JP4539151B2 - Tree structure differential data creation device - Google Patents

Description

  The present invention relates to a device for expressing a relationship between a group of documents having a hypertext structure in a tree structure, and more particularly to creation of difference data indicating a change in the tree structure by updating the hypertext structure.

  A document group on the WWW (World Wide Web) has a hypertext structure (also called a hyperlink structure) in which a document is a node and a link between documents is an arc. This hypertext structure has the advantage that any document can be easily linked by a link. However, since it is basically a network structure, when the number of documents increases and the relationship between documents becomes complicated, humans can create documents. It becomes difficult to grasp the relationship between them.

  On the other hand, a method of displaying the hypertext structure in a simplified tree structure is conventionally used. To convert a hypertext structure to a tree structure, follow the links described in the document in order from the first document, and if a linked document is found, incorporate it into the tree structure as a child node of the link source document. Can be repeated.

  One such tree structure display is disclosed in Patent Document 1. In this method, when a link is extended from a plurality of nodes to one node (assumed to be “A”) in the hypertext structure, when converting this to a tree structure, Select one of them, place node A as a child of the selected node, and other nodes linked to node A include a reference node (also referred to as a shortcut node) indicating node A. Place as a child. In this process, when following a hypertext structure link, the link source node when the node A is first reached is the parent of the real node A on the tree structure. The reference node is displayed differently from the real nodes constituting the tree structure. These reference nodes point to the same document as the real node A. However, since the relationship on the tree structure is representatively shown by the real node A, the reference node has no child nodes.

  In addition, “GoLive”, a website management tool of Adobe Systems, Inc., when specifying a starting document on the WWW, examines the hypertext structure by following links in order from that document, and 1 in the hypertext structure. When a document is specified, it has a function of creating and displaying a tree structure having the document as a root node from the hypertext structure.

  On the WWW, web pages are frequently generated, deleted, and updated. As a result, the hypertext structure formed by the web page group changes, and it is necessary to update the tree structure information corresponding thereto. Such updating of the tree structure may be performed by following the links again and visiting the web page group.

  By the way, there is a case where it is desired to share information of the tree structure created / updated in this way by a plurality of devices. In such a case, information on the tree structure created and updated by a certain device may be transmitted to another device. However, when the number of documents constituting the tree structure is large, the amount of data of the tree structure to be transmitted is enormous. Become a thing. For such a problem, Patent Document 2 discloses a system in which a transmission-side apparatus creates difference information between a tree structure created last time and a tree structure created this time and transmits the difference information to the reception-side apparatus. .

  The method of Patent Document 2 is effective in reducing the amount of transmission data, but has the following problems. That is, in the method of Patent Document 2, the transmission side apparatus performs a process of extracting the difference between the tree structure created last time and the tree structure newly created this time, in order to create difference information. For this process, first, there is a problem that a huge amount of data indicating the entire tree structure created last time must be held. Second, in this process, the current tree structure must be traced twice, that is, when the current tree structure is created and when the created tree structure is compared with the previous tree structure. Processing load is heavy.

JP-A-6-35657 JP 2001-67260 A

  An object of the present invention is to provide an apparatus for efficiently creating a difference in tree structure data representing a hypertext structure in accordance with a change in the hypertext structure formed by a document group.

Tree apparatus according to the present invention, in response to changes in the hypertext structure documents on the network is configured, in order to update the tree structure data representing the hypertext structure, corresponding to a change in the hypertext structure An apparatus for creating tree structure difference data indicating a change in structure data, in which information indicating whether each document exists on the hypertext structure at a time before the change in the hypertext structure is registered A pre-update document table, a pre-update link table in which links between documents existing on the hypertext structure at the time before the change of the hypertext structure are registered, the pre-update document table, and the pre-update link table By examining the document group after the change while referring to the The change of the tree structure data is shown by detecting the added and deleted documents and links and tracing the inter-document links of the hypertext structure after the change in order from the predetermined first document while referring to the detection result. A difference processing unit for creating tree structure difference data, wherein the difference processing unit collects document groups registered in the pre-update document table and link destination document groups of the respective documents from the network. , based on whether they each document is acquired, as they each document, those present and document content from before the change is not updated, the present and document content from before the change is updated , A first process for creating a post-collection document table that indicates which one of the one added by the change and the one deleted by the change; The links between documents exist from before the change based on the comparison between the links between documents included in each document collected from the document and the links between documents registered in the pre-update link table. , A second process of creating a post-collection link table that indicates which one of the one added by the change and the one deleted by the change, and after the change in order from the top document When tracing an inter-document link having a hypertext structure, the state of the inter-document link and the state of the link destination indicated by the inter-document link are retrieved from the post-collection document table and the post-collection link table, and obtained by retrieval. was based on a combination of the states of the state and the linked document of the document links, (a) the link source of the statements listed in the tree structure data of link between the document (B) Corresponding to the link destination document indicated by the inter-document link in the tree structure data, in which the node corresponding to the link destination document indicated by the inter-document link is added as a child of the node corresponding to A change that a node is deleted from a child of a node corresponding to the link source document indicated by the inter-document link, and (c) a node corresponding to the link destination document indicated by the inter-document link in the tree structure data When it is determined that a change corresponding to any one of the change contents in which the contents of the linked document are updated is added to the tree structure data, the data indicating the change contents is changed to tree structure difference data. And the difference processing unit includes a combination of the state of the inter-document link and the state of the link-destination document, A change rule table indicating whether or not the tree structure data corresponding to the combination is changed and the change contents when there is a change, and the change contents include (a), (b), and (c ) And is obtained by a search from the post-collection document table and the post-collection link table when tracing the inter-document link of the hypertext structure after the change in the third process. If the changes corresponding to the combination of the states of the state and the linked document links between the documents is determined from the change rule table, the changes corresponding to the changes that were required the tree structure data It is determined that it has been added .

  In a preferred aspect of the present invention, the difference processing unit collects the document group registered in the pre-update document table and the link destination document group of each document from the network, and can acquire each document. A post-collection document table indicating whether each of these documents exists before the change, has been added by the change, or has been deleted by the change. Based on a comparison between a link group included in each document created and collected from the network and a link group registered in the pre-update link table, each link exists before the change, the change A post-collection link table is created to indicate which one is added by (1) or deleted by the change, and the change is made in order from the first document. When following the inter-document link of the hypertext structure, the inter-document link and the link destination document indicated by the inter-document link are searched from the post-collection document table and the post-collection link table and obtained by the search. It is determined whether or not the tree structure data is changed with respect to the linked document based on the combination of the link state between the documents and the linked document state. Data indicating the contents is added to the tree structure difference data. -

  In a further preferred aspect, the difference processing unit changes the combination of the inter-document link state and the link-destination document state, whether or not the tree structure data corresponding to the combination is changed, and when there is a change. And a change rule table indicating the contents, and the change rule table is searched when tracing the inter-document link of the hypertext structure after the change.

  The best mode for carrying out the present invention (hereinafter referred to as “embodiment”) will be described below with reference to the drawings.

  FIG. 1 is a diagram showing an example of the overall configuration of a system to which the present invention is applied.

  In this system, each company A, B, and C has in-house networks 14a, 14b, and 14c, and a large number of computers such as personal computers and various servers are connected to the in-house networks 14a, 14b, and 14c. Yes. The in-house networks 14a, 14b, and 14c are respectively provided with firewalls so that unauthorized access to the in-house network from devices on the Internet 10 is prevented. Many documents exist as web pages on the in-house networks 14a, 14b and 14c. These web pages are linked together by a link description to form a hypertext structure. A user in each company can browse a document existing on the company's in-house network 14a, 14b or 14c together with a document published by a server on the Internet 10 using a web browser on a personal computer.

  Here, since the hypertext structure formed by the document group is difficult for humans to understand, if this can be converted into a tree structure easy to understand for humans and presented to the user, the user can easily find a desired document. The tree structure information providing server 12 connected to the Internet 10 is a server that provides such tree structure information. The tree structure information providing server 12 is an ASP (Application Service Provider) server, for example. In this case, companies A, B and C are ASP customers. The tree structure information providing server 12 provides the company A, B, and C with tree structure data that represents the link relationship of the document groups in each of the in-house networks 14a, 14b, and 14c in a tree structure.

