JP5476867B2 - Mashup program, mashup device, and mashup method - Google Patents

Description

  The present invention relates to a mashup program, a mashup device, and a mashup method, and more particularly to a mashup program, a mashup device, and a mashup method that use a technique called mashup that matches resources on a network.

  In recent years, a technique called mashup of information resources (for example, Web pages) on a network such as the Internet or an intranet is often used. The mashup is to form new contents and services by combining contents and functions of a plurality of different providers (see, for example, Patent Document 1).

  In web mashups, search, photos, videos, maps, and other service sites publish APIs (Application Program Interface) to combine photos and shooting location maps, or zip code data and maps. Service to do has been built.

  Web mashups not only build such services at the program level, but also become easier at the end-user level, if you can share awareness, collaborate on web pages, progress management, portal creation, advertising It is expected to be used in various situations such as systems.

  A part of content or function on a certain web page can be cut and pasted on another web page to make it available. That is, when accessing a certain web page, the contents of another web page at that time are displayed. And enabling a function such as search and video browsing require a technique for cutting a part of a specified Web page and a technique for pasting a part of a Web page to another specified Web page. Become.

  As a technique for cutting out a part of a Web page, the following techniques are well known as clipping techniques or scraping techniques. For example, as a Web clipping technique, one that cuts out a part of a specified Web page as an image (snapshot) is known. As a web scraping technique, a technique is known in which data is extracted from a designated web page by pattern matching or position designation by Xpath (see, for example, Patent Document 2).

  On the other hand, a technique for giving information to a certain Web page is known as a Web annotation technique, and there are various formats such as text, images, HTML fragments (Note: In addition to HTML, there are various formats such as XHTML, XML, and CHTML for Web page descriptions. Data that can be handled in HTML, such as those described as HTML on behalf of them, is embedded in another Web page (HTML).

  In this way, if the technology for cutting data from a web page and the annotation technology for adding data to another web page are combined, data clipped or scraped from one web page can be mashed up (composited) on another web page. ) Can be displayed.

JP 2008-299414 A JP 2004-110427 A

  However, in normal clipping and scraping, the original Web page is saved as an image as a snapshot, or a fragment of the original Web page structured document (HTML) itself, or a structure (tag) is removed from the structured document. Since the data is stored, the latest information at the time when the mashup screen is accessed is not reflected, and the functions such as search and moving image that were in the original Web page cannot be used.

  As an existing system, there is a system in which a gadget (widget) prepared in advance is pasted on a specific Web page. In addition, as an existing system, there is a system in which a fragment of a structured document is prepared in advance by saying “Please paste this on your blog”. In these systems, a screen having a specific function can be pasted on another Web page, but only a prepared one can be pasted, and a part of an arbitrary web page cannot be pasted.

  Further, the user commonly expects that some function is handed over to another Web page by cutting a part of the content or function on the Web page and pasting it on another Web page. However, in order to transfer the function expected by the user to another Web page, it is necessary to cut out an appropriate range of the HTML fragment from a certain Web page.

  However, it is not easy for a general user to cut out an appropriate range of an HTML fragment from a certain Web page in order to transfer the expected function to another Web page.

  One embodiment of the present invention has been made in view of the above points, and provides a mashup program, a mashup device, and a mashup method for transferring a function of an arbitrary resource on a network to a new resource formed by mashup. The purpose is to provide.

One embodiment of the present invention is a mashup program for operating a computer that pastes a cut range cut from a cut-off Web resource to a paste-destination Web resource, and that means that the computer is independent for realizing a function. If a node given as a cut-out range is not an unresolved part based on a table in which a component tag that does not form a component and a configuration tag that realizes a function by combining the component tags, the tree structure of the node An unresolved part confirmation unit for confirming whether there is an unresolved part node at a lower level, and when the unresolved part node is confirmed, the unresolved part node is an upper node of the tree structure of the node given as the cut range. Find the configuration node corresponding to the solution part node, and use the elements below the found configuration node as the normalization cut range. It operates as a related configuration acquisition unit.

  In addition, what applied the component, the expression, or the arbitrary combinations of the component of one Embodiment of this invention to a method, an apparatus, a system, a computer program, a recording medium, a data structure, etc. is also effective as an aspect of this invention. .

  According to an embodiment of the present invention, it is possible to provide a mashup program, a mashup device, and a mashup method that allow a function of an arbitrary resource on a network to be transferred to a new resource formed by mashup.

It is an image figure of an example of the Web page of cut origin. It is an image figure of an example of the web page of a paste destination. It is an image figure of an example of the web page of the paste destination where the part cut from the web page of cut origin was pasted. It is an image figure of an example of the Web page of cut origin. It is a block diagram of an example of a components and a structure tag list table. It is explanatory drawing which shows the 1st method of extending a cutting range. It is an image figure of an example of the Web page of cut origin. It is a block diagram of an example of a document tree. It is explanatory drawing which shows the 2nd method of extending a cutting range. It is a block diagram of an example of an action element table. It is explanatory drawing which shows the 3rd method of extending a cutting range. It is explanatory drawing which shows the 4th method of extending a cutting range. It is a block diagram of one Example of the system containing the mashup apparatus of a present Example. It is a block diagram of an example of a computer system. It is a block diagram of an example of the web browsing device to which the mashup program of a present Example is applied. It is a flowchart showing the process of the cutting source selection part. It is a flowchart showing the process of the attachment destination selection part. It is a flowchart showing the process of the range presentation selection part. It is a flowchart showing the process of the web mashup realization apparatus. It is a flowchart of an example showing the process sequence of the unresolved parts confirmation part. It is a flowchart of an example showing the processing procedure of the option of an unresolved parts confirmation part. It is an example flowchart showing the processing procedure of the related configuration acquisition unit. It is a flowchart of an example showing the process sequence of the 2nd method of extending a cutting range. It is a block diagram of an example of a script source description attribute table. It is an example of the flowchart showing the process sequence of the 3rd method of extending a cutting range. It is a flowchart of an example showing the process sequence which determines the relevance degree of the normalization cut range and an action element. It is a flowchart of an example showing the process sequence of the 4th method which expands a cutting range. It is a block diagram of an example of the reference / definition table of an action element. It is a block diagram of an example of an identifier management table. It is a block diagram of an example of an identifier management table.

  Next, modes for carrying out the present invention will be described based on the following embodiments with reference to the drawings. In this embodiment, a Web page will be described as an example of a resource on the network.

(Overview of this example)
In the present embodiment, in order to correctly execute the functions of the cut-out web page, such as form input, in the mashup device that cuts the screen on the web page and pastes it on another designated web page. In addition, by cutting out and holding the minimum necessary information from the cut-out range of the cut-off web page by configuration tag search, action element search, and action element narrowing, the correct function is executed on the paste-up web page I try to do it. In this embodiment, the necessary functions are executed on the Web page of the pasting destination by performing necessary normalization on the Web page of the cutting source.

  First, in order to facilitate understanding of the present embodiment, a part of contents and functions on a certain Web page is cut out and pasted on another Web page, that is, when a certain Web page is accessed. A mashup technique for displaying the contents of another Web page at that time and enabling functions such as search and video browsing will be described.

  FIG. 1 is an image diagram of an example of a Web page to be cut out. The cut-out Web page in FIG. 1 has a search window portion 1. The user moves, for example, the red frame 2 presented by the system on the Web page, selects an appropriate part, and cuts out the part on the Web page with a predetermined command (ctrl-c) or the like.

  In the Web page of FIG. 1, a search window portion 1 is designated by a red frame 2. The user can cut out the search window portion 1 from the cut-out Web page by inputting a predetermined command for cutting out the portion on the Web page in the state of FIG.

  FIG. 2 is an image diagram of an example of a Web page to be pasted. The user selects, for example, an appropriate portion by moving the red frame that the system presents on the pasting Web page, and the search window cut from the original Web page by a predetermined command (ctrl-v) or the like. Paste part 1. FIG. 3 is an image diagram of an example of a pasting destination Web page to which a portion cut from the original Web page is pasted.

