CN1860483A - Web content adaptation process and system - Google Patents

Abstract

An apparatus and method for adapting web page content are described. The adaptation of web page content for display on smaller intended display devices often requires the splitting of the content over a number of smaller pages. The apparatus and method relate to a procedure which integrates the process of splitting the content with applying transformations (for example, reducing the font size, images, etc) so as to optimise this process. The procedure is carried out systematically over the entire web page content, recursively splitting the content into smaller and smaller portions whilst simultaneously alternating this with various transformations so as to minimise the amount of white space visible on the smaller pages. Additionally, the preferred embodiment also tracks the transformations which have been applied to the objects and ensures consistency by applying them later to any similar objects.

Description

Web content adjustment processing and system

Technical field

The present invention relates to come web page contents is adjusted to be presented at equipment and the method on the desired display device by web page contents being divided into a plurality of less webpages.

Background technology

Sending Web content to different device is the process that described content is understood, rebuild and makes, its mode can be sent content source to different device (for example, Desktop PC, PDA and mobile phone) in the mode that is fit to the different device characteristic with understanding.(for example referring to Current Technologies for Device Independent, Mark HButler, HP Labs Technical Report HPL-2001-83, April 4 calendar year 2001) in the prior art, the present known three kinds of implementations that have:

At first, in the Web content development phase, network development person/author can use webpage development software manually to make this content to adapt to different devices.By doing like this, can generate the different editions (for example HTML/CSS, WML, XML/XSL) of single source based on device performance.This method is the raw mode that Web content is sent to different device, and a large amount of if desired version, is time-consuming and a hard row to hoe for network development person/author then.

Second kind of more automatic method is to use the code conversion method based on the agency by acting server, and when the final user had submitted the URL link to by the HTTP request, this acting server (on the fly) apace carried out adjustment work.This method is done a large amount of calculating at the acting server place, the result is that the reaction time of system is slack-off.In addition, initial network development person/author does not have to get involved the adjustment to Web content, and this can cause law and copyright problem in some countries.

The third known technology is to use by the method for adjustment based on client computer (final user's device) at client-side mounting adjustment system software.Adjustment System based on client computer is adjusted Web content apace after receiving the result who is sent it back by the requested webserver.This method is done a large amount of calculating at client-side, and this can expend the handling property with the deterioration client computer.Equally, initial network development person/author does not have to get involved the adjustment to Web content, and this can cause law and copyright problem equally.In addition, small-sized mobile device can not be used this method owing to the limitation of computing power.

As a part of adjusting, (for example, PDA and wap phone) size is less, may need one page Web content is divided into a plurality of less pages or leaves (be called as content and divide (content splitting)) in view of the client computer display device.For example, for the content of PC webpage (for example, 800 * 600 pixels) being offered less display (PDA that for example, has 240 * 320 pixels) with readable form, it is desirable to this content is adjusted into the less page or leaf that is suitable for minimum number, also attempt to make white space amount minimum simultaneously.

The existing known method that is used for the content division is included in Name is called " System, Method and Computer Program Product for Page Rendering Utilizing Transcoding " U.S. Patent application No.09/942,051 (being published as US 2003/0050931 A1)Middle those methods of describing.The document has been described and a kind ofly can have been adjusted to be presented at the system on the different display device Web content by Web content being divided into a plurality of page or leaf.This system is at first by using XML to make up the hierarchical tree structure with Web content, then with its format to be presented on the device (for example, by changing into the text font of being supported by display device, and reference variable is come replacement redundancy information).To be divided into multipage to export display device to through formative structure then.Yet although this method has realized content is divided into the purpose that is fit to the client computer display device, it does not satisfy the expectation purpose that makes the white space minimum.

In view of the foregoing, need a kind of web page contents of adjusting to be presented at the other method on the desired display device, it does not have the shortcoming of prior art, particularly attempts to make the amount minimum of the white space that shows with described content.