  Here, in the in-house networks 14a, 14b and 14c, documents are newly created, deleted, and contents are changed as needed, and the link structure (hypertext structure) formed by the document group changes accordingly. In order to provide the latest tree structure data as much as possible to the users in the client company in such a situation, the tree structure information server 12 responds to the change of the document group in each in-house network 14a, 14b and 14c. It is necessary to update the tree structure data held by the server 12 itself.

  However, since the in-house networks 14a, 14b and 14c are generally protected by firewalls, the tree structure information providing server 12 cannot acquire documents (web pages) on the in-house networks 14a, 14b and 14c.

  Therefore, in this embodiment, the tree structure information collection device 16 is provided in the firewall of each company's internal networks 14a, 14b and 14c. The tree structure information collecting device 16 investigates a document group existing in the company's in-house network 14a, 14b or 14c, and creates difference data indicating how the tree structure constituted by the document group has changed since the previous survey. And transmitted to the tree structure information providing server 12. Thereby, the tree structure information providing server 12 can obtain difference data regarding the tree structure formed by the document group in the firewall.

  FIG. 2 is a diagram illustrating an example of functional module configurations of the tree structure information providing server 12 and the tree structure information collection device 16.

  The tree structure information providing server 12 includes a storage device 122, a client request processing unit 126, and a tree structure updating unit 128. The storage device 122 stores tree structure data 124 indicating a tree structure formed by a document group in the in-house networks 14a, 14b and 14c of each customer company.

  The client request processing unit 126 receives a tree structure request 180 from a client device 18 (typically a user's PC in each customer company) connected on the Internet 10 and receives tree structure data 182 corresponding to the request 180. Created and provided to the client device 18. Here, the tree structure data 124 in the storage device 122 includes a tree structure group constituted by a document group on an in-house network of all customer companies and a document group disclosed on the Internet. The client request processing unit 126 selects, from such large-scale tree structure data 124, the tree structure browsing range designated by the user who issued the tree structure request 180, and the tree structure belonging to the range of the user's browsing authority. The data is extracted and transmitted as tree structure data 182 to the client device 18.

  The tree structure update unit 128 is a functional module for updating the tree structure data 124 in the storage device 122. The tree structure update unit 128 converts the tree structure data 124 currently stored in the storage device 122 into the latest hypertext structure based on the tree structure difference data 220 sent from the tree structure information collection device 16 of each customer company. Update to reflect.

  The tree structure information collection device 16 provided in each customer company has a storage device 200, a collection unit 162, and a difference data creation unit 164.

  The storage device 200 stores a circulation list 202, a previous document table 204, a previous link table 206, and a difference information determination table 208. The circulation list 202 is a list showing documents to be acquired for the update when the update timing of the tree structure comes. The previous document table 204 is a table indicating the presence / absence state of each document on the in-house network obtained by the collection process at the previous tree structure update timing. The previous link table 206 is a table showing the existence / non-existence state of links between documents on the in-house network obtained by the collection processing at the previous tree structure update timing. The difference information determination table 208 is a table showing a rule for obtaining a tree structure difference from a combination of document and link state changes before and after the update timing.

  Further, a post-collection document table 210 and a post-collection link table 212 are created on the storage device 200 in accordance with the collection of the document and link information by the collection unit 162. The post-collection document table 210 is a table indicating the state of each document that has been found as a result of the collection process of the collection unit 162. Further, the post-collection link table 212 is a table showing the state of each link that has been found as a result of the collection process of the collection unit 162. The detailed contents of these lists and tables and how to use them will be described later with specific examples.

  The collection unit 162 is a functional module that collects a document group existing on the in-house network 14 provided with the tree structure information collection device 16 and links described in the document. By this collection, a post-collection document table 210 and a post-collection link table 212 are created.

  The difference data creation unit 164 uses the post-collection document table 210 and post-collection link table 212, which are the results of collection by the collection unit 162, and other information stored in the storage device 200, and uses the tree structure difference data 220. Is a functional module that creates

  Details of the processing performed by the collection unit 162 and the difference data creation unit 164 will be described later with specific examples.

  The tree structure information collection device 16 operates the collection unit 162 and the difference data creation unit 164 to create the tree structure difference data 220 every time the tree structure update timing determined by a predetermined rule arrives, and the tree structure information providing server 12 to send.

  Next, the operation of the system of this embodiment will be described with a specific example.

  FIG. 3A is a diagram showing a hypertext structure formed by a group of documents in a certain in-house network at the previous tree structure update timing. FIG. 3B is a diagram showing a hypertext structure formed by a group of documents in the same in-house network at the current tree structure update timing. 3A and 3B, nodes 300 indicated by circles indicate individual documents constituting the hypertext structure, and the numbers indicated in the circles indicate document IDs assigned to the documents. An arc 302 indicated by an arrow indicates a link between the documents. The node at the base of the arrow indicates the link source document, and the node at the tip (arrowhead side) of the arrow indicates the link destination document.

  In FIG. 3A, the frame line of the node with the document ID “7” is indicated by a dotted line, which means that the actual data of the document corresponding to the node does not exist on the in-house network. That is, in the document with the document ID “3”, although the link to the document with the document ID “7” is described, the document with the document ID “7” does not exist. In the following description, in order to simplify the description, “document n” is simply described instead of “document ID“ n ”(n is a natural number)”.

  Comparing FIG. 3A and FIG. 3B, it can be seen that there is the following change in the hypertext structure between the previous update timing and the current update timing.

(1) The link from document 2 to document 4 is deleted (2) The link from document 4 to document 1 is newly generated (3) The content (content) of document 5 is updated (4) Document 6 was deleted (the link from document 6 to document 8 was also deleted)
(5) Document 7 is newly generated (6) A link from document 7 to document 8 is newly generated.

  Roughly speaking, information indicating such a change may be transmitted as the tree structure difference data 220 from the tree structure information collecting apparatus 16 to the tree structure information providing server 12.

  3A and 3B are very simple examples created for the purpose of explanation, so the difference data can be easily obtained. However, in an actual in-house network, thousands, tens of thousands, or more documents have a hypertext structure. It is not easy to obtain tree-structured difference data corresponding to the structure formed by such an enormous document group. Therefore, the tree structure information collection apparatus 16 of this embodiment efficiently creates tree structure difference data by the following processing.

  FIG. 4 is a diagram showing the flow of the tree structure data update process in the system of this embodiment.

  In this process, when the update timing of the tree structure data arrives, the collection unit 162 of the tree structure information collection apparatus 16 provided in the in-house network first executes the document group acquisition process with reference to the circulation list 202. Then, the collection unit 162 compares the result of the acquisition process with the previous document table 204 and the previous link table 206, thereby indicating the post-collection document table 210 indicating the state change of each document and each link before and after the update timing. Then, the post-collection link table 212 is created (S10).

  FIG. 5 is a diagram illustrating an example of data contents of the circulation list 202. As shown in FIG. 5, the circulation list 202 includes a document ID assigned to each document and a URL of the document for each document constituting the hypertext structure in the in-house network where the tree structure information collection device 16 is installed. (Uniform Resource Locator) is registered. The document ID may be unique on the in-house network, but of course may be unique in the entire system including a plurality of in-house networks. In the circulation list 202, information on documents that have been confirmed to exist by processing at the previous tree structure update timing (ie, documents that can be acquired) among the documents constituting the hypertext structure in the in-house network. It is remembered. Note that the example of FIG. 5 corresponds to the hypertext structure shown in FIG. 3A, so there is no entry corresponding to the document with the document ID “7”. However, this is merely an example, and instead of a document that could not be acquired, if a link pointing to the document is described in another document, it can be registered in the circulation list 202.

  In S <b> 10, the collection unit 162 first tries to acquire a document indicated by the URL of each document registered in the circulation list 202 using an HTTP (Hypertext Transfer Protocol) GET method. If the content of the document acquired by this acquisition processing includes a link description indicating a URL that is not included in the circulation list 202, the collection unit 162 assigns a new document ID to the document indicated by the link description. It registers in the circulation list 202 and tries to acquire the document. In response to this acquisition request, if the server indicated by the URL of the acquisition target document has the document indicated by the requested URL, the server transmits the document to the tree structure information collection device 16 and has the document indicated by the URL. If not, data indicating that the document does not exist is transmitted to the tree structure information collecting device 16. In this acquisition process, for example, by using an If-modified-since header, the entity data of the document is sent to the tree structure information collection device 16 only when the document to be acquired is updated. You can also In this case, when the document to be acquired has not been updated, data indicating that there is no update is returned to the tree structure information collection device 16. With such an acquisition process, it can be determined whether or not each document to be acquired exists. If the If-modified-since header is used, it can be determined whether or not the content of the existing document has been updated since the previous acquisition. Note that if only the difference between the tree structures is obtained, it is not necessary to check whether or not the contents of the document have been updated. However, if the If-modified-since header is used, the tree structure information collection device 16 has already acquired the difference. There is an advantage that it is possible to avoid obtaining the document again.