  In the Web page of FIG. 3, the pasting portion is designated by a red frame 3. The user inputs a predetermined command to be pasted on the part on the Web page, and the part 1 of the search window cut out from the cut-out Web page is pasted to the destination Web page specified by the red frame 3. Can be pasted on the attachment part. In the Web page of FIG. 3, the search window portion 1 of the cut-out Web page shown in FIG. 1 is embedded in the pasting portion designated by the red frame 3.

  What is pasted on the web page to be pasted is not a mere snapshot, but has the function of the cut portion cut from the web page to be cut. In the example of FIG. 3, the search function of the cut-out Web page is held. The user can obtain a search result by inputting a query in the search window portion 1 pasted on the Web page of FIG.

The reason why the search function of the cut-out web page can be held in the paste-destination web page is as follows. First, the Web page is created by HTML or the like. The HTML around the search window portion 1 is specifically described as follows.
<h5><label for = "searchInput"> Search </ label></h5>
<div id = "searchBody" class = "pBody">
<form action = "/ w / index.php" id = "searchform">
<div>
<input type = “hidden” name = "title" value = "Special: Search"/>
<input id = "searchInput" name = "search" type = "text" accesskey = "f" value = ""/>
<input type = “submit” name = "go" class = "searchButton" id = "searchGoButton" value = "display"/>
<input type = “submit” name = "fulltext" class = "searchButton" id = "mw-searchButton" value = "search"/>
</ div>
</ form>
</ div>

  The search function of the cut-out Web page does not capture the HTML as described above as a simple text string, but cuts it out in an appropriate range in consideration of the document tree, and includes it in the HTML text of another Web page. By embedding in, it is realized on the Web page of the pasting destination.

  However, when embedding a portion cut from a cut-out web page in a paste-destination web page, depending on the range cut from the cut-out web page, the function expected by the user is retained (the same function in the paste-destination web page May not be available).

  FIG. 4 is an image diagram of an example of a Web page to be cut out. For example, the cut-out Web page in FIG. 4 specifies a range (search window) in which the red frame 4 is narrower than that in FIG. The user can cut out the search window from the cut-out web page by inputting a predetermined command for cutting out the portion on the web page in the state of FIG. It is assumed that the user moves the red frame on the pasted Web page, selects an appropriate part, and pastes a search window cut from the cut-out Web page with a predetermined command or the like.

  In FIG. 4, the red frame 4 is a search window. However, even if the search window designated by the red frame 4 is pasted on another Web page, the function expected by the user is not retained. Here, it is common sense that the function expected by the user is a search function rather than simply pasting a search window. The reason why the search function expected by the user is not maintained is as follows.

First, the web page is created by HTML or the like. The HTML around the search window is specifically described as follows.
<h5> Search </ h5>
<div id = "searchBody" class = "pBody">
<form action = "/ w / index.php" id = "searchform">
<div>
<input type = “hidden” name = "title" value = "Special: Search"/>
<input id = "searchInput" name = "search" type = "text" accesskey = "f" value = ""/>
<input type = “submit” name = "go" class = "searchButton" id = "searchGoButton"
value = "display"/>
<input type = “submit” name = "fulltext" class = "searchButton" id = "mw-searchButton" value = "search"/>
</ div>
</ form>
</ div>

  In the example of FIG. 1, the red frame 2 indicates the range of <form> to </ form>. In the example of FIG. 4, a red frame 4 indicates a range of <input id = "searchInput" name = "search" type = "text" accesskey = "f" value = "" />. The above HTML requires an entire <form> to </ form> in order to maintain an actual search function.

  When an HTML fragment is cut out and pasted on another Web page, an appropriate range must be cut out in order to maintain a function that a user expects in common sense. In the example of Fig. 4, the range of the search window is limited to make the explanation easy to understand, but the search function is retained even if <div> ~ </ div> inside <form> ~ </ form> is cut off. Can not. The range visible on the browser at this time (for example, the range specified by the red frame 2) is the same as the range specified by the red frame 2 in FIG.

  As for other examples, as will be described later, in order to retain the functions expected from the cut-out range of the cut-out web page in the paste-destination web page, the cut-out range cut from the cut-out web page is appropriate. Need to control.

  In addition, as a method of retaining the function expected from the cut-out range of the Web page of the cut-off even in the Web page of the paste-destination, instead of cutting the cut target part of the cut-off Web page, the cut-off Web page It is also conceivable that the entire page is taken out and pasted on the web page to be pasted, and the parts other than the cut target part are hidden.

  In this method, since the part other than the cut target part is not displayed, it looks like the specified part is cut off. In addition, since the entire Web page that is the cut source is held in this method, the function expected of the Web page that is the paste destination can be held without impairing the function of the Web page that is the cut source.

  However, holding the entire cut-out Web page is wasteful because it holds up to unnecessary portions. This is not only wasteful in capacity, but also causes a phenomenon that a function (explained) other than the cut target part of the cut-out Web page is executed as a side effect. For example, as a side effect, when a certain Web page is accessed, another information (such as an advertisement) is displayed in a pop-up, or music flows.

  In addition, the above method cannot respond to a case where two Web pages are to be merged, such as matching the display style to the Web page to be pasted while maintaining the functions of the Web page to be cut off. This can be considered as a kind of side effect, because only the function of the Web page of the cut-out source is necessary, but the style of the Web page of the cut-out source is inherited as it is.

  As described above, when information on an arbitrary Web page is cut out and pasted on a new Web page, an appropriate range is expected to be cut out so that the function of the arbitrary Web page can be transferred to the new Web page. Or I had to expect (or ignore) any side effects.

  In this embodiment, even if an arbitrary cutout range is specified on an arbitrary Web page, the function expected by the user is estimated from the cutout range, without causing side effects in a broad sense on the pasted Web page. This increases the rate at which the functions expected by the user work correctly. In other words, in this embodiment, by specifying an arbitrary cutout range in an arbitrary Web page, the function expected by the user is estimated from the cutout range, and not the entire original Web page is held but estimated. The part considered to be the minimum necessary for the execution of the function is cut out.

  The mashup device cuts the screen on the web page that is the cut source, holds the position as position information, and accesses the web page that is the pasting destination when the specified paste destination web page is accessed. By extracting the information of the current cut-out web page from the position information of the image, normalizing it to match the paste-destination web page, and temporarily embedding it in the paste-destination web page currently accessed, There are some which mash up the functions of a plurality of Web pages as they are.

  However, the above mashup device assumes that the clipping range is valid (the clipping range has all the information necessary to execute the function expected by the user). In view of this, in this embodiment, the cutout specified so as to include all information necessary for executing the function suggested by the cutout range (the function expected by the user from the cutout range) with respect to the cutout range specified by the user. It extends the range.

  Below, based on the cutting range designated by the user, a method for extending the cutting range so as to include all information necessary for executing the function suggested by the cutting range will be described.

(How to extend the cutting range)
Here, a case where a Web page search window having a search function is pasted on another Web page will be described as an example. When providing a search function as a service on a Web page, an input form is often used. To display the input form, <form> tag is used and form element (<form> ~ </ form>) becomes one form. Form parts such as an input element, a select element, and a textarea element are arranged in the form element.

  For example, the input element displays each form part such as a text input in the form and an execution button. Forms usually have a single submit (execute) button. After a value is entered for each input component, the submit button is pressed, and the action attribute is determined by the transfer method specified in the method attribute of the <form> tag. Call the action specified in.

  In other words, since the input element is a part of the form element, even if <input> ~ </ input> is specified as the cutting range, it does not make sense unless it is cut in the entire form of <form> ~ </ form>. It will be. Like the input element, the select element and the textarea element are parts of the form element, so if the entire form of <form> ~ </ form> is not cut, it will not make sense.

  There are other examples, <param> tag is <applet> tag (tag for inserting Java applet in HTML document) and <object> tag (image, video, audio, various plug-in data, This tag specifies the parameters required when data is executed in a Java applet, other general-purpose tags for embedding data in various formats such as HTML documents, etc. If you don't cut the entire </ applet> or <object> ~ </ object> node, it won't make sense.

  Therefore, in the following, the <input>, <select>, <textarea>, and <param> tags are referred to as component tags. <form>, <applet>, and <object> tags are called configuration tags. The relationship between the component tag and the component tag is held as a component / component tag list table as shown in FIG. FIG. 5 is a configuration diagram of an example of a component / configuration tag list table.