Summary of the invention

Be used to adjust web page contents to be presented at equipment and the method on the expectation device in order to satisfy the demand, to the invention provides.Here, provide in the mode of iteration and to divide the integrated form that combines with conversion and handle, with adjustment content display size.This allows content is divided into the less webpage of suitable quantity, and the white space amount that will show on webpage simultaneously remains minimum.In this context, should understand size that term " conversion " can comprise minimizing/increase picture/text for example, remove/replace content etc.

According to a first aspect of the invention, provide a kind of and be used to adjust web page contents to be presented at the equipment on the desired display device, comprise adjustment component, be used for described content is divided into a plurality of less webpages to be presented at described device that this adjustment component is arranged for:

(i) described content is divided into a plurality of content parts, and in the described content part at least one repeatedly repeating step (ii) to (vi);

Whether (ii) analyze described content is suitable for being presented on the described device with the size of determining described content part;

If (iii) the size of described content part is unsuitable for being presented on the described device, then this content part is carried out a plurality of conversion;

(iv) analyze through the conversion content to determine whether be suitable for being presented on the described device through the size of conversion content part; And

If (vi) the size through the conversion content part is unsuitable for being presented on the described device, then described content part further is divided into a plurality of content parts.

According to a second aspect of the invention, provide a kind of and be used to adjust web page contents, comprised by carrying out following steps described content is divided into a plurality of less webpages to be presented on the described device to be presented at the method on the desired display device:

(i) described content is divided into a plurality of content parts, and in the described content part at least one repeatedly repeating step (ii) to (vi);

Whether (ii) analyze described content is suitable for being presented on the described device with the size of determining described content part;

If (iii) the size of described content part is unsuitable for being presented on the described device, then this content part is carried out at least one content conversion;

(iv) analyze through the conversion content to determine whether be suitable for being presented on the described device through the size of conversion content part; And

If (vi) the size through the conversion content part is unsuitable for being presented on the described device, then described content part further is divided into a plurality of content parts.

According to a third aspect of the invention we, provide a kind of computer program or package, it is arranged to when being carried out by computer system, it/they make this system carry out said method.Described computer program or package may be implemented as modulated carrier signal (for example signal that transmits by network (for example, the internet)), it comprise with described computer program or package at least one corresponding data.

In addition, another aspect of the present invention also provides a kind of computer-readable recording medium, and it stores according in described computer program of the third aspect or the package at least one.This computer-readable recording medium can be magnetic, light, magneto-optic, solid-state or other storage medium that can be read by computing machine.

It will be apparent to one skilled in the art that the aspect of the present invention that proposes above being combined in any suitable array mode and describe by claims and any preferred feature that this paper and claims were proposed.

Description of drawings

With reference to accompanying drawing, from following description to the embodiments of the invention that only provide in the mode of example, further feature and advantage of the present invention will be apparent, and identical Reference numeral is represented identical part in the accompanying drawing, and in the accompanying drawings:

Fig. 1 is the parts of the expression embodiment of the invention and the system chart of the signal flow between them;

Fig. 2 is a processing flow chart, in more detail expression in operation between the parts of the embodiment of the invention information how to flow;

Fig. 3 is the process flow diagram of algorithm that is used for detecting the characteristic of Web content display object in the embodiment of the invention;

Fig. 4 is the decision tree of function that is used for detecting the display object of Web content in the embodiment of the invention;

Fig. 5 is a process flow diagram of representing in embodiments of the present invention how application content conversion; And

Fig. 6 represents how to carry out in embodiments of the present invention content to divide the process flow diagram of handling.

Embodiment

With reference to Fig. 1 to Fig. 6 embodiments of the invention are described below.

Fig. 1 is the system chart of the system that provided by the embodiment of the invention.Describe as the back, this system is made of 8 subassemblies.The whole operation of this system will be described in the back.

At first provide client capabilities to find module 12.The purpose of this module be find such as type of device final user's equipment energy characteristic and such as the device performance of the resolution of screen size/supported, processing power etc., similarly, this module can receive the information about its performance from final user's display device.Client capabilities finds that module 12 sends to determination module 14 with final user's device information.