  The collection unit 162 compares the obtained result with the previous document table 204 to determine how the state of each document has changed from the previous update timing, and creates a post-collection document table 210 indicating the state change. To do.

  FIG. 6A is a diagram showing an example of data contents of the previous document table 204 corresponding to the hypertext structure of FIG. 3A. In this figure, the document ID and the state of the document are registered for each document. The “state” of the registered document is “present” or “non-existent”. That is, the previous document table 204 registers whether or not a document corresponding to each document ID exists on the network at the time of the previous tree structure update timing. Note that the example of FIG. 6A follows a method in which a “non-existent” document is not registered in the previous document table 204, and thus an entry with a document ID “7” does not exist. However, this is only an example, and instead of this, a method of registering in the previous document table 204 can be considered if the document is shown in the link description of any document. In this case, the document ID “7” is registered as “non-existing” in the previous document table 204.

  FIG. 6B is a diagram illustrating an example of data contents of the post-collection document table 210 created by the collection unit 162 when the hypertext structure of FIG. 3A is changed to the structure of FIG. 3B. The post-collection document table 210 shows changes in the state of each document before and after collection. The state of each document before collection is shown in the previous document table 204, and the state of the document after collection is obtained by document acquisition processing by the collection unit 162. Changes in the document status before and after collection include (1) “New” (documents that were not collected before are newly detected at the time of collection), (2) “Delete” (documents that existed before collection are present at the time of collection) (3) “Update” (it was found at the time of collection that the contents of the document that existed before the collection were updated). Moreover, although it cannot be said to be “change” in the literal sense, (4) “existence” (contents existing before collection are not updated as “change” in a broad sense indicating the relationship between the states of the documents before and after collection. And (5) “Non-existent” (documents that were not present before collection were not present at the time of collection). A document that does not exist before collection and does not exist at the time of collection (“non-existence”) does not appear on the tree structure, and thus does not need to be registered in the table 204. Further, when only the change in the tree structure is desired, “update” may be handled in the same manner as “existence”.

  The collection unit 162 matches the previous document table 202 and the current acquisition processing result to obtain a change in the state of each document attempted to be acquired, and registers it in the post-collection document table 210. In the example of FIG. 6B, for example, the document 5 is “update”, the document 6 is “deleted”, and the document 7 is “new”.

  The collection unit 162 completes the post-collection document table 210 by repeating the above processing until all URLs registered in the circulation list 202 and acquisition results of all URLs newly detected in this processing are found. Let

  In S10, the collection unit 162 detects link state changes before and after the current update timing. The state of the link group before the update timing arrives is shown in the previous link table 206.

  FIG. 7A is a diagram showing an example of the data contents of the previous link table 206. As shown in this figure, an entry indicating one link in the link table 206 is composed of four items: a link source ID, a link destination ID, a state, and a shortest path flag. The link source ID is the document ID of the link source document of the link (that is, the document including the description of the link). The link destination ID is the document ID of the link destination document of the link (that is, the document pointed to by the link).

  The state is information indicating whether or not the link exists. If there is a link description in any of the document groups acquired at the previous update timing, the link state indicated by the link description is “existing”. In the example of FIG. 7A, since a method of not registering links that do not exist in the table is used, the state of all links is “present”.

  The shortest path flag is a flag indicating whether or not the link is the shortest path that reaches the link destination document when the tree structure is created. In this embodiment, in order to create differential data of a tree structure, a hypertext structure at the update timing is searched according to a link search rule at the time of tree structure creation. In general, a breadth-first search rule is known as one of search rules for creating a tree structure. Here, the creation process of the tree structure by the breadth-first search in the present embodiment will be described. In this process, in order to create a tree structure, the information of each document shown in the post-collection document table 210 and the link information between the documents shown in the post-collection link table 212 are examined.

  In order to perform a breadth-first search of a hypertext structure, one queue for storing links and a search flag corresponding to each document are prepared. At the start of creating a tree structure, the search flags of all documents are reset, and the search is started from a predetermined top document (for example, the top page of a company Web page). That is, first, a search flag for the first document is set, and one or more links indicated in the first document are added to the queue in a predetermined order (for example, the appearance order of each link in the description of the document). To do. Then, “acquires one link from the queue, searches the linked document, sets a search flag for the document reached by this search (referred to as“ search document ”), and queues the link indicated in the search document. "Add to" until the queue is empty. In the search process, the document that has already been searched once may be reached again. At this time, the search flag is already set, and the state of the search flag remains unchanged. Here, since the post-collection document table 210 and post-collection link table 212 that acquire links in this search also contain documents and links that do not exist on the WWW at this time, such absent documents and links Links should not be built into the tree structure. Therefore, when the state change of the search document shown in the post-collection document table 210 is “deleted” or “non-existing”, the link shown in the search document is not added to the queue. In addition, when the state change of the link taken out from the queue (this is shown in the post-collection link table 212. Details will be described later) is “deleted” or “non-existent”, the search flag for the linked document is not set. The link indicated in the search document is not added to the queue. A series of documents and links traced in the search process shows a tree structure. However, in a tree structure, one node (referred to as “attention node”) cannot be a child of multiple parent nodes, so when generating a tree structure, a certain attention node (document) It is a child of the node of the link source document of the link traced when it arrives (that is, when the search flag for the node of interest is not set). Then, when the node of interest is reached from another node after that (that is, when the search flag for the node of interest is set), the reference node (shortcut node) indicating the node of interest is included in the “other node”. ) As a child. On the other hand, the child node corresponding to the link when the node of interest is first reached is handled as a real node. When the link destination document is a reference node, the search for the link described in the link destination document is stopped (that is, the link indicated in the document is not added to the queue). This is because the search from the link described in the link destination document can be performed by the route of the real node corresponding to the link destination document, and therefore, a duplicate search is avoided. When displaying the tree structure data, the real node and the reference node can be displayed in different display forms.

  The shortest path flag indicates whether or not the document indicated by the “link destination ID” has reached the earliest in such a hypertext structure search. As can be seen from the above description, when the shortest path flag indicates “O”, that is, “is the shortest path”, in the tree structure data, the node of the link destination ID is a child of the node of the link source ID, and Become a real node. On the other hand, when the shortest path flag indicates “x”, that is, “not the shortest path”, in the tree structure data, the node of the link destination ID is a child of the node of the link source ID, but is a reference node Means.

  In the network structure of FIG. 3A, the document 4 can be reached from both the document 2 and the document 5 paths. However, when searching for the document 1 as a starting point, the route passing through the document 2 can reach the document 1 earlier. In the table 7A, the link from documents 2 to 4 is the shortest path, and the link from documents 5 to 4 is not the shortest path.

  Note that the document 7 does not exist in the network structure of FIG. 3A. For a document that does not exist, it is impossible to determine whether the shortest path for each link that reaches the document, so the shortest path flag is set to “unknown”.

  The previous link table 206 indicating the state of the link group before the update timing has been described above. The collection unit 162 associates the table 206 with the link description described in each document collected this time to obtain the state change of each link before and after the collection, and the post-collection link indicating this state change. A table 212 is created.

  There are four types of link status changes: (1) “new”, (2) “deleted”, (3) “existing”, and (4) “non-existing”. The meanings indicated by these four kinds of state changes are the same as in the case of the state changes of the document described above.