  The part / configuration tag list table of FIG. 5 indicates that the <input>, <select>, and <textarea> tags are part tags of the <form> tag. The part / configuration tag list table of FIG. 5 indicates that the <param> tag is a part tag of <applet> or <object> tag.

  The part / configuration tag list table of FIG. 5 indicates that the <form> tag is a configuration tag of <input>, <select>, and <textarea> tags. The part / configuration tag list table of FIG. 5 indicates that <applet> and <object> tags are configuration tags of <param> tags.

  FIG. 6 is an explanatory diagram showing a first method of extending the cutout range. FIG. 6 expands the cut-out range by finding a configuration node for the part designation. The unresolved part confirmation unit 12 is given a cut-out range (HTML fragment or document tree subtree).

  For example, the unresolved part confirmation unit 12 is provided with the search window indicated by the red frame 15 in FIG. FIG. 7 is an image diagram of an example of a cut-out Web page. The Web page in FIG. 7 is shown in HTML, and its document tree is as shown in FIG.

  FIG. 8 is a configuration diagram of an example of a document tree. The search window indicated by the red frame 15 given as the cutout range is an input element corresponding to the first input node 20 of the p element in the substructure of the form element. The unresolved part confirmation unit 12 is given a cutting range. Specifically, the cut range is given as a node of the document tree. A node on the tree structure is considered to indicate a subtree having the node as a root.

  That is, the input node 20 in FIG. 8 indicates input elements of <input> to </ input> on HTML. In the unresolved parts confirmation unit 12, a part element such as input that is not a subordinate structure of a component such as a form element (part tag on the tree structure) To determine whether it has a node labeled with

  For example, the input node 20 in FIG. 8 is itself a component element. Therefore, the unresolved component confirmation unit 12 determines that there is an unresolved component element because the input node 20 is an unresolved component element.

  When the span node 21 that is a sibling of the input node 20 is given, the span node 21 itself is not a component element, and there is only an a element even if the sub structure in the tree structure is traced from the span node 21. Therefore, the unresolved component confirmation unit 12 determines that there is no unresolved component element because the a element is not a component element.

  When the p node 22 which is the parent of the input node 20 is given, the p node 22 itself is not a component element. However, when a substructure in a tree structure is traced from the p node 22, there is no component such as form. An input node 20 that is a component element appears. Therefore, the unresolved component confirmation unit 12 determines that there is an unresolved component element. When the ancestor form node 23 of the p node 22 is given, the form node 23 itself is a constituent element. Therefore, the unresolved component confirmation unit 12 determines that there is no unresolved component element.

  When a body node 24 that is an ancestor of the form node 23 is given, even if a subordinate structure in a tree structure is traced from the body node 24, a component element does not appear without a component. Therefore, the unresolved component confirmation unit 12 determines that there is no unresolved component element. In FIG. 8, a form node 23 appears below the div node 25 below the body node 24. Therefore, the unresolved component confirmation unit 12 does not go to the input node 20 that is lower than the form node 23.

  As described above, the unresolved component confirmation unit 12 confirms whether or not there is an unresolved component element in a given range (node). If there is no unresolved component element, the unresolved component confirmation unit 12 returns the given range (node) as it is. When there is an unresolved part element, the unresolved part confirmation unit 12 calls the related configuration acquisition unit 13.

  The related configuration acquisition unit 13 traces the upper structure in the tree structure with respect to the given range (node) and finds a component. When the configuration element is found, the related configuration acquisition unit 13 returns the found configuration element (subtree having the configuration node as a root) as a normalized range (normalized cut range).

  In the case of the input node 20, a component element is found at the form node 23 as the tree structure is traced. Therefore, the form element (subtree having the form node 23 as a root node) is returned as a normalized cut range. It is. In the case of the p node 22 as well, when the upper structure in the tree structure is traced, the component is found at the form node 23, so the form element is returned as the normalized cutout range.

  In the above, when a part element subordinate structure (for example, a character string described in a button) is specified as a cut-out range, the substructure is not regarded as a component element, but is regarded as a part of the component element. It is also possible. In that case, the unresolved parts confirmation unit 12 traces the lower structure to search for the component elements, and traces the upper structure to confirm the existence of the part elements that are not in the higher structure than the constituent elements.

  In the first method of extending the cut range, even if only a component element is selected due to a low literacy person or a user interface restriction problem that gives the correspondence between the HTML source and the browser screen, A component that uses the component element is selected as a normalized cutout range (a mashup target).

  Thereby, in the first method for extending the cut range, a function (a function expected by the user) as generally intended is saved as compared with a case where a simply cut portion (cut range) is pasted. Greatly increase the likelihood of being Further, in the first method for extending the cutout range, the increase in the pasted portion (normalized cutout range) is minimal, and there is little possibility of causing side effects.

  However, if the function related to the cutting range is not realized by the local cutting, the function as generally intended cannot be stored only by the first method of extending the cutting range.

Next, a second method for extending the cutout range will be described. For example, it is assumed that the cutting range is the following element (HTML description).
<form id = "fm1" onsubmit = "return f ()">
<input type = "text" name = "q" value = "" size = "35">
<input type = "submit" value = "search">
</ form>

  The onsubmit attribute describes a script that is activated when the send button is pressed. In the above HTML description, when “search” is pressed, the function f () is called. In the above element, the action is not completed with only the form element (<form> to </ form>), and the function f () must also be incorporated.

  The function f () should be described directly somewhere in the Web page or in a script taken from the outside. Therefore, the second method of extending the cut range is to capture all of the script description of the Web page and all the description of reading the script from the outside, and all of the script description of the Web page to the pasted Web page. And the function f () can be executed by reflecting all the descriptions that read the script from the outside.

  FIG. 9 is an explanatory diagram showing a second method of extending the cutout range. Fig. 9 shows the functions that are generally intended by saving all the actions in the Web page and expanding the cutting range when the functions related to the cutting range are not realized by local cutting. To do.

  The action acquisition unit 32 acquires an action element that may include a function or variable definition called from the cut range (or a function or variable definition called repeatedly from the called function). The action element is determined by an element name (tag) in the action element table. FIG. 10 is a configuration diagram of an example of the action element table. The action acquisition unit 32 receives the document tree of the cut-out Web page as an input, and cuts out nodes (action elements) including actions for each node of the document tree.

  A typical action element is a script element. The <script> tag describes a script such as JavaScript. A script may be directly written in <script> ~ </ script>, or <script type = "text / javascript" src = "xxx. It may be described in an external file called by src attribute like js "/>.

Action elements include form elements in addition to script elements. For example, suppose it is the following form element.
<form id = "f1" name = "f1" method = "post" action = "">
<input name = "q" type = "hidden" value = "">
<input name = "r" type = "hidden" value = "s">
</ form>

  In this form element, all input elements are hidden, do not require any input from the user, and have an auxiliary function of other action elements. Therefore, it is necessary to cut out the form element as described above as an action element.

  The action element management unit 33 converts the extracted action element into an appropriate format and holds it in the action element set table 31. For example, the action element can be stored in its description itself, or the position of the action element may be stored in a format such as XPath. When the action element set output unit 34 has obtained the action elements from the entire document tree, the action element set output unit 34 outputs the action element set to be added as an extension of the cut position from the action element set table 31.

  In the second method of extending the cut range, all actions in the Web page are cut, so that a function that is not realized by local cut can be saved. In addition, the second method of extending the cut range extends the cut range not by the entire cut-out web page but by action, so that the entire cut-out web page is taken out and pasted to the paste-destination web page. The cut-out range is limited, and capturing of styles and the like can be controlled.

  However, since the second method of extending the cut range cuts out all parts of the action, side effects such as display of standard information such as advertisements may remain. Next, a third method for extending the cutout range will be described.

The script describes the definition of a function that is executed when called from an instruction or event handler. However, the following script is an example with only a command and no function definition.
<div>
<script type = "text / javascript">
document.write ("Hello")
</ script>
world
</ div>

  The above script is evaluated and executed when it is read. Therefore, the execution result of the above script (document tree after executing the script) is Hello world.