Determination module 14 includes existing customer's machine performance profile (profile) ID that is detected in advance or be scheduled to by Adjustment System.Client capabilities profile ID is the ensemble of communication relevant with the display characteristic of display device.In use, determination module 14 at first will be based on being compared by final user's equipment energy characteristic of finding the information that module sends and performance (CC) scope and existing performance profile.If client capabilities and existing profile coupling, then the profile ID with the profile of coupling sends to content caching 10 (storing the different editions of pregenerated Web content through adjusting in content caching 10).If the CC that is received set not with existing CC commensurate in scope (that is) not with the existing CC profile of the Performance Match of the device of current request, then adjusting module 16 will be triggered.In addition, adjusting module 16 can also be by manual triggers, to generate the different editions of primitive network content under not from the situation of final user's specific request.

No matter how adjusting module 16 is triggered, after the triggering, the http head of requested Web content is checked in adjusting module 16 actions, and also action comes control content analysis module 20 from the requested Web content of Web content source storage part (store) 22 retrievals.Content analysis module 20 is moved subsequently and is analyzed the Web content that is instructed to from content source storage part 22, and to the adjusting module 16 passbacks parameter area relevant with the characteristic of this Web content, as the input value to adjusting module 16.The input parameter that adjusting module 16 receives from content analysis module 20 makes adjusting module 16 can be adjusted at the Web content of being asked in the primitive network content of storage the Web content storage part 22.Therefore, the output of adjusting module 16 is versions through adjusting of the Web content of raw requests, and its set with client capabilities (CC) information (asking the set of one or more characteristic of display of the device of this Web content) is sent to content caching 10 with suitable SGML (for example html).As previously mentioned, described display characteristic information finds that by client capabilities module 12 is definite.

Except above-mentioned module, also be provided with content caching 10 as previously mentioned, its action is with the different version through adjusting of storage primitive network content.In certain embodiments, buffer memory 10 can also be stored the client capabilities feature as the ensemble of communication relevant with the display characteristic of different client computer display device.Also provide the content source 22 that wherein stores primitive network content to be adjusted, and content sorting module 18, this module is used under the control of content analysis module 20 it being put in order before analyzing the primitive network content that receives from content source storage part 22.In addition, also be provided with customized module 24, it only is a front end system, provides the preview from the Web content through adjusting of content caching.Customized module 24 is the off-line modules that allow author's preview and further customize the content through adjusting.

The various system modules that provide by the embodiment of the invention have been provided, have been described the operation of these modules below with reference to Fig. 2 to Fig. 6 in more detail.

Equipment provided by the present invention greatly may be implemented as the computer system as webserver etc.Alternatively, this equipment may be implemented as a plurality of computer system parts (as shown in Figure 1) that can implement in discrete computer system.Preferably, one of them computer system is as the webserver of other computer system.No matter how this equipment to implement, described computer system is not only as pure server, also allows the Web content author to develop Web content and it is checked.In addition, this system also is used to generate the different editions of the primitive network content that is used for different desired display devices.

In view of these functions, there are three kinds of different operator schemes in this system:

First pattern, wherein its action is to serve the request of user to Web content, and this request receives by network;

Second pattern, wherein its action generates the version through adjusting at the Web content of different desired display devices with before receiving the request of user to the version through adjusting of this Web content; And

Three-mode is wherein carried out the adjustment of Web content, apace so that the other version through adjusting to be provided in response to user's request.The operation of each pattern of these patterns will be described now.

At first handle second pattern in the aforesaid operations pattern, imagine that for this reason this system is used for generating the different editions of primitive network content before user's request.For example, can during Website development, use this pattern, with the original source Web content is provided each at different desired display devices (all having different display characteristics) and special adjusted a plurality of versions.