  FIG. 7B shows an example of the post-collection link table 212 corresponding to the change in the hypertext structure from FIG. 3A to FIG. 3B. The post-collection link table 212 has an entry for each link registered in the previous link table 206 and each link newly detected by the current collection. One entry includes a pair of a link source ID and a link destination ID that specify a link, a state change of the link, and a shortest path flag. The link state change can be obtained by comparing the previous link table 206 with the presence / absence of each link found by the current collection. For example, the link from the document 2 to the document 4 is “existing” in the previous link table 206 but cannot be detected when searching the hypertext structure in FIG. Further, the link from document 4 to 1 does not exist in the previous link table 206, but is detected when searching the hypertext structure in FIG. 3B, so the state change becomes “new”. The shortest path flag is “unknown” for the link detected as “new” this time, and the shortest path flag of the previous link table 206 for the link corresponding to other state changes (“exist” and “deleted”). Take over the value.

  The collection process (S10) by the collection unit 162 has been described above. When this collection process is completed and the post-collection document table 210 and the post-collection link table 212 are completed, the process by the differential data creation unit 164 is executed (see FIG. 4 again).

  That is, the difference data creation unit 164 creates tree-structured difference data using the collected document table 210, the collected link table 212, and the difference information determination table 208 (S12). In this creation processing, the difference data creation unit 164 traces the document on the hypertext structure by following the links on the hypertext structure in order from a predetermined first document in the order of breadth-first search for creating a tree structure. I will go around. Since the post-collection link table 212 covers all the links existing at the current update timing, the search can be executed by referring to this table 212. Each time the difference data creation unit 164 finds a new link in the width direction priority search (this link is referred to as an attention link), the state change of the attention link and the state change of the link destination document indicated by the attention link Are obtained from the post-collection document table 210 and the post-collection link table 212, respectively. Then, the combination of the link of interest and the state change of the linked document is searched in the determination table 208, and if difference information is obtained by this search, it is added to the tree structure difference data. In S12, the search for the link, the search of the determination table 208, and the addition of difference information based on the search result are repeated until all links in the hypertext structure are traced.

  A specific example of the difference information determination table 208 is shown in FIGS. In these four tables, the contents of the change in the tree structure corresponding to the combination are registered in the cell indicated by the combination of the state change of the target link and the state change of the linked document indicated by the target link. In these tables, a hatched cell indicates that the combination indicated by the cell cannot occur or the tree structure does not change for the combination indicated by the cell.

  These four tables are prepared according to changes in the shortest path state of the link of interest before and after the breadth-first search is started. Here, the shortest path state of each link before the search is started is shown as a “shortest path flag” in the post-collection link table 212. Further, the state of the link after the search can be determined by the search flag of the link destination document. That is, if the search flag of the linked document reached by following the attention link is not set (this indicates that the document has been reached for the first time in the search process), the attention link is the shortest path, and the search flag If is standing, it is not the shortest path. Therefore, when a certain attention link is traced, the state change of the attention link is fixed. In response to this state change, an appropriate table is selected from FIGS.

  FIG. 8A shows that the link of interest is the shortest path before the start of breadth-first search (hereinafter referred to as “past” for simplicity), and the shortest path even when the current breadth-first search is completed (hereinafter referred to as “present” for simplicity). An example of the difference information determination table 208A in a case is shown.

  In this table 208A, when the document state change is “delete” and the link state change is “delete” or “exist”, the difference information of the tree structure is “delete”. This “deletion” of the difference information means that, in the tree structure, the real node indicating the link destination document of the target link is “deleted” from the child of the node indicating the link source document of the link. That is, in this case, the linked document that arrived by following the attention link exists as a child real node of the link source document at the previous update timing, but the document itself is deleted at the current update timing. It turns out that the real node is deleted. In this case, difference information indicating that the link-destination real node whose parent is the link source of the target link has been deleted is added to the tree structure difference data 220. Of the tree structure difference data 220 shown in FIG. 9, the entry corresponding to the last document 6 is an example of the difference information. Here, when the link state change is “existing”, only the attention link on the hypertext remains after updating (however, this is not related to the tree structure).

Also, when the document state change is “update” and the link state change is “exist”, the difference information is “update”. In other words, considering that the link state change is “existing” and the document state change is “update”, the document is updated this time even at the previous update timing, considering that it is the “shortest path” before and after the search. This means that it is a real node that is a child of the link source document at the timing, and the content of the document has been updated. Therefore, the tree structure the difference data 220, in the tree structure, add the difference information indicating that Oh lapis lazuli link destination document a child of the link source document of attention link is content update. The entry corresponding to the first document 5 in the tree structure difference data 220 in FIG. 9 is an example of the difference information. If it is sufficient to know only the change in the tree structure (that is, it is not expressed in the tree structure data 182 to be passed to the client device 18 until the document content is updated), it is assumed that the tree structure does not change for this combination. Good.

  It should be noted that the situation where the link state change is “new” is inconsistent with the fact that the link was the shortest path in the past, and thus cannot occur. Also, if the link status change is “non-existent”, there is no link of interest in the past or present, so the parent-child relationship corresponding to the link of interest on the tree structure does not exist in the past or present. There will be no change between the past and the present. Therefore, in this case, it is not necessary to incorporate information about the target link into the tree structure difference data 220. Even when the state change of a document is “non-existing”, the document itself does not exist in the past or the present, and therefore the document does not exist in the tree structure or in the past or the present. That is, there is no change in the tree structure after all.

  Further, when the state change of both the document and the target link is “existing”, the link destination document of the link is the link source document in both the past and the current in the situation where the target link is the shortest path both in the past and the present. This indicates that this is a real node of the child, so there is no change in the tree structure.

  Also, the situation where the state change of the document is “new” cannot occur because the link contradicts the shortest path in the past.

  When the state change of the target link is “deleted” and the state change of the link destination document is “exist” or “update”, the link destination document is the link source document on the tree structure at the previous update timing. Although it exists as a real node that is a child, the parent-child relationship is lost because the link of interest is deleted at this timing. Therefore, according to one idea, in the difference information table 208A, the difference information corresponding to the two combinations may be “deleted”. However, in this case, if the tree structure information providing server 12 is of a type in which the tree structure difference data 220 is processed one difference information at a time from the top, the difference information “deleted” is used as the tree structure difference data 220. If the tree structure information providing server 12 reads the difference information, it deletes the actual node indicating the link destination document indicated by the attention link. There is a problem like this.

  That is, in this case, the deletion of the real node also deletes the node group of descendants of the real node in the tree structure (this kind of treatment is common because of the tree structure). However, either the real node or its descendant nodes (hereinafter collectively referred to as “descendant nodes”) may later be found to be linked from another document. In such a case, for a descendant node that has been found to be linked from another document, only the parent changes and other attributes do not change. Also, the descendants of the descendant node have no change in the parent or other attributes. In such a case, most simply, the difference information indicating the new parent of the descendant node may be sent to the server 12. For the descendants of the descendant node, there is no need to send difference information because neither the parent nor other attributes have changed.

  However, once the real node indicating the link destination document of the link of interest is deleted, if it turns out that the descendant node is a child of another node, the difference information indicating that the real node is deleted and the descendant node are It becomes necessary to send to the server 12 difference information indicating that it is newly added as a child of another node. In addition, since the descendants of the descendant nodes have been deleted even if there is no change in the direct parent node, it is necessary to send difference information indicating that they are added to “new”. This increases the amount of data compared to the simple difference information described above. When attributes other than the parent information are also included in the difference information, the amount of difference information is further increased.

  Therefore, in the present embodiment, in the case of FIG. 8A, when the state change of the target link is “deleted” and the state change of the link destination document is “update” or “exist”, the target link is traced. At this time, the difference information indicating that the real node of the link destination document is “deleted” is not incorporated in the tree structure difference data 220. Instead, at this point, a flag indicating that there is a possibility that the actual node of the linked document may be deleted is turned on, and the remaining search is continued. If a link pointing to the real node or its descendants is found in the remaining search, the flag is turned off and difference information corresponding to the link is created. Thereby, the above-mentioned problem is avoided. Also, if the real node has never been visited at the end of the search, the deletion flag remains on, so the difference information to delete the real node is created and the tree structure difference data Add to 220. In this way, if the document itself continues to exist even after the link is deleted, whether or not the node corresponding to the document is finally deleted on the tree structure is determined only after the search of the hypertext structure is completed. Therefore, generation of useless difference information can be reduced, and the data amount of the tree structure difference data 220 can be reduced.

  The above is processing when the state change of the target link is “deletion” and the state change of the link destination document is “update” or “existence”. By incorporating information at the end of the tree structure difference data 220, the data amount of the tree structure difference data 220 can be reduced.