  In this way, if there is no function definition in the script, it is an independent thing that is not referenced by other actions, so there is no need to add it as a cutting range. In addition, if there is no function definition in the script, leaving it in reverse may cause side effects such as writing extra text that is not originally cut out like the script above or displaying advertisements. unknown.

  Therefore, action elements that do not seem to be related to the realization of functions in the original cut range, such as scripts without function definitions as described above, should not be put in the cut list.

Other examples are shown below. For example, assume that the following form element is extracted as an action.
<body>
<form id = "f1" name = "f1" method = "post" action = "">
<input name = "q" type = "hidden" value = "">
<input name = "ie" type = "hidden" value = "UTF-8">
<input name = "ref" type = "hidden" value = "sa">
</ form>
<h1> Header </ h1>
...

  In this form element, the type of all inputs is hidden. This means that no input windows or buttons are displayed on the browser. This form element exists independently before h1 (not inside any block). From this, it can be determined that this form element is likely to be attached to another form element (called from another script) (roughly independent) rather than used independently.

On the other hand, assume that the following form element is cut out as an action.
<form action = "http: //xxx/search.js" name = "search" method = "get">
<B> Web site search </ B>
<input type = "text" size = "28" name = "search"><input type = "submit" value = "search">
<input type = "hidden" name = "target" value = "combined">
<input type = "hidden" name = "look" value = "yyy">
</ form>

  In this form element, the type of some inputs is hidden, but there are type = "text" and type = "submit", and it is necessary to execute some processing such as search for normal input It can be judged that the possibility of the form is high (independence is high). Some of these highly independent actions have no side effects unless they are submitted, but they are also used to output typical information such as advertisements.

  FIG. 11 is an explanatory diagram showing a third method for extending the cutout range. FIG. 11 shows that when the function related to the cut range is not realized by local cut, the cut range is expanded by narrowing down the action based on the relationship between the cut range and the action from all the actions in the Web page. It preserves functionality as generally intended.

  The third method for extending the cut range is to calculate the relevance of each action element with the action elements in the cut range (the higher the degree of independence of the action elements, the lower the relevance). Sift whether to leave as action element set. FIG. 11 is obtained by adding an action element sorting selection unit 35 to the configuration of FIG.

  The action element sorting and selection unit 35 calculates the degree of independence of the given action element and the degree of association with the action in the cut-out range from the program form, location, variables, and the like. The action element sorting and selection unit 35 leaves only the action elements that are recognized as having a low degree of independence and a high degree of association with the cut-out range (for example, when the value is greater than a certain threshold). Details of the processing of the action element sorting and selection unit 35 that leaves only action elements that are recognized as having a low degree of independence and a high degree of association with the cut-out range will be described later.

  The action element sorting and selection unit 35 may automatically determine an action element having a low degree of independence and a high degree of association with the cut-out range, or a person may give some instruction.

  In the third method of extending the cut range, the relationship with the action elements in the cut range is calculated for each action element in the Web page, and the portion to be left as the cut range is narrowed down. The possibility of (execution of an unnecessary action such as independent advertisement display independent of the cut-out range) can be greatly reduced.

  However, since the action relevance calculation by the third method for extending the cut-out range is based on heuristics, there may be cases where side effects remain or the action elements that were originally required are screened off. Next, a fourth method for extending the cutout range will be described.

For example, if you cut the following script
<form style = "margin: 0px;" id = "fm1" name = "fm1" onSubmit = "return goF2 ('fm1')">
<input name = "q" autocomplete = "off" value = "" size = "35" class = "j12" accesskey = "e" type = "text"/>
<input value = "search" class = "j12" onclick = "return goF1 ('fm1')" type = "button"/>
<input value = "Clear" onclick = "javascript: wordClear ('fm1')" class = "j12" type = "button"/>
</ form>

  In this script, the functions called goF2 and goF1 are called. Therefore, in order to narrow down the action elements more accurately, it is only necessary to extract the action element set narrowed down to the action elements in which the function being called is defined. However, since the function or variable described in another action element may actually be called from the function definition, the action to be taken in while managing the unresolved identifier (function or variable) It will be necessary to determine the elements.

  FIG. 12 is an explanatory diagram showing a fourth method for extending the cutout range. FIG. 12 shows that when the function related to the cutout range is not realized by local cutout, the action is analyzed from all the actions in the Web page and the cutout range is expanded by narrowing down the action more accurately. It preserves the functions as intended. 12 is obtained by adding an identifier reference definition function acquisition unit 36, an identifier definition source action element acquisition unit 37, and an identifier management unit 38 to the configuration of FIG.

  The identifier management unit 38 manages a resolved identifier (defined in the captured action element) (function or variable) and an unresolved identifier (not yet defined only in the captured action element). is there.

  The identifier reference definition relationship acquisition unit 36 first captures a script for a given action element. The script to capture is itself if the script is described inside an action element. In addition, when an external file is referred to, the script to be acquired is the script described in the external file by acquiring the external file. The identifier reference definition relationship acquisition unit 36 analyzes the fetched script and analyzes what identifier is defined in the script and what external identifier is referenced.

  The identifier definition source action element acquisition unit 37 uses an analysis result of the identifier reference definition relationship acquisition unit 36 for an unresolved identifier to identify an action element for resolving the identifier.

  In actual operation, for example, first, an unresolved and resolved identifier for an action element within the cutout range is given to the identifier management unit 38. The identifier reference definition relationship acquisition unit 36 analyzes in advance which identifier is defined by which action element and which identifier is referenced for all action elements, and holds a list thereof.

  The identifier definition source action element acquisition unit 37 extracts an unresolved identifier from the identifier management unit 38, refers to the analysis result of the identifier reference definition relationship acquisition unit 36, and identifies an action element for resolving the identifier. The identifier definition source action element acquisition unit 37 notifies the action element management unit 33 of the identified action element, and information on the identifier defined by the taken action element, information on the identifier that is externally referenced, and Is sent to the identifier management unit 38. Then, the identifier definition source action element acquisition unit 37 repeats such an operation until there are no unresolved identifiers, thereby specifying a set of action elements for which all identifiers have been resolved. Details of the processes of the identifier reference definition function acquisition unit 36, the identifier definition source action element acquisition unit 37, and the identifier management unit 38 will be described later.

(Details of this example)
FIG. 13 is a configuration diagram of an embodiment of a system including the mashup device of the present embodiment. The system in FIG. 13 includes a web mashup setting device 100, a web mashup realization device 110, and an annotation management server 120.

  The web mashup setting apparatus 100 allows the user to specify a cutout part and a pasting destination. The web mashup realizing apparatus 110 displays the mashup result based on the cut portion and the pasting destination designated by the user. The annotation management server 120 manages data for mashup. The web mashup setting device 100 and the web mashup realization device 110 are implemented as extended functions of the web browsing devices 101 and 111. Although the web mashup setting device 100 and the web mashup realization device 110 are separated in FIG. 13, they may be realized simultaneously as an extended function of one web browsing device.

  The web mashup setting apparatus 100 includes a web browsing device 101, a cutting source selection unit 102, a range presentation selection unit 103, a paste destination selection unit 104, and an annotation management unit 105.

  The range presentation selection unit 103 presents a user's specific operation in the display of the Web page by the normal Web browsing device 101, thereby presenting the cutting range and the paste destination position candidate by the user's keyboard or mouse operation. . The range presentation selection unit 103 returns a position on the tree structure corresponding to the currently displayed range by a designated operation (for example, ctrl-c or ctrl-v).

  The cut source selection unit 102 starts presenting and acquiring cut candidates by a specific operation of the user, and cuts the designated range (actually acquires the cut position). The pasting destination selection unit 104 starts presenting and obtaining pasting candidates by a user's specific operation (for example, cutting end or specific key operation), and determines the instructed position as a pasting destination (actually pasting) The cut-out information (URI of the cut Web page and the XPath of the cut-out portion) and the paste information (the paste-destination URI that is the URI of the Web page to be pasted and the XPath of the paste position) are obtained. A set is passed to the annotation management unit 105.