Therefore, when in this pattern, pending first step is a plurality of set that generate the display characteristic profile of predetermined desired display device, and each set has unique ID and corresponding with the set of one or more display characteristic, and each characteristic is got the value of certain limit.For example, the set of the first client capabilities profile can have a plurality of fields of client type by name (client_type), screen resolution (screen_resolution) and color depth (colour_depth).The first profile IDCC1 for example can have value " PC ", have value " 800 * 600 " or " 1024 * 768 " in the screen resolution field in the client type field, and has value " 16 " in the color depth field.As another example, another client computer profile set with ID " CC2 " can have value " PDA ", have value " 200 * 300 " in the screen resolution field in the client type field, and has value " 32 " in the color depth field.It should be understood that above-mentioned only is the non-limiting example that this class can be used for the information of profile set, forms set by the various combination with equipment energy characteristic and performance and can easily generate a large amount of different profiles.In case generate, just the set of client capabilities profile be stored in the profile server 26, as shown in Figure 2.This is only used for the physically Database Systems of the profile of memory storage.Profile server 26 can be by determination module 14 visit, so that determination module can compare the characteristic of carrying out the requesting users display device and the client capabilities profile set of storage.

After a plurality of client capabilities profile set that generated the various combination that respectively has desired display equipment energy characteristic and value, in the operator scheme of current description, described system is subsequently at the version through adjusting that is stored in each client capabilities profile set generation original source Web content in the profile server 26.This realizes the original source Web content is adjusted into be complementary with each client capabilities profile set by triggering adjusting module 16.Usually trigger adjusting module 16 respectively at each client capabilities profile set, thereby when triggering for arbitrary time, with the single version through adjust of generation with corresponding this Web content of single client capabilities profile set.Be given in the detail operations of adjusting module 16 when triggering below.

The first step of being carried out by adjusting module 16 is to trigger content analysis module 20 by the details that transmits original source Web content to be adjusted to content analysis module 20.Content analysis module 20 is retrieved the original contents source from content source storage part 22 (the all-network content source that its storage is created by developer/author) then, and sends institute content retrieved source to content sorting module 18 to convert the xHTML file to.The function of content sorting module 18 is structures of arrangement SGML (Web content), and converts thereof into xHTML architecture.The xHTML form provides succinct and neat structure with the execution analysis task for content analysis module 20.Can provide content sorting module 18 by using third party software (for example, the TIDY that can obtain from http://tidy.sourceforge.net/) at priority date.Like this, here, will no longer provide the more detailed operation of content sorting module 18.Content sorting module 18 is to the xHTML file of content analysis module 20 passbacks through arrangement.

Receiving after the Web content of arrangement task below content analysis module 20 is carried out then in order:

I) total pixel of display object and pixel separately and characteristic in the computational grid content;

Ii) detect the function of each object in the Web content (they are the signs in the webpage usually).For example, object can have style (styling), structure and show label (tag);

Iii) object-based architectural characteristic [L1] (information of object tag) makes up single object; Then

Iv) mate display object display mode (pattern), and they are grouped in together with formation group (using pattern matching algorithm to realize).

These four tasks are realized by tailor-made algorithm respectively, will provide their details below.

For first task, its objective is pixel/size of calculating display object (for example, text, image etc.).The algorithm of carrying out this task will at first detect the type of display object.This algorithm is used different analysis logics to dissimilar display object then.For example, if display object is a text object, then it obtains length, font and size and based on these inputs and calculating pixel.If display object is image/applet (applet)/object, then this algorithm is with object-based width with highly calculate total pixel.For other display object, this algorithm will calculate total pixel based on the width, height and/or the wide/high attribute that are provided with in the image parameter (if it is specified) in the HTML content.Figure 3 illustrates the performed definite step of this algorithm, and will be described below.

With reference to Fig. 3,, treat by the first step that this algorithm is carried out it is to detect in each display object in the Web content of arrangement in step 3.2.In step 3.4, estimate then to determine whether the display object that is detected is text.If this evaluation result determines then that for certainly the display object that is detected is a text,, obtain the length of each text string in the display object then in step 3.6.Next in step 3.8,,, will be set at the attribute of textual indicia at the character quantity of the definite string of step 3.6 in step 3.10 for each concatenates into textual indicia.

Then, determine the font and the font of each text string,, also determine the size of each text string then in step 3.14 in step 3.12.In step 3.16, use this information to calculate the height and the width of text string, and, these height that calculate and width value are set at the respectively further attribute of the textual indicia of string in step 3.18 based on font, font and the size attribute of text string.Then in step 3.50, the processing that is used to be confirmed as the specific display object of text finishes, and begins once more with the next display object in the detection Web content in step 3.2 this processing.In case by described algorithm process all display object, then no longer repeat this algorithm.