  FIG. 8B shows an example of the difference information determination table 208B when it is determined that the link of interest has been the shortest path in the past but is not currently the shortest path.

  In this table 208B, the contents of each column corresponding to the case where the link state change is “new” and “non-existing” and the row corresponding to the case where the state change of the document is “new” and “non-existence” are as follows: This is the same idea as in the case of the table 208A described above.

  Further, when the document state change is “delete” and the link state change is “delete” or “exist”, the difference information of the tree structure is “deleted”. is there.

  Further, when the state change of the link of interest is “deleted” and the state change of the link destination document is “update” or “exist”, the node state is not changed in the same way as in the case of the table 208A.

  On the other hand, the cell in which the state change of the target link is “existing” and the state change of the link destination document is “update” or “existing” is different from the case of the table 208A. Such a combination of state changes means that the linked document is a real node on the tree structure at the past time point, and has changed to a reference node at the present time point. Therefore, in these combinations, the difference information is “new” and “not the shortest path”. This difference information means that, in the tree structure, a reference (shortcut) node indicating the link destination document is added as a child to the link source document node of the target link.

  FIG. 8C shows an example of the difference information determination table 208C in the case where the link of interest has not been the shortest path in the past but is now found to be the shortest path.

  In this table 208C, the contents corresponding to the column corresponding to the case where the link state change is “non-existing” and the row corresponding to the case where the document state change is “non-existing” are the same as those in the case of the above-described table 208A. It is.

  When the link state change is “new” and the document state change is “new”, both the document and the link appear newly. Here, since the link is the shortest path, the difference information corresponding to the combination is “new” and “is the shortest path”. This difference information means that, in the tree structure, a real node indicating the link destination document is added as a child to the link source document node of the link of interest. In the example of FIG. 9, the entry corresponding to the document 7 on the second line corresponds to the difference information in this case.

  If the link status change is “new” and the document status changes are “update” and “exist”, the document existed in the past, but the link was not the shortest path. Although it was not a node, it is now the shortest path, meaning that the document has become a real node. Information indicating “update” and “is the shortest path” is registered in the cell corresponding to such a combination. In the tree structure difference data 220 shown in FIG. 9, the entry corresponding to the document 8 on the sixth line is an example of difference information created from this information. Since this difference information indicates that the document 8 is a real node and has been “updated” to be a child of the document 7, the server 12 changes the real node of the document 8 to the document 7 in the tree structure. Is a child of the real node that indicates

  In the combination in which the state change of the target link is “new” and the state change of the linked document is “deleted”, there is no change in the tree structure, so it is not necessary to create difference information.

  If the state change of the link of interest is “deleted” and the state change of the link destination document is “update”, “delete”, or “exist”, the content of the corresponding cell is “deleted”. In this case, difference information indicating that a node indicating a link destination document existing in the past is deleted is created.

  The contents of the cell corresponding to the combination in which the state change of the target link is “existing” and the state change of the link destination document is “new” are “new” and “is the shortest path”. In this case, difference information indicating that the real node indicating the link destination document is added as a child of the link source document is created.

  When the state change of the link of interest is “existing” and the state change of the linked document is “update” or “exist”, the contents of the corresponding cell are “main body update”, “shortcut deletion” and “It is the shortest path”. “Main body update” means that, in the past, the parent of the link destination node of the link of interest is a node other than the link source node, and is updated to the link source node. Also, “shortcut deletion” means that the reference node (shortcut) of the link destination node existing as a child of the link source node is deleted along with the update. Accordingly, in this case, in the example of FIG. 9, two entries corresponding to the shortcuts of the document 4 and the document 4 shown in the fourth and fifth lines are created as difference information.

  The contents of the cell corresponding to the combination in which the state change of the target link is “existing” and the state change of the link destination document is “deletion” is “deletion”. In this case, difference information indicating that the reference node indicating the link destination document is deleted from the child of the link source document is created.

  FIG. 8D shows an example of the difference information determination table 208D when it is determined that the link of interest is not the shortest path in the past and is not the shortest path at present.

  In this table 208D, the column corresponding to the case where the link state change is “non-existing” and the content of the horizontal row corresponding to the case where the document state change is “non-existing” are the same as in the case of the above-described table 208A. It is.

  If the status change of the attention link is “new” and the status change of the linked document is “new”, “update”, or “existence”, the attention link did not exist in the past, so the tree structure There was no node corresponding to the linked document. On the other hand, since the attention link is newly generated and the link destination document currently exists, a new node is added to the tree structure. However, since the link of interest is not the shortest path, the newly added node is a reference node. For this reason, the contents of the cells corresponding to these combinations are “new” and “not the shortest path”. In the example of FIG. 9, the entry corresponding to the shortcut of the document 1 on the third line is an example of such difference information.

  In the combination in which the state change of the target link is “new” and the state change of the linked document is “deleted”, there is no change in the tree structure, so it is not necessary to create difference information.

  If the state change of the link of interest is “deleted” and the state change of the link destination document is “update”, “delete”, or “exist”, the content of the corresponding cell is “deleted”. In this case, since the target link itself is deleted, the tree-structured node corresponding to the link destination of the link disappears. Therefore, difference information indicating that a node indicating a linked document existing in the past is deleted is created.

  The content of the cell corresponding to the combination in which the state change of the target link is “existing” and the state change of the link destination document is “new” is “new”. In this case, difference information indicating that the reference node indicating the link destination document is added as a child of the link source document is created.

  When the state change of the target link is “existing” and the state change of the link destination document is “update” or “exist”, the link destination document is not a reference node (not a past shortest path) in the past. Since the current node is still a reference node (since it is not the shortest path at present), the tree structure does not change. For this reason, it is not necessary to create difference information.

  The contents of the cell corresponding to the combination in which the state change of the target link is “existing” and the state change of the link destination document is “deletion” is “deletion”. In this case, difference information indicating that the real node indicating the link destination document is added as a child of the link source document is created.

  FIG. 9 shows an example of the tree structure difference data 220 created by the breadth-first search of the hypertext structure using the difference information determination tables 208A to 208D described above and the post-collection link table 212 of FIG. 7B.

  Each row of the difference data 220 corresponds to one difference information. One difference information includes “document ID”, “operation”, “parent document ID”, and “other attributes”. A set of “document ID” and “parent document ID” indicates that the former is the latter child in the tree structure. “Operation” indicates an operation content for updating the tree structure.

  When the content of “operation” is “new”, a node of “document ID” is newly added to the tree structure as a child of “parent document ID”.

  When the content of “operation” is “update”, the parent of the real node indicated by “document ID” in the tree structure is changed to the node of “parent document ID”. This implies that the parent-child relationship between the real node of the “document ID” and the parent at the previous update timing is deleted at the same time. For example, the fourth line in FIG. 9 shows an “update” process for changing the real node of the document 4 to a child of the document 5, which includes the document that existed at the previous update timing (FIG. 3A). 2 and document 4 are deleted. When the “operation” content of the difference information is “update”, the server 12 determines that the parent of the actual node of the document indicated in the difference information is different from the parent of the actual node of the document at the previous update timing. The parent-child relationship is changed as described above.

  Note that when the “operation” content of the difference information is “update”, the case where the content of the document is changed is also included. If the “operation” content of the difference information is “update”, the server 12 determines whether or not the parent-child relationship has been changed to the actual node of the document indicated by the difference information, and if not, the content of the document Is determined to have been updated.

  When the content of “operation” is “delete”, the node of “document ID” that is a child of “parent document ID” is deleted from the tree structure.

  Since the relationship between the entry of each row of the difference data and the structural change shown in FIGS. 3A and 3B has already been described, the description is omitted here.

  In this way, if the tree structure difference data 220 as shown in FIG. 9 is created and transmitted to the server 12, the server 12 can update the tree structure.

  Returning to FIG. 4 again, the tree structure information collection device 16 transmits the tree structure difference data 220 created in this way to the tree structure information providing server 12 (S14).

  The tree structure information providing server 12 receives the tree structure difference data 220 (S20), and in accordance with the data 220, the tree structure corresponding to the tree structure information collecting apparatus 16 of the difference data transmission source stored in the storage device 122. Data is updated (S22).