  The annotation management unit 105 requests the annotation management server 120 to register annotation information that is a set of the given cut information and pasting information. It should be noted that during the cutting and pasting, a movement between general Web pages or movement within the Web page by the Web browsing device 101 is performed.

  When it is assumed that operations such as general movement between Web pages by the Web browsing device 101 and movement within the Web page are performed as a series of operations, the cut information may be temporarily stored in the memory. Although it is good, the cutting information may be written to an external file at the end of the cutting operation and the appropriate cutting information may be taken out from the external file at the start of the pasting operation.

  The annotation management server 120 includes an annotation database management unit 121 and an annotation database 122. In response to a request from the web mashup setting device 100, the annotation database management unit 121 registers annotation information in which the cut information and paste information are set as a set in the annotation database 122.

  Also, the annotation database management unit 121 receives a request from the Web mashup realization apparatus 110 and returns annotation information having the specified URI as a pasting destination URI. Note that the annotation management server 120 may be realized on the same computer as the Web mashup setting device 100 and the Web mashup realizing device 110, or may be realized on another computer connected via a network. Absent.

  The web mashup realizing apparatus 110 includes a web browsing device 111, a cutout information acquisition unit 112, a normalization unit 113, a pasting position acquisition unit 114, a synthesis unit 115, and an annotation management unit 116.

  The web browsing device 111 takes in the content (web page) of the designated URI from the network. At this time, the annotation management unit 116 requests the annotation management server 120 to receive the annotation information given to the designated URI. For each piece of received annotation information, the cutout information acquisition unit 112 passes the URI of the cut source described in the annotation information to the Web browsing device 111, imports the content of the cut source URI from the network, and cuts out the cutout described in the annotation information. Cut out the original XPath part (HTML fragment).

  The normalizing unit 113 normalizes the cut-out HTML fragment so as to be compatible with the environment of the Web page to be pasted. The pasting position acquisition unit 114 obtains a pasting position for pasting the HTML fragment from the pasting XPath described in the annotation information.

  The combining unit 115 pastes the normalized HTML fragment at the pasting position obtained by the pasting position acquisition unit 114. Note that the DOM tree of the Web page on which the normalized HTML fragment is pasted by the synthesis unit 115 is rendered and displayed by a rendering engine, as will be described later.

  The web mashup setting device 100, the web mashup realizing device 110, and the annotation management server 120 in FIG. 13 can be realized by a computer system shown in FIG. 14, for example.

  FIG. 14 is a configuration diagram of an example of a computer system. The computer system includes an input device 151, an output device 152, a drive device 153, an auxiliary storage device 154, a main storage device 155, an arithmetic processing device 156, and an interface device 157 that are mutually connected by a bus B.

  The input device 151 is a keyboard or a mouse. The input device 151 is used for inputting various signals. The output device 152 is a display device or the like. The output device 152 is used to display various windows and data. The interface device 157 is a modem, a LAN card, or the like, and is used for connecting to a network.

  The mashup program of this embodiment is included in various programs for controlling the web mashup setting device 100 and the web mashup realization device 110. The mashup program is provided, for example, by distributing the recording medium 158 or downloading from a network. The recording medium 158 on which the mashup program is recorded is information such as a CD-ROM, a flexible disk, a magneto-optical disk, etc., a recording medium for recording information optically, electrically or magnetically, a ROM, a flash memory, etc. Various types of recording media, such as a semiconductor memory that electrically records data, can be used.

  When the recording medium 158 on which the mashup program is recorded is set in the drive device 153, the mashup program is installed from the recording medium 158 to the auxiliary storage device 154 via the drive device 153.

  The mashup program downloaded from the network is installed in the auxiliary storage device 154 via the interface device 157. The auxiliary storage device 154 stores the installed mashup program and also stores necessary files, data, and the like.

  The main storage device 155 reads out and stores the mashup program from the auxiliary storage device 154 when the computer is activated. The arithmetic processing unit 156 implements various processes described later in accordance with a mashup program stored in the main storage unit 155.

(Example of mashup program applied to a web browser)
FIG. 15 is a configuration diagram of an example of a Web browsing device to which the mashup program of the present embodiment is applied. The web browsing device 160 shown in FIG. 15 shows a typical configuration, and details such as CSS processing and PDF plug-in that are unnecessary for the description are omitted.

  The HTML parser 162 analyzes the HTML document 161 received via the network, and generates a tree structure (DOM tree) 163. At the same time, the Javascript engine 165 executes the Javascript program 164 in the HTML document 161 and rewrites the contents of the DOM tree 163. The rewritten DOM tree 163 is sent to the rendering engine 166.

  The rendering engine 166 converts the received DOM tree 163 into a form that can be seen on the screen. Then, the JavaScript engine 165 rewrites the DOM tree 163 and changes the content to be displayed according to the Javascript program 164 by a user operation or the like via the user interface 167. The mashup program 168 is typically held in the web browser 160 (ie, implemented in a form incorporated into the web browser 160 or added as a function extension).

  Then, in addition to the Javascript program 164 of the read HTML document 161 being executed, the mashup program 168 is executed, and the DOM tree 163 is rewritten.

  Note that data such as an annotation database referred to by the mashup program 168 is held outside the web browsing device 160. As another embodiment for holding the mashup program 168, it is conceivable to hold the mashup program 168 in the HTML document 161 to be read. In this embodiment, applicable Web pages are limited, but the Web browser 160 can have the same effect as holding the mashup program 168.

  Hereinafter, processing of the web mashup setting device 100 and the web mashup realization device 110 in the system of FIG. 13 will be described.

(Web mashup setting device 100)
A typical web browsing device 101 incorporating the web mashup setting device 100 includes a cut-out source selection unit 102, a range presentation selection unit 103, a paste destination selection unit 104, and an annotation management unit 105.

  FIG. 16 is a flowchart showing the processing of the cutting source selection unit. The cut source selection unit 102 is activated by a command (key operation, menu, button, or the like) embedded in the web browsing device 101.

  Proceeding to step S21, the activated cut-off source selection unit 102 activates the range presentation selection unit 103 with the mode cut. In step S 22, the cut source selection unit 102 receives a cut position node (XPath) from the range presentation selection unit 103. In step S23, the cut source selection unit 102 temporarily stores the cut position node (XPath) received from the range presentation selection unit 103 as cut information, and the process ends.

  FIG. 17 is a flowchart showing the processing of the paste destination selection unit. The paste destination selection unit 104 is activated by the end of the cut source selection unit 102 or by a command (key operation, menu, button, or the like) embedded in the web browsing device 101.

  In step S31, the started paste destination selecting unit 104 starts the range presentation selecting unit 103 with the mode being pasted. In step S 32, the paste destination selection unit 104 receives a paste position node (XPath) from the range presentation selection unit 103.

  In step S 33, the paste destination selection unit 104 calls the annotation management unit 105 and pastes the temporarily stored cut information and the paste position node (XPath) received from the range presentation selection unit 103. A set of information is registered in the annotation database 122.

  FIG. 18 is a flowchart showing the processing of the range presentation selection unit. The range presentation selection unit 103 returns the cut position or paste position node (XPath) to the cut source selection unit 105 or the paste destination selection unit 104 according to a user action or the like. Specifically, the range presentation selection unit 103 selects a node on the tree structure that is currently focused on, for example, by moving the mouse cursor, and determines the node on the tree structure by clicking the mouse.

  In step S41, the range presentation selection unit 103 receives a mode indicating whether the present is a range presentation for cutting work or a range presentation for pasting work. For example, the range presentation selection unit 103 accepts a cut mode indicating that the range is presented for cut work from the cut source selection unit 102. Further, the range presentation selection unit 103 receives a pasting mode indicating that it is a range presentation for pasting work from the pasting destination selection unit 104. The range presentation selection unit 103 activates the accepted mode.

  In step S42, the range presentation selection unit 103 obtains the position of the mouse cursor on the screen. In step S43, the range presentation selection unit 103 obtains a position on the tree structure of the Web page corresponding to the position of the mouse cursor on the screen. In step S44, the range presentation selection unit 103 obtains an appropriate node (element) as the current node from the position on the tree structure of the Web page corresponding to the position of the mouse cursor on the screen. For example, if the range presentation selection unit 103 is a character or an inter-character (#text node), the upper node (p node, td node, h1 node, etc.) is the current node.