Return step 3.4,, then carry out estimating the second time to determine whether institute's detected object is image, applet or object in step 3.20 if determine that here the display object that is detected is not a text.If this evaluation result is for affirming, be that display object is image, applet or object, then handle and proceed to step 3.22, further estimate to determine whether to have specified the width of image, applet or object in this step, if this is the case, then processing proceeds to step 3.24.If not this situation, then handle proceeding to step 3.28, determine the original width of object in this step, and aftertreatment also proceeds to step 3.24.

In step 3.24, further estimate to determine whether to have specified the height of this detected image, applet or object.If this evaluation result is then handled and is proceeded to step 3.26 for certainly.On the contrary, if the evaluation result of step 3.24 then handles proceeding to step 3.30 for negating, determine the original height of object in this step.Handle from step 3.30 then and proceed to step 3.26.

In step 3.26, will be by the width of abovementioned steps definite image, applet or object and width and the height attributes that height attributes is set at object tag in the Web content.Afterwards, the processing in this specific display object of step 3.50 finishes.As before, handle other object if desired, then begin once more to handle in step 3.2.

Return step 3.20,, then handle proceeding to step 3.32, estimate to determine whether to have specified the width and the height of institute's detected object in this step if the evaluation of being carried out here determines that the display object that is detected is not image, applet or object.If this is the case, then handling proceeding to step 3.34, is the parameter of style attribute that is used for each control sign of this display object that is detected with specified width and height setting in this step.Finish in this processing of step 3.50 then, and then can repeat as previously mentioned if desired.

If alternatively, determine both do not had specified width, which width also not have specified altitude assignment in step 3.32, then handle further to estimate in step 3.36, determined whether to specify the size of detected display object in this step.If this is the case, then handle and proceed to step 3.34, size is set at the parameter of the style attribute of each control type that is used for this object in this step.Equally, handle proceeding to step 3.50 then, finish in this step process, and if pending other display object, then can repeat this processing.

At last, if in step 3.36 evaluation result is to negate, then estimate at last determining whether to have specified the value of the display object that detected, if then specified value is set at the parameter of style attribute of each control type of this object in step 3.34 in step 3.38.On the contrary,, then handle and proceed to step 3.40, in this step, retrieve the acquiescence width and the height of each control, in step 3.34 they are set at default value then from storage part if do not specify this value.

Therefore, this algorithm is used for determining dimensional parameters through each display object of Web content of arrangement, for example the width of the length of text string or image and height.As hereinafter described, this information can be used for adjusting processing then.

For second task, another algorithm that provides as described below is to carry out this task.

At first carry out the functional category of the algorithm of second task from the single object of angle predefine of SGML.Single object (O) is the unit that embeds in the SGML, and it carries himself characteristic, for example shows style, static or dynamic and structural style.

Defined following predetermine class:

Information (I)

Message header (T)

Control (C)

Decorate (D)

Replaceable omniselector (RN)

Not replaceable omniselector (UN)

Replaceable omniselector title (RNT)

Not replaceable omniselector title (UNT)

These functional categories are defined as follows:

Information (the I)-object of information displaying contents is provided, it is important and not replaceable.This object can be text, image, video, audio frequency or any object (for example, a JAVA program) file.

The object of message header (T)-descriptor object, it can be image or the text header with information attribute.

Control (C)-the be meant object that is used for mutual purpose, for example button (send or submit to), input text district, form, drop-down menu, check box, list box etc.

Decoration (D)-not rises information function is provided, and only is used to improve the object of visual effect.This object can be image or text.

Replaceable omniselector (RN)-omniselector is the URL linked object.Replaceable omniselector be can be replaced the omniselector object replaced of text.It must be to be provided with the image that substitutes text.

Not replaceable omniselector (UN)-as described, omniselector are the URL linked objects.Therefore not replaceable omniselector is the omniselector object of cannot replaced text replacing.It must be image or the text that does not substitute text.