  When the tree structure information collection device 16 transmits the tree structure difference data 220 to the server 12, the “previous” document table 204 and “previous” at the next update timing are retrieved from the post-collection document table 210 and post-collection link table 212. A link table 206 is created. That is, since the post-collection document table 210 and the post-collection link table 212 indicate the transient “state change” of the document and link, the “state change” is rewritten to “state”. FIG. 10 shows the rewriting rules. In the rule of FIG. 10, when the state changes are “new”, “update”, and “exist”, they are converted into a state of “exist”. If the state change is “deleted” or “non-existing”, it is converted into a state of “non-existing”.

  In this conversion, information on whether or not the shortest path is obtained when the hypertext structure is searched by the breadth-first search is reflected in the link table.

  The document table 204 ′ and the link table 206 ′ obtained by the rewriting as described above are as shown in FIGS. 11 and 12. These are created by rewriting the tables in FIGS. 6B and 7B, respectively.

  The embodiments of the present invention have been described above. As can be seen from the above description, according to the present embodiment, when the tree structure information collection device 16 creates the tree structure difference data 220, the process corresponding to the tree structure search includes a hypertext structure breadth-first search. Therefore, the processing load is reduced as compared with the conventional technique in which the tree structure is searched twice. Further, since it is not necessary to save a huge amount of data such as the previous tree structure data, the storage capacity of the device 16 can be saved.

  The tree structure information providing server 12 and the tree structure information collecting apparatus 16 described above execute a program describing the processing contents of each functional module described in this specification using hardware of a general computer system. This can be realized.

  The embodiments described above are merely examples, and various modifications can be made within the scope of the present invention. For example, in the above example, the tree structure information providing server 12 provided the tree structure information of a document in the company's internal network to a certain customer company, but the tree structure of the company document is disclosed on the Internet 10. It can also be merged into the tree structure of the existing documents and provide the merge result. In the above example, the hypertext structure is circulated according to the breadth-first search rule to create the tree structure difference data 220. Alternatively, a method using another search rule such as a depth-first search rule may be used. Good.

  In addition, the above explanation did not touch on the display of the tree structure on the client device side that acquired the tree structure difference data in detail, but actually the tree structure is not displayed on the client device side. In order to identify individual documents constituting the document, a document name is displayed for each document icon on the tree structure display (tree display). There are some points to be noted when displaying this document name.

  That is, the file name or URL of the document is first considered as the document name of the web document. However, there is a problem that it is difficult for the viewer to identify the document. As a solution to this, there is a method of displaying a title (indicated by a <TITLE> tag) included in an HTML description of a web document as a document name. This is a good solution because the title is often a phrase that is generally easy for the viewer to understand. However, this method has drawbacks. This will be described with reference to FIG.

  FIG. 13 is a diagram illustrating an example of the browser screen 400 displayed on the display device of the client device. The browser screen 400 includes a document display screen 410 that displays a web document and a tree display screen 420 that displays a tree structure of a document group for navigation. In the tree display screen 420, the tree structure data provided by the tree structure information providing server 12 and the tree structure of the document configured based on the difference data are displayed in the same manner as the tree structure formed by folders and files in the file system. It is shown in The file system tree shows the inclusion relationship between folders and files, whereas the document tree structure is different only in the relationship between the link source and link destination between documents. In the tree display screen 420, each document is indicated by an icon 422 indicating a title character string in the HTML description.

  The problem here is that, when the icon 422 displaying the title character string is displayed as a document being displayed in the tree, there is a discrepancy in wording with the link description 412 in the Web document display. In a web document, the linked document is an anchor text (a character string sandwiched between anchor tags <A href=URL> and </A>. On the display, it is generally distinguished from others by an underline) However, since the anchor text is freely described by the creator of the link source document and is independent of the title described in the link destination document, they generally do not match. If the document name on the tree display is different from the link description appearing on the actual web document, it is difficult for the user to grasp which actual web document the document on the tree display indicates.

  This is the case where the title is displayed as the document name in the tree display. However, it is considered that the file name is the same as the document name, or it becomes more difficult for the user to understand.

  On the other hand, as shown in FIG. 14, when the document name indicated by the icon 422A of each document in the tree display screen 420A matches the link description 412 in the web document, the tree display screen 420 is displayed to the user. There is an advantage that it is easy to grasp which document is an individual document to be displayed.

  Here, in the case of the system that transmits the differential data of the tree structure from the individual tree structure information collecting device 16 to the central tree structure information providing server 12 as in the present embodiment, the display as shown in FIG. 14 is realized. Requires a mechanism in which the collection device 15 detects a change in the anchor text of the link description in the web document and transmits the change to the server 12. A modified example of such a mechanism will be described below.

  In this modification, the configuration of the system and apparatus may be the same as that of the above embodiment shown in FIGS. 1 and 2, and the processing contents of the collection unit 162, the difference data creation unit 164, and the tree structure update unit 128 are different. Only.

  In this modification, the collecting unit 162 collects not only the URL of the link but also the accompanying anchor text at the time of collecting the link by visiting the web, and records this in the link table as one of the link attributes. For users in web surfing, the anchor text functions as an identification name for identifying the link destination document of the link. In this example, the anchor text is used as the “reference name” of the link destination document. handle. That is, the link anchor text is registered in the link table as the value of the reference name, which is one of the link attributes. Then, the difference data creation unit 164 compares the reference name (anchor text) of each link in the previous link table 206 with the reference name of each link in the post-collection link table 212 obtained by the collection, and the same When the reference name for the link changes before and after the update, information indicating this change is incorporated into the tree structure difference data 220 and sent to the tree structure information providing server 12.

  In order to provide a user-friendly tree display, the tree structure information providing server 12 provides the client device 18 with tree structure data 124 indicating the reference name of the link leading to the document as the document name of each document. Here, since the hypertext structure is expanded into a tree structure, when links are made from a plurality of documents to the same document, a plurality of nodes corresponding to each link on the tree structure (in the above embodiment, of which In this modification, the anchor text indicated by the link to each node is registered as a reference name for each of these nodes. In order to maintain such tree structure data 124, the tree structure update unit 128 updates the tree structure data 124 based on the tree structure difference data 220 received from the tree structure collection device 16. The reference name of the document in the tree structure data 124 is changed according to the reference name change information.

  The above process will be described using an example. For example, assume that the hypertext structure in the state shown in FIG. 15A at a certain update timing becomes as shown in FIG. 15B at the next update timing. In this modification, the anchor text of each link is registered as a reference name in the link table 210, so the previous link table 206a (see FIG. 16A) corresponding to the state of FIG. 15A and the post-collection link table corresponding to the state of FIG. 15B. A data item “reference name” is provided in 212a (see FIG. 16B). The example of FIGS. 15A and 15B is obtained by adding the document with the document ID “9” to the example of FIGS. 3A and 3B. The difference between the previous time and the current time indicates that the document ID has the document ID “5”. It is the same as the example of FIG. 3A and FIG. 3B except that the reference name of the link “9” is updated. Similarly, each link table shown in FIGS. 16A and 16B is obtained by adding a record about a link to the document with the document ID “9” to each link table in FIGS. 7A and 7B. In FIGS. 15A and 15B, and FIGS. 16A and 16B, the difference is clearly shown by attaching a thin ink to the same contents as FIGS. 3A and 3B, and FIGS. 7A and 7B, respectively.

  In this example, as can be seen from FIGS. 16A and 16B, when the reference names for the link from the document with the document ID “5” to the document with the document “9” are collected at the previous update timing, the “proposal form” is used. In contrast, when it was collected at this collection timing, it changed to an “idea sheet”. The difference data creation unit 164 finds a link from the document “5” to “9” whose reference name has changed as described above, by comparing the previous link table 206a and the post-collection link table 212a. Then, as shown in FIG. 17, the difference data creation unit 164 uses the change information of the reference name as difference information (the bottom row of the table of FIG. 17) regarding the document “9” that is the link destination document of the link. It is incorporated in the tree structure difference data 220a. In this tree structure difference data 220a, in addition to the data items of the tree structure difference data 220 shown in FIG. 9, a data item of “reference name” is provided, and the reference name of the link to itself has changed. In the difference information record of the document, “update” is incorporated as the value of “operation”, and the changed reference name “idea sheet” is incorporated as the value of “reference name”.