  In step S45, the range presentation selection unit 103 displays a frame surrounding the current node according to the mode. The range presentation selection unit 103 changes, for example, the color of the cutting work frame and the color of the pasting work frame. Further, the range presentation selection unit 103 enlarges or reduces the frame according to the movement of the mouse cursor.

  In step S46, if there is an interruption due to a preset action (such as clicking a mouse button or a specified key input), the range presentation selection unit 103 displays a node (frame is displayed) at that time. Range) is returned to the cutting source selection unit 102 or the pasting destination selection unit 104. If there is no interruption by a preset action, the range presentation selection unit 103 returns to step S42 and continues the process.

(Web Mashup Realization Device 110)
A Web browsing device (or Web client in general) 111 incorporating the mashup program 168 of this embodiment performs an operation as shown in the flowchart of FIG. FIG. 19 is a flowchart showing processing of the Web mashup realizing apparatus.

  In FIG. 19, in step S51, the Web browsing device 111 receives the URI of the Web page to be displayed. In step S52, the web browsing device 111 reads the web page specified by the received URI from the network. In step S53, the web browsing device 111 analyzes the read web page with a parser and extracts the structure. The extracted structure takes the form of a tree structure (tree) (for example, a DOM tree).

  In step S54, the annotation management unit 116 accesses the annotation management server 120 using the received URI as a key, and acquires all annotation information in which the received URI is specified as the paste destination. Then, the web mashup realization apparatus 110 performs the following processing for each annotation information acquired in step S54.

  First, the process proceeds to step S55, where the cut-out information acquisition unit 112 acquires the cut-out URI described in the received annotation information. In step S56, the cutout information acquisition unit 112 passes the acquired URI of the cut source to the Web browsing device 111, and acquires the Web page of the cut source URI from the network. In step S 57, the web browsing device 111 analyzes the web page of the acquired URI of the cut-out source using the HTML parser 162.

  In step S58, the cutout information acquisition unit 112 obtains a cutout position from the cutout XPath described in the annotation information, and takes out an HTML fragment from the cutout position of the cutout Web page. In step S59, the normalization unit 113 normalizes the extracted HTML fragment so as to be compatible with the environment of the Web page to be pasted.

  In step S60, the paste position acquisition unit 114 determines the paste position on the tree structure to which the HTML fragment is to be pasted based on the XPath of the paste destination described in the annotation information and the tree structure extracted in step S53. To win. In step S61, the synthesis unit 115 embeds (grafts) the normalized HTML fragment at the pasting position acquired by the pasting position acquisition unit 114.

  When the processing of steps S55 to S61 for each piece of annotation information acquired in step S54 is completed, the web browsing device 111 proceeds to step S62, and passes the tree structure in which each HTML fragment of the annotation information is embedded to the rendering engine 166. Render.

(Three embodiments of the mashup method in the present embodiment)
As typical embodiments in which the mashup method in the present embodiment is incorporated into the mashup device of FIG. 13, there are first to third embodiments. The first embodiment is a method of expanding a position at the time when a position to be cut is designated. The second embodiment is a method of expanding the position when Web mashup is realized. The third embodiment is a method for preventing clipping when the minimum necessary information is not included as a user interface.

  Hereinafter, the first to third embodiments will be described.

  In the first embodiment, the cut-out source selection unit 102 of the Web mashup setting device 100 is extended. In the first embodiment, after the cut range is specified, the cut range is normalized. By giving this normalized cut range to the annotation management unit 105 as a normalized cut range, the first embodiment provides a normalized cut range regardless of the cut range specified by the user. The mashup with the function to be performed becomes possible.

  In the second embodiment, the cut-out information acquisition unit 112 of the Web mashup realizing apparatus 110 is extended. In the second embodiment, the range (before normalization) actually cut out is held as annotation information. In the second embodiment, the cutout information acquisition unit 112 of the Web mashup realization apparatus 110 provides a cutout range as annotation information. The cutout information acquisition unit 112 normalizes a given cutout range and obtains an actual cutout using the normalized range (normalized cutout range) before acquiring an HTML fragment from the cut-out Web page.

  As described above, the second embodiment differs from the first embodiment in normalization timing. In the first embodiment, normalization may be performed only once at the time of cutting. In the second embodiment, it is necessary to perform normalization every time a Web mashup that displays mashup information is realized. Instead, in the second embodiment, when a script is changed between the time of cutting and the time of pasting, the cutting range and the like are determined based on the new script.

  The third embodiment extends the range presentation selection unit 103 of the web mashup setting device 100. In the third embodiment, when presenting a range that can be cut out as a user interface, a node (cutout range) indicated internally on the tree structure and a normalized cutout range are provided at the same time.

  In the third embodiment, normally (unless a special operation is performed), only the normalized range is presented to the user so that only the normalized range can be cut. The third embodiment makes it easier to show what is happening to the user. The first to third embodiments may be combined.

(First method for extending the cutting range)
The component / configuration tag list table as shown in FIG. 5 holds the relationship between the component tag and the configuration tag. The component tag is a component for realizing a function such as search, and has no meaning by itself. The configuration tag plays a role of calling a script (program) for realizing the function by collecting the component tags.

  FIG. 20 is a flowchart illustrating an example of the processing procedure of the unresolved component confirmation unit. In step S71, the unresolved part confirmation unit 12 receives the cut range as a node of the document tree, and sets a search list including one of the nodes.

  In step S72, the unresolved component confirmation unit 12 returns no unresolved component node when the search list becomes empty. If there is a member in the search list, the unresolved parts confirmation unit 12 proceeds to step S73, extracts one member (node), and removes the extracted member from the search list.

  In step S74, the unresolved component confirmation unit 12 confirms whether or not the extracted member is a component element. Specifically, the unresolved component confirmation unit 12 confirms whether the tag name of the extracted node is registered as a component tag name in the component / configuration tag list table 11.

  If the tag name of the extracted node is registered as a component tag, the unresolved component confirmation unit 12 determines that the extracted member is a component element. If it is determined that it is a component element, the unresolved component confirmation unit 12 returns presence of an unresolved component node. If it is not determined to be a component element, the unresolved component confirmation unit 12 proceeds to step S75, and confirms whether or not the extracted member is a component as in step S74. If it determines with it being a component, the unresolved components confirmation part 12 will return to step S72.

  That is, the unresolved component confirmation unit 12 does not search for the subordinate structure of the node extracted in step S73. If it is determined that it is not a component, the unresolved parts confirmation unit 12 adds the child nodes as members of the search list if there are child nodes of the node extracted in step S73, and then returns to step S72. That is, when the component is not a component, the unresolved component confirmation unit 12 searches for the component element in the subordinate structure of the node extracted in step S73.

  FIG. 21 is a flowchart illustrating an example of an optional processing procedure of the unresolved component confirmation unit. In the flowchart of FIG. 20, when the inside of a part element such as a character string in the search window is selected, it is not a part element. However, if it is considered that an element inside the part is also a part, the unresolved part confirmation unit 12 further executes the process of the flowchart of FIG. 21 when it is determined that the part is not a part element.

  In other words, the unresolved component confirmation unit 12 proceeds to step S81 and receives the original node (cutout range). In step S82, the unresolved component confirmation unit 12 determines whether the received original node is a top node. If it is not the top node, the unresolved part confirmation unit 12 proceeds to step S83 and confirms whether or not the original node is a component. If it is not determined that the component is a component, the unsolved component confirmation unit 12 proceeds to step S84, and determines whether the original node is a component element. If not determined to be a component element, the unresolved component confirmation unit 12 proceeds to step S85, moves the original node to one higher node, and then returns to step S82.

  In step S82, if the unresolved part confirmation unit 12 is the received original node or the node moved in step S85 is the top node, the unresolved part confirmation unit 12 determines that the original node is not a part element part. In step S83, if the received original node or the node moved in step S85 is a constituent element, the unresolved part confirmation unit 12 determines that the original node is not a part element part.

