is normally made by the following 1
Takes line format. command arg1 [arg2 ...] At this time, the command is a character string indicating the type of the inquiry instruction, and arg1 (its argument) is its argument. However, the arguments cannot contain whitespace except the last one (except for the last argument, which can include whitespace other than line breaks, depending on the instruction).

The response or notification from the server usually takes the following one-line format. "Status type version resp
once-reason "

However, exceptionally, the result may be plural lines.

The respective meanings are shown below.

Status: A three-digit integer indicating whether the processing for the inquiry was successful. The meaning of the status value is shown in Table 1 together with the reason field.

[0149]

[Table 1]

Type: any one of Q, N, and L characters representing the type of server and service, shown in Table 2.

[0151]

[Table 2]

Version: Protocol version of server response Response: Response (notification) data Reason: Additional explanation (mainly used for debugging) In this section, user manager that manages user information and supports collaborative search system. Will be described.

The user manager manages all kinds of data related to the collaborative search system. Its roles are roughly divided into management of a participant database and acquisition and management of WWW access information.

The user manager responds to various inquiries from the collaboration river client as a response.
Returns information about it. This is the inquiry service. In addition, the user manager must notify other members of important changes to the database by members of the group. That is the notification service. Below is a summary of these two services.

(1) Inquiry service: Processes an inquiry about a database managed by the user manager. The client connects to this inquiry service and requests processing when it wants to obtain and set the necessary information about the database. This connection does not have to be left on all the time, rather it is better to reconnect with each query. (2) Notification Service: When significant changes are made to other members of the group or system databases,
Notify all clients currently connected of the change. The client must connect to the notification service at the beginning of the session for collaborative search. Once connected, the server will be logged on thereafter, and information necessary for cooperative search will be sent one after another from the server until the cooperative search is logged off and the connection is disconnected.

When a connection request is made to the user manager, the user manager uses the Internet identification protocol (Identifier).
user (RFC1413) to identify the user of the connection partner. Then, the alias database is searched to obtain the registered mail address of the user for the user manager. After that, this mail address is used as a user name (user identifier).

The user manager manages information about the users of the group for cooperative search as attribute values. The attribute value can be set and obtained by a set instruction and a get instruction, respectively.

As the ID for identifying the user, the mail address of the user obtained from the identification protocol and the alias database is used. The user name public can be used as a user name for storing data that can be read and written by anyone.

The attribute values include the following.

# LOGON #: The time when the user has recently connected to the notification service. # LOGOFF #: The time when the user recently disconnected the notification service. ! NAME! : User's name. The user name displayed in the list of clients (collaboration river). ! URL! : A page that the user has recently viewed with a WWW browser. ! status! : User status. Displayed on the client. Horn: The phone address of the user. * NRECV *: The number of messages sent from other members and stored in the server (user manager) for itself. The system handling regarding attribute values can be classified as follows. * Name *: Set by the system and cannot be rewritten by the user. # Name #: Set by the system, cannot be rewritten by the user, and when there is a change, the notification service is used to notify all clients of it. ! name! : It can be set freely by the user (the system can be rewritten), and when there is a change, it is notified to all clients. Name: Normal attribute. What the user can set freely.

There is no limit to the number of attribute values, and the client can prepare as many attribute values as necessary. The implementation of the attribute value is realized by the hash table used by GDBM.

In order to manage the access destination of the member of the group by the WWW browser, it is necessary to acquire the access information of the user's WWW on the server side and prepare it so that the member of another group can use it.

Therefore, in this system, each user (W
WW browser is a dedicated proxy server (proxy se)
access to WWW information via an intermediary of rber). Delegated as a proxy server (del
The user manager can know the access information of the user by using the egated) and passing the log directly to the user manager by the FIFO.

On the user manager side, this log information is analyzed and each user's! URL! Set a value for the attribute and notify all clients through the notification service.

[0165]

According to the present invention, two nodes, which are information units, are used.
Since the two-dimensional planes are lifted and expanded in the three-dimensional space by arranging them in the three-dimensional plane and linking them respectively, it is possible to observe the specific node in detail.

Since the specific node in the three-dimensional space can be rotated or partially enlarged for observation, the necessary information can be accurately and easily retrieved from a large amount of complicated information.

[Brief description of drawings]

FIG. 1 is an overall hardware view of an implementation system of the present invention.

[Fig. 2] A view of the view browser of Natto.

[FIG. 3] A view of the view of natto seen from directly above.

FIG. 4 is a view of the view control panel of natto.

FIG. 5 is a view of the system menu of Natto view.

FIG. 6 is a diagram of a wire frame mode in the natto view.

FIG. 7 is a diagram of a popup for selecting a node in the natto view.

FIG. 8 is a diagram of a popup when the cursor in the natto view is also selected.

FIG. 9 is a diagram showing a multi-step pulling operation of the natto view in the same manner.

FIG. 10 is a diagram showing a display of fellows in the natto view group.

FIG. 11 is a view in which the nut view is also lifted and rotated to a position that is easy to see.

FIG. 12 is a diagram of the collaboration river and the collaboration river communication window.

FIG. 13 is a diagram of user registration for the collaboration river.

FIG. 14 is a diagram of sending a message of the collaboration river.

FIG. 15 is a diagram of receiving a message from the collaboration river.

FIG. 16 is a diagram of the system architecture.

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Kenichi Okada 4-25-12 Hongo, Bunkyo-ku, Tokyo (72) Inventor Takahiro Soma 3-14-1, Hiyoshi, Kohoku-ku, Yokohama-shi, Kanagawa Keio University Faculty of Science and Engineering Measurement Department of Engineering Matsushita Lab (72) Inventor Haruhiko Nishiyama 3-14-1, Hiyoshi, Kohoku Ward, Yokohama City, Kanagawa Prefecture Keio University Faculty of Science and Engineering Matsushita Lab (72) Inventor Hidekazu Shiozawa Hiyoshi Hikita, Kanagawa Prefecture 3-14-1 Keio University Faculty of Science and Engineering Department of Measurement Engineering Matsushita Laboratory

Claims (3)

[Claims] 1. A hypermedia visual system characterized in that nodes, which are information units, are arranged in a two-dimensional plane and linked to each other, and then a specific node among the nodes in the two-dimensional plane is lifted to form a three-dimensional model. System. 2. A node, which is an information unit, is arranged in a two-dimensional plane and linked to each other, and then a specific node among the nodes in the two-dimensional plane is lifted to make a three-dimensional model, and the lifted node is lowered to another. A hypermedia visualization system that enables you to take a bird's eye view of information structures from different viewpoints by interactively lifting nodes. 3. A node, which is an information unit, is arranged in a two-dimensional plane and linked to each other, and then a specific node selected from the nodes in the two-dimensional plane is lifted to make a three-dimensional model, and this three-dimensional model is rotated. Or, a hypermedia visualization system characterized by partially enlarging and overlooking the specific node for easy viewing.