  Although omitted in the example of FIG. 17 to avoid complication, if the “operation” value of the difference information of a document having the tree structure difference data 220a is “new”, the document (or reference) This means that the reference name (anchor text) of the document obtained by the current collection is included in the difference information record. Also, if the value of “Operation” of a document is “update” for reasons other than the change of the anchor text, it may be due to the change of the parent document, and if the parent document changes, the anchor text that generally points to the document is changed. The contents also change. Therefore, in general cases where the value of “operation” is “update”, the reference name of the document obtained at the time of the current collection is incorporated into the difference information record. In the example of FIG. 9, “operation” is “update” in the difference information record of the document corresponding to the real node. However, in this modification, “operation” is “update” even when the anchor text is changed. Therefore, the “operation” may be “update” for the difference information record corresponding to the reference shortcut node.

  The tree structure information providing server 12 that has received the tree structure difference data 220a from the tree structure information collection device 16 uses the value of the reference name (if any) of the difference information record of each document included in the difference data 220a. Compared with the reference name of the node of the document in the current tree structure data 124, if both are different, the reference name in the tree structure data 124 is changed to the reference in the difference information record. For a “new” document in the tree structure difference data 220a, a node indicating the document is created in the tree structure data 124 as a child node of the node indicated by the “parent document ID”, and the node of the document is determined. The reference name shown in the tree structure difference data 220a is set.

  As described above, in this modification, the tree structure information collection device 16 detects a change in the anchor text set for the link in the document in the same manner as the change in the link structure of the hypertext, and detects the change information as a tree structure difference. Since it is incorporated in the data 220a and transmitted to the tree structure information providing server 12, the tree structure information providing server 12 matches the display of the link (anchor text) during web surfing to the client device 18. A tree display showing highly likely document names can be provided (see FIG. 14).

  In this modification, the anchor text is displayed as a reference name for each node (document) in the tree structure, so that the consistency with the web surfing is easy to understand, but the real node and the reference shortcut node There is a demerit that it becomes difficult to understand whether or not the documents pointed to by these nodes are the same entity. For this, in addition to the anchor text, at least one of the file name (or URL) of the document and the title in the HTML description of the document is set as a reference name for the document node in the tree structure. In the tree display, it is preferable to display the plural types of names for one document side by side. It is also preferable to provide both a mode for displaying side by side and a mode for displaying only anchor text in the browser of the client device 18 so that the mode can be selected. In this case, the tree structure information collecting device 16 collects the titles of the linked documents of the links in the document in the same manner as in the case of the anchor text, and if there is a change in the title, the changed title is displayed. The difference information record of the link included as an attribute may be incorporated into the tree structure difference data 220a. The file name (or URL) of the linked document may be added by referring to the circulation list 202. That is, for each update timing, if the difference of the circulation list 202 from the previous update timing (only the first timing is the entire circulation list 202) or the circulation list 202 is provided from the tree structure information collecting device 16 to the tree structure information providing server 12. The file name and URL information can be added to each document node of the tree structure data 124.

  This modified example can also be realized by the following processing procedure in addition to the processing procedure described above.

  In this processing procedure, a state of “update” is newly introduced as a link state, and tree structure difference data is created using a difference information determination table including a rule for a case where the link state is “update”.

  Therefore, the collection unit 162 of the tree structure information collection device 16 determines the link based on the comparison between the link description anchor text in the collected document and the link reference name shown in the previous link table 206a corresponding to the link description. When it is detected that the reference name is updated, “update” is set in the state change column of the corresponding link in the post-collection link table 212a as shown in FIG.

  When the collection unit 162 finishes creating the post-collection link table 212a and the like in this way, the difference data creation unit 164 then uses the difference information determination tables 208E to 208H as shown in FIGS. Create data. The difference information determination tables 208E to 208H correspond to the difference information determination tables 208A to 208D shown in FIGS. 8A to 8D are different from the case where the link state change corresponds to “update”, only the case corresponding to the case of that column will be described below.

  In the link search when creating the tree structure difference data, if the target link is the shortest path in the past and is still the shortest path, the difference information determination table 208E in FIG. 19A is referred to. At this time, let us consider a case where the state change of the attention link is “update”. In this case, the state change of the linked document cannot be “new” or “non-existent”. On the other hand, if the status change of the linked document is “updated”, it means that the content of the document has been updated and the reference name of the document is the same as when the status change of the target link is “exist”. Also means it has been updated. Therefore, for this attention link, the difference information in which the value of “operation” is “update” and the reference name is the updated value is incorporated into the tree structure difference data (see FIG. 17). The updated reference name may be retrieved from the post-collection link table 212a. Further, when the state change of the linked document is “deleted”, the document itself is deleted, and thus the difference information of the tree structure is “deleted”. Also, if the status change of the linked document is “existing”, it means that the document itself has not changed, but the reference name has been changed. In this case, “operation” becomes “update”, and the link after collection Difference information including the reference name of the link retrieved from the table 212a is created.

  Next, when the link of interest has been the shortest path in the past, but is not currently the shortest path, the difference information determination table 208F in FIG. 19B is referred to. At this time, let us consider a case where the state change of the attention link is “update”. In this case, the state change of the linked document cannot be “new” or “non-existent”. On the other hand, when the state change of the linked document is “deleted”, the document itself is deleted, and thus the difference information of the tree structure is “deleted”. In addition, when the state change of the link destination document is “update” or “exist”, the document becomes a reference (shortcut) node as in the case where the state change of the target link is “exist”, and the link has the shortest path. Is not. Therefore, difference information including the information that the operation is “new” and “not the shortest path” and including the reference name obtained from the post-collection link table 212a is created.

  Next, when the link of interest has not been the shortest path in the past, but is currently the shortest path, the difference information determination table 208G in FIG. 19C is referred to. At this time, if the state change of the target link is “update”, the state change of the linked document cannot be “non-existent”. On the other hand, when the state change of the linked document is “deleted”, the document itself is deleted, and thus the difference information of the tree structure is “deleted”. Also, if the status change of the linked document is “new”, it means that the document has become a new real node and the reference name has been updated. Therefore, as the difference information, the operation is “new” and “shortest” A record including the information “is a path” and the updated reference name is created and incorporated in the tree structure difference data. Further, when the state change of the linked document is “update” or “exist”, it means that the reference (shortcut) node has become a real node, so the operation is “update” and “ Difference information including the information “is the shortest path” and including the reference name obtained from the post-collection link table 212a is created.

  Finally, when the link of interest is not the shortest path in the past and is not the shortest path at present, the difference information determination table 208G in FIG. 19D is referred to. At this time, if the state change of the target link is “update”, the state change of the linked document cannot be “non-existent”. On the other hand, when the state change of the linked document is “deleted”, the document itself is deleted, and thus the difference information of the tree structure is “deleted”. If the status change of the linked document is “new”, the contents of the cell corresponding to the combination in the table 208G are “new”. In this case, difference information including the fact that the reference node indicating the link destination document is added as a child of the link source document and the reference name is created. If the status change of the linked document is “update” or “exist”, the linked document remains the reference (shortcut) node and the reference name has changed, so the operation is “update”. And the difference information including the reference name obtained from the post-collection link table 212a is created and incorporated in the tree structure difference data.

  Through the processing described above, tree structure difference data including information on changing a reference name can be created.

It is a figure which shows an example of the whole structure of the system with which this invention is applied. It is a figure which shows an example of a functional module structure of the tree structure information provision server 12 and the tree structure information collection device 16. It is a figure which shows the hypertext structure which the document group in a certain internal network comprised at the last tree structure update timing. It is a figure which shows the hypertext structure which the document group in the same company network comprises at this tree structure update timing. It is a figure which shows the flow of the update process of the tree structure data in the system of this embodiment. It is a figure which shows an example of the data content of a circulation list. It is a figure which shows an example of the data content of the last document table. It is a figure which shows an example of the data content of the document table after collection. It is a figure which shows an example of the data content of the last time link table. It is a figure which shows an example of the data content of the link table after collection. It is a figure which shows an example of a difference information determination table. It is a figure which shows an example of a difference information determination table. It is a figure which shows an example of a difference information determination table. It is a figure which shows an example of a difference information determination table. It is a figure which shows an example of tree structure difference data. It is a figure which shows the rewriting rule of the post-collection document table and the post-collection link table. It is a figure which shows an example in the data content of the "previous document table" used at the next update timing obtained by rewriting. It is a figure which shows an example in the data content of the "previous link table" used at the next update timing obtained by rewriting. It is a figure for demonstrating the problem of the document name display in the document navigation using a tree structure. It is a figure which shows the example of a document name display easy for a user to understand. It is a figure which shows the hypertext structure in a certain time. FIG. 15B is a diagram showing the structure after the hypertext structure of FIG. 15A has been changed (including change of anchor text in the document). It is a figure which shows an example of the data content of the last link table in the modification for implement | achieving a document name display easy for a user. It is a figure which shows an example of the data content of the link table after collection in the modification for implement | achieving a document name display easy for a user. It is a figure which shows an example of the tree structure difference data produced in the modification for implement | achieving a document name display easy for a user. It is a figure which shows an example of the data content of the post-collection link table produced | generated in another process sequence of a modification. It is a figure which shows an example of the difference information determination table used in another process procedure of a modification. It is a figure which shows an example of the difference information determination table used in another process procedure of a modification. It is a figure which shows an example of the difference information determination table used in another process procedure of a modification. It is a figure which shows an example of the difference information determination table used in another process procedure of a modification.