  In step S84, if the unresolved part confirmation unit 12 is the received original node or the node moved in step S85 is a part element, the unresolved part confirmation unit 12 determines that the original node is a part of the part element.

  In the process of the flowchart of FIG. 21, the unresolved part confirmation unit 12 sequentially traces the nodes from the original node to the upper node, and if it reaches the top node without finding any component or part element, the original node becomes the part. It is determined that it is not an element part. Even when a constituent element is found first, the unresolved part confirmation unit 12 determines that the original node is not a part element part. When a component element is found first, the unresolved component confirmation unit 12 determines that the original node is inside the component element. In the present embodiment, for example, a character string between search buttons is not considered as a component element.

  FIG. 22 is a flowchart illustrating an example of the processing procedure of the related configuration acquisition unit. In step S91, the related configuration acquisition unit 13 receives the original node (cutout range). In step S92, the related configuration acquisition unit 13 moves the original node to one higher node. In step S93, the related configuration acquisition unit 13 returns to step S92 and repeats the process if the node moved in step S92 is not a component. If the node moved in step S92 is a constituent element, the related configuration acquisition unit 13 returns the node moved in step S92 as a normalized range (normalized cut range).

  In the processing of the flowchart of FIG. 22, the related configuration acquisition unit 13 is called when the unresolved component confirmation unit 12 determines that there is an unresolved component element in the cut out range. The related configuration acquisition unit 13 traces one node at a time starting from the node of the document tree that is the clipped range. If the traced node is a component, the related configuration acquisition unit 13 returns the component (a subtree having the component node as a root) as a normalized cut range.

(Second method of extending the cut range)
FIG. 23 is a flowchart illustrating an example of a processing procedure of the second method for extending the cutout range. Since the process of step S101 is the same as the first method for extending the cutout range, the description thereof is omitted.

  Proceeding to step S102, whether the action acquisition unit 32 includes a reference to an identifier such as a function or a variable defined in another action element in the structure below the range normalized in step S101 (normalized cut range). judge. Identifiers such as functions and variables are described inside the script. Therefore, the action acquisition unit 32 simply determines whether or not a script source is described.

Specifically, whether or not the reference in step S102 is included is determined according to the presence / absence of elements / attributes in the script source descriptible attribute table. FIG. 24 is a configuration diagram of an example of a script source descriptible attribute table. Note that the comments in the script source descriptible attribute table in FIG. 24 are given for explanation and are not essential.
<form>
<input type = "button" value = "Click!" onClick = "source input">
</ form>

The above form element is an example in which the script source is described as the value of the onClick attribute. The script source may be described in the form element event attributes such as onreset attribute and onsubmit attribute described in the form element shown in the script source descriptive attribute table of FIG. 24, html, head, title, etc. There are mouse event attributes such as an onclick attribute and an onmousedown attribute that can be described in various elements except some elements. The script may be written as the value of the href attribute of the a tag as follows:
<a href="JavaScript: Fill in the source"> Click here </a>

  If none of the above scripts are included in the normalized cutout range, the action acquisition unit 32 does nothing and returns the normalized cutout range as it is. When the script as described above is included in the normalized cutout range, the action acquisition unit 32 determines that there is a possibility of including a reference, and performs the following processing.

  In step S103, the action acquisition unit 32 focuses on the top of the document tree, and sequentially traces each node below the top of the document tree. The tree structure search is a well-known algorithm and will not be described. Here, it is assumed that the action acquisition unit 32 sequentially follows the nodes of the tree structure by the depth-first search.

  Proceeding to step S104, the action acquisition unit 32 determines whether or not the search is completed depending on whether or not there is a node to be visited. If there is a node to be traced, the action acquisition unit 32 proceeds to step S105 and moves the focus to the next node. In step S106, the action acquisition unit 32 determines whether the node focused in step S105 is an action element.

  Specifically, based on the action element table, the action acquisition unit 32 determines a form element constituting an input form, a script element describing a script, and the like as action elements.

  If the node focused at step S105 is an action element, the action element management unit 33 proceeds to step S107, and additionally registers the nodes below the focus focused at step S105 as action elements in the action element set. Further, since the nodes below the focused node in step S105 are registered as action elements, the action acquisition unit 32 does not need to individually search the tree structure below the focused node in step S105. Move to the last node of the node you hit.

  The action acquisition unit 32 repeats the processes of steps S104 to S107 until there are no nodes to follow. When there are no nodes to be traced, in other words, when the search of the entire document tree is completed, the action element set output unit 34 outputs the registered action element set from the action element set table 31.

(Third method for extending the cutting range)
FIG. 25 is a flowchart illustrating an example of the processing procedure of the third method for extending the cutout range. In the flowchart of FIG. 25, it is determined whether or not the degree of association between the normalized cutout range and the action element is high, and action elements having a high degree of association with the normalized cutout range are registered in the action element set.

  The processing in steps S111 to S116 is the same as the processing in steps S101 to S106 in FIG. If the node focused in step S115 is an action element, the action element sorting and selection unit 35 proceeds to step S117, and determines whether or not the degree of association between the normalized cutout range and the action element is high in the flowchart of FIG. It is determined by the processing shown.

  FIG. 26 is a flowchart illustrating an example of a processing procedure for determining the degree of association between the normalized cutout range and the action element. Proceeding to step S121, the action element sorting and selection unit 35 determines whether or not the action element focused at step S115 is a script having neither function definition nor variable definition therein.

  If the action element focused in step S115 is a script that has neither function definition nor variable definition therein, the action element sorting selection unit 35 determines that the degree of association between the normalized cutout range and the action element is low.

  If the action element focused in step S115 is not a script having no function definition or variable definition therein, the action element sorting unit 35 proceeds to step S122. In step S122, the action element sorting and selection unit 35 determines whether or not the action element focused in step S115 is a form element whose type of all inputs is hidden. If the action element focused in step S115 is a form element in which all input types are hidden, the action element sorting selection unit 35 determines that the degree of association between the normalized cutout range and the action element is high.

  If the action element focused in step S115 is not a form element in which all input types are hidden, the action element sorting / selecting unit 35 proceeds to step S123. Proceeding to step S123, the action element sorting selection unit 35 determines whether or not the action element focused at step S115 is a form element having an input that is not hidden.

  If the action element focused in step S115 is a form element having an input that is not hidden, the action element sorting selection unit 35 determines that the degree of association between the normalized cutout range and the action element is low.

  If the action element focused in step S115 is not a form element having an input that is not hidden, the action element sorting / selecting unit 35 performs the remaining processing for determining the degree of association between the normalized cutout range and the action element. Steps S121 to S123 are an example of a process for determining the degree of association between the normalized cutout range and the action element.

  Returning to step S117 of FIG. 25, if the action element sorting selection unit 35 determines that the degree of association between the normalized cut-out range and the action element is high, the process proceeds to step S118, and the nodes below the node focused in step S115 are set as action elements. Add to the action element set. In addition, since the nodes below the focused node in step S105 are registered as action elements, the action acquisition unit 32 does not need to individually search the tree structure below the focused node in step S115, so that the focus is focused in step S115. Move to the last node of the hit node.

  The action acquisition unit 32 repeats the processes of steps S114 to S118 until there are no nodes to follow. When there are no nodes to be traced, in other words, when the search of the entire document tree is completed, the action element set output unit 34 outputs the registered action element set from the action element set table 31.

  In the second method of extending the cut range, if a script description exists in the normalized cut range, a reference is included, and a set of all form elements and script elements in the document tree is set as an action element set in the normalized cut range. Added. However, in the second method of extending the cut range, unnecessary (sometimes causing side effects) action elements are also included in the normalized cut range.

  In practice, some heuristics can screen the elements that should be included in the normalized clipping range. Some simple heuristics are determined only by the action element itself. For example, a script is included, but an element in which no function definition or variable setting exists in the script or a form element having a non-hidden input is not included as an action element. A third method for extending the cropping range is to screen the elements to be included in the normalized cropping range with some heuristics.

(Fourth method for extending the cutting range)
FIG. 27 is a flowchart illustrating an example of the processing procedure of the fourth method for extending the cutout range. In step S131, the identifier management unit 38 creates a reference / definition table for action elements.