Replaceable omniselector title (RNT)-replaceable omniselector title is the URL linked object that information is provided of describing the omniselector object.It can be replaced by replaced text.It must be to be provided with the image that substitutes text.

Not replaceable omniselector title (UNT)-it is the URL linked object that information is provided of describing the omniselector object.It cannot be replaced by replaced text.It must be image or the text that does not substitute text.

By these predetermined function classifications are provided, algorithm can start one scan and comparison mechanism, and this mechanism analysis embeds the single Properties of Objects in the SGML (for example HTML).As shown in Figure 4, the reasoning of analysis is based on decision tree (scanning and Compare Logic order).

This algorithm begins by scan for networks content-label language from top to bottom.When scan process began, each single object of search, certification mark language also compared itself and predefined functional category.Carry out the end of this comparison process up to SGML.

In scan cycle, the single object of this algorithm search is also determined its function based on characteristic that object had.This algorithm stops relatively this single object (On) characteristic after first single object On is restricted to the specific function classification, and searches for next single object (On+1).

With reference to Fig. 4, the applied decision tree of this algorithm is handled as follows.This algorithm begins by at first searching for single object.In case detected single object 40, then algorithm just checks in the detected object whether be embedded with the hyperlink characteristic in step 4.2.

If this object has the hyperlink characteristic, then check to determine by searching any alternative text whether this object is arranged whether this object is interchangeable in step 4.4.If alternative text and title attribute (as being determined by step 4.6) are arranged, then be replaceable omniselector title (RNT) 48 with this object class for this object.And if for this object alternative text is arranged but does not have title attribute, then this algorithm is a replaceable omniselector (RN) 46 with this object class.

If do not substitute text and title attribute is arranged for this object, then this algorithm is not replaceable omniselector title (UNT) 52 with this object class.If also do not have title attribute and substitute text for this object, then this algorithm is not replaceable omniselector (URN) 50 with this object class.Estimate this difference in step 4.16.The title attribute of RNT and UNT is based on following condition:

It has title attribute; With

It is a display object; With

It must be URL hyperlink (image or a text); And

The display object that is adjacent is compared it and is had different styles.

After step 4.2,4.4,4.6 and 4.16 is with this object and the comparison of hyperlink characteristic, if object also is not classified, then the present invention will be used for the inspection logic of non-hyperlink attribute.Next step will compare the user side interaction characteristic.Determine that factor that whether single object has a user side interaction characteristic is whether single object is in following: button (send or submit to), input text district, form, drop-down menu, check box or list box, couple this result estimates in step 4.8.

Have the user side interaction characteristic if detect single object in step 4.8, then it will be classified as control (C) 42.Otherwise this algorithm will further relatively whether it be the object with video, object of classification or audio attribute in step 4.10.If then this single object will be included in information (I) functional category 44.

If this object also is not classified, then this algorithm is further checked this single object in step 4.12 by determining this single object whether to have the attribute of decoration.Determine to decorate attribute based on following standard:

The empirical value of the size of single object-decorate the size of attribute from expression is best derived the size of single object; Or

Between current single object (On) and next single object (On+1), there are symbol, line and separator (separator).

Object size (width and height) is based on empirical value (subjective value).The inventor has carried out testing experiment to 100 webpages, the result show have Pixel Dimensions wide＜=20 and height＜=image of 20 tend to be decorative object.

If current single object satisfies above-mentioned condition, then it will be classified as decoration (D) function 54.If in single object, find to decorate attribute, then in step 4.14 the present invention with further fox message title attribute.

In case determine that single object does not have the decoration attribute, then it be categorized as informational function or message header function in step 4.14.Based on following standard single object only is defined as message header (IT) 58:

It has title attribute; With

It is a display object; With

It only can be text or image; And

The display object that is adjacent is compared it and is had different styles.

If determine that this single object does not have title attribute, then it be categorized as information (I) function 56.

Therefore, as understanding from above, based on scan process and comparison mechanism by this algorithm carried out, all single objects in the Web content of arrangement have all obtained to represent its role's in SGML the specific function of being assigned, and