Explanation of symbols

  10 Internet, 12 Tree structure information providing server, 14a, 14b, 14c Internal network, 16 Tree structure information collection device, 162 Collection unit, 164 Difference data creation unit, 200 Storage device, 202 Travel list, 204 Previous document table, 206 Previous time Link table, 208 difference information determination table, 210 post-collection document table, 212 post-collection link table.

Claims (5)

In response to changes in the hypertext structure documents on the network is configured, in order to update the tree structure data representing the hypertext structure, tree showing the change of the tree structure data corresponding to a change in the hypertext structure An apparatus for creating structural difference data,
A pre-update document table in which information indicating whether each document exists on the hypertext structure at a time before the change of the hypertext structure is registered;
A pre-update link table in which links between documents existing on the hypertext structure at the time before the change of the hypertext structure are registered;
By examining the document group after the change while referring to the pre-update document table and the pre-update link table, a document and a link added to and deleted from the hypertext structure due to the change are detected. A differential processing unit that creates a tree structure difference data indicating a change in the tree structure data by tracing an inter-document link of the hypertext structure after the change in order from a predetermined head document while referring to a detection result;
With
The difference processing unit
Said pre-update document table registered document group, and collects the linked document group thereof each document from over the network, based on whether or not acquired their respective documents, they each document, the change Among those that existed before and the document content has not been updated, those that existed before the change and the document content has been updated, those that have been added due to the change, and those that have been deleted due to the change A first process for creating a post-collection document table that indicates which is,
Based on a comparison between an inter-document link group included in each document collected from the network and an inter-document link group registered in the pre-update link table, each inter-document link exists before the change. A second process of creating a post-collection link table that indicates which one of the one added by the change and the one deleted by the change;
When tracing the inter-document link of the hypertext structure after the change in order from the top document, the post-collection document table and the post-collection state of the inter-document link state and the link destination document indicated by the inter-document link are indicated. Based on a combination of the state of the link between documents and the state of the link-destination document retrieved from the link table, and (a) the link source document indicated by the link between documents in the tree structure data A change that a node corresponding to the link destination document indicated by the inter-document link is added as a child of the corresponding node; (b) a node corresponding to the link destination document indicated by the inter-document link in the tree structure data Is deleted from the child of the node corresponding to the link source document indicated by the inter-document link, and (c) the tree structure data And a change corresponding to any one of the change contents that the contents of the link destination document are updated with respect to the node corresponding to the link destination document indicated by the inter-document link is added to the tree structure data A third process for adding data indicating the change contents to the tree structure difference data when determined;
As well as
The difference processing unit is a change indicating a combination of the inter-document link state and the link-destination document state, presence / absence of a change in the tree structure data corresponding to the combination, and a change content when there is a change. The change contents include at least three types (a), (b), and (c), and the hypertext structure after the change in the third process is included in the change contents. When the inter-document link is traced, the change contents corresponding to the combination of the state of the inter-document link obtained by the search from the post-collection document table and the post-collection link table and the state of the link destination document are When determined from the change rule table, it is determined that a change corresponding to the determined change content has been added to the tree structure data .
A tree structure difference data creation device characterized by that. The tree structure difference data creation device according to claim 1,
The pre-update link table has shortest path information indicating whether each inter-document link corresponds to a shortest path from the first document to a link destination document of the inter-document link,
The difference processing unit has a plurality of change rule tables, and when tracing the inter-document link of the hypertext structure after the change, each inter-document link is linked to the inter-document link from the first document. A tree structure using a change rule table corresponding to a combination of the determination result and the shortest path information of the inter-document link in the pre-update table Differential data creation device.   The tree structure difference data creation device according to claim 1, wherein the difference processing unit corresponds to the link destination document when following the hypertext structure inter-document link after the change in order from the first document. When it is determined that a change to delete the node from the tree structure data is necessary, the data indicating the change content is not added to the tree structure difference data at that time, and all remaining inter-document links are traced. A tree structure difference data creation device that adds data indicating the change contents to the tree structure difference data only when no inter-document link to the link destination document is found at the time. The tree structure difference data creation device according to claim 1,
In the pre-update link table, for each inter-document link, an anchor text described in association with the inter-document link in the document in which the inter-document link is described is registered,
The difference processing unit detects an anchor text attached to each inter-document link included in each document collected from the network, and registers the anchor text in the post-collection link table in association with the inter-document link. When the anchor text for the same inter-document link is different between the pre-update link table and the post-collection link table, the anchor text of the inter-document link is changed to the anchor text indicated in the post-collection link table. Adding data indicating the change to the tree structure difference data;
A tree structure difference data creation device characterized by that. In order to update the tree structure data representing the hypertext structure according to the change of the hypertext structure formed by the documents on the network, the computer system changes the tree structure data corresponding to the change of the hypertext structure. A program for causing a computer system to function as a device for creating tree structure difference data indicating:
A pre-update document table in which information indicating whether each document exists on the hypertext structure at a time before the change of the hypertext structure is registered;
A link table before update in which links between documents existing on the hypertext structure at the time before the change of the hypertext structure are registered;
By examining the document group after the change while referring to the pre-update document table and the pre-update link table, a document and a link added to and deleted from the hypertext structure due to the change are detected. A differential processing unit that creates a tree structure difference data indicating a change in the tree structure data by following a hypertext structure inter-document link after the change in order from a predetermined head document while referring to a detection result,
As well as
The difference processing unit
Document groups registered in the pre-update document table and link destination document groups of the respective documents are collected from the network, and based on whether or not each of the documents has been acquired, the respective documents are changed before the change. Existing in the document and the document content has not been updated, one that has existed before the change and the document content has been updated, one that has been added due to the change, and one that has been deleted due to the change A first process of creating a post-collection document table indicating whether or not
Based on a comparison between an inter-document link group included in each document collected from the network and an inter-document link group registered in the pre-update link table, each inter-document link exists before the change. A second process of creating a post-collection link table that indicates which one of the one added by the change and the one deleted by the change;
When tracing the inter-document link of the hypertext structure after the change in order from the top document, the post-collection document table and the post-collection state of the inter-document link state and the link destination document indicated by the inter-document link are indicated. Based on a combination of the state of the link between documents and the state of the link-destination document retrieved from the link table, and (a) the link source document indicated by the link between documents in the tree structure data A change that a node corresponding to the link destination document indicated by the inter-document link is added as a child of the corresponding node; (b) a node corresponding to the link destination document indicated by the inter-document link in the tree structure data Is deleted from the child of the node corresponding to the link source document indicated by the inter-document link, and (c) the tree structure data And a change corresponding to any one of the change contents that the contents of the link destination document are updated with respect to the node corresponding to the link destination document indicated by the inter-document link is added to the tree structure data A third process for adding data indicating the change contents to the tree structure difference data when determined;
As well as
The difference processing unit is a change indicating a combination of the inter-document link state and the link-destination document state, presence / absence of a change in the tree structure data corresponding to the combination, and a change content when there is a change. The change contents include at least three types (a), (b), and (c), and the hypertext structure after the change in the third process is included in the change contents. When the inter-document link is traced, the change contents corresponding to the combination of the state of the inter-document link obtained by the search from the post-collection document table and the post-collection link table and the state of the link destination document are When determined from the change rule table, it is determined that a change corresponding to the determined change content has been added to the tree structure data .
A program characterized by that.

Images