  The identifier reference definition relationship acquisition unit 36 determines, for each action element, what kind of identifier is defined in the action element and what kind of identifier is referenced, and an identifier management unit 38. To set the action element reference / definition table.

  The action element is specifically held as a position on the document tree (for example, XPath). For example, the action element reference / definition table is as shown in FIG. FIG. 28 is a configuration diagram of an example of an action element reference / definition table. Action elements are not only directly written scripts. When the action element refers to an external file, the action element takes in the external file and takes out an identifier from the script that is the content of the external file. Note that the script analysis technique is well known as a parsing technique, and therefore details are omitted.

  In step S132, the identifier reference definition relationship acquisition unit 36 extracts the identifier definition and the identifier reference from the action elements in the normalized cutout range. In step S133, the identifier management unit 38 sets an identifier management table. The identifier management table has an identifier field, a reference field, and a defined field. The identifier management unit 38 sets the identifier management table from the definition of the identifier extracted from the action element in step S132 and the reference of the identifier.

For example, assume that the normalized cutout range is the following form element.
<form style = "margin: 0px;" id = "fm1" name = "fm1" onSubmit = "return goF2 ('fm1')">
<input name = "q" autocomplete = "off" value = "" size = "35" class = "j12" accesskey = "e" type = "text"/>
<input value = "search" class = "j12" onclick = "return goF1 ('fm1')" type = "button"/>
<input value = "Clear" onclick = "javascript: wordClear ('fm1')" class = "j12" type = "button"/>
</ form>

  In the case of the above form element, the identifier management table defines goF2, goF1, and wordClear as identifiers as shown in FIG. FIG. 29 is a configuration diagram of an example of an identifier management table.

  In step S134, the identifier definition source action element acquisition unit 37 checks whether or not there is an identifier having “◯” in the reference field but not having “◯” in the defined field in the identifier management table of FIG. An identifier having “◯” in the reference column but not having “◯” in the defined column is an undefined identifier.

  If there is an undefined identifier, the identifier definition source action element acquisition unit 37 proceeds to step S135 and extracts one of the undefined identifiers. Here, goF2 is selected. In step S136, the identifier definition source action element acquisition unit 37 extracts an action element that resolves (defines) the undefined identifier (goF2) from the action element reference / definition table (in the above example, / html / head). / script [3] is retrieved).

  In step S137, the identifier definition source action element acquisition unit 37 adds the action element extracted in step S136 to the action element set. Proceeding to step S138, goF1, goF2, and goF3 are defined in the action element extracted in step S136, and checkWord is referenced, so goF1 and goF2 in the identifier management table are defined.

  Also, the identifier definition source action element acquisition unit 37 adds goF3 to the identifier management table and defines it. Further, the identifier definition source action element acquisition unit 37 adds wordClear to the identifier management table and puts “◯” in the reference column. As a result, the identifier management table is as shown in FIG.

  Proceeding to step S138, the identifier definition source action element acquisition unit 37 repeats the addition of action elements of steps S134 to S138 until there is no undefined identifier. When there are no undefined identifiers, the action element set output unit 34 outputs the registered action element set from the action element set table 31.

  The present invention is not limited to the specifically disclosed embodiments, and various modifications and changes can be made without departing from the scope of the claims. The mashup device includes the web mashup setting device 100 of FIG. 13, the web mashup setting device 100 and the web mashup realization device 110, the web mashup setting device 100, the web mashup realization device 110, and Various forms such as a form having the annotation management server 120 can be taken.

DESCRIPTION OF SYMBOLS 1 Search window part 2,3,4,15 Red frame 11 Parts / configuration tag list table 12 Unresolved parts confirmation unit 13 Related configuration acquisition unit 20-25 Node 31 Action element set table 32 Action acquisition unit 33 Action element management unit 34 action element set output unit 35 action element selection selection unit 36 identifier reference definition relation acquisition unit 37 identifier definition source action element acquisition unit 38 identifier management unit 100 Web mashup setting device 101, 111 Web browsing device 102 cut source selection unit 103 range Presentation selection unit 104 Paste destination selection unit 105, 116 Annotation management unit 110 Web mashup realization device 112 Cutout information acquisition unit 113 Normalization unit 114 Paste position acquisition unit 115 Compositing unit 120 Annotation management server 121 Notation database management unit 122 Annotation database 151 Input device 152 Output device 153 Drive device 154 Auxiliary storage device 155 Main storage device 156 Arithmetic processing device 157 Interface device 158 Recording medium 160 Web browsing device 161 HTML document 162 HTML parser 163 DOM tree 164 Javascript Program 165 Javascript engine 166 Rendering engine 167 User interface 168 Mashup program

Claims (6)

A mashup program for operating a computer for pasting a cut range cut from a cut-off Web resource to a paste-destination Web resource,
The computer,
When a node given as a cut-out range is not an unresolved part based on a table that associates a component tag that does not make sense to realize a function and a configuration tag that implements the function by combining the component tags. , An unresolved part confirmation unit for confirming whether there is an unresolved part node below the tree structure of the node;
When it is confirmed that there is the unresolved part node, a component node corresponding to the unresolved component node that is an upper node of the tree structure of the node given as the cut-out range is found, and the elements below the found component node A mashup program that operates as a related configuration acquisition unit having a normalized cutout range. Said computer further
When the normalized cut range includes a reference to an identifier of a function or variable defined in another action element, a document tree of the entire Web resource of the cut source including the normalized cut range is received, and the document tree Action acquisition means for acquiring the included action element node based on the action element table defining the action element node;
The mashup program according to claim 1, wherein the mashup program is operated as an action element set output unit that outputs the action element node acquired by the action acquisition unit as an action element node set to be added to the normalized cutout range. Said computer further
The action element acquired by the action acquisition means for the action element node having a high degree of association is determined for each action element node by determining the degree of association between the action element node acquired by the action acquisition means and the normalized cut range. The mashup program according to claim 2, wherein the mashup program is operated as an action element sorting / selecting means for narrowing down nodes. Said computer further
Identifier reference definition relationship acquisition means for retrieving an identifier defined from the action element node and an identifier to be referenced;
The action element node is defined based on an identifier defined from the action element node included in the document tree of the whole Web resource of the cut-out source extracted by the identifier reference definition relationship acquisition means and an identifier to be referenced. The identifier and the identifier to be referred to are set in the reference / definition table of the action element that associates the identifier and the identifier to be referred to, and are referred to the identifier defined from the action element node included in the normalized cutout range extracted by the identifier reference definition relationship acquisition unit An identifier management means for setting an identifier management table having an identifier field, a reference field, and a defined field based on the identifier, and managing the unresolved identifier;
An action element to be taken out from the identifier management table, the unresolved identifier that has not been defined, and to add the action element node defining the identifier from the reference / definition table of the action element to be added to the normalized cutout range The mashup program according to claim 2, wherein the mashup program is operated as an identifier definition source action element acquisition means to be added to a node set. A mashup device for pasting a portion cut from a cut-off Web resource to a paste-destination Web resource,
When a node given as a cut-out range is not an unresolved part based on a table that associates a component tag that does not make sense to realize a function and a configuration tag that implements the function by combining the component tags. , An unresolved part confirmation unit for confirming whether there is an unresolved part node below the tree structure of the node;
When it is confirmed that there is the unresolved part node, a component node corresponding to the unresolved component node that is an upper node of the tree structure of the node given as the cut-out range is found, and the elements below the found component node A mashup device having a related configuration acquisition means having a normalized cut-off range as a reference. A mashup method executed by a computer for pasting a portion cut from a cut-out web resource to a paste-destination web resource,
The computer is
When a node given as a cut-out range is not an unresolved part based on a table that associates a component tag that does not make sense to realize a function and a configuration tag that implements the function by combining the component tags. , An unresolved component confirmation step for confirming whether there is an unresolved component node below the tree structure of the node;
When it is confirmed that there is the unresolved part node, a component node corresponding to the unresolved component node that is an upper node of the tree structure of the node given as the cut-out range is found, and the elements below the found component node A mashup method for executing a related configuration acquisition step having a normalized cutout range as a reference.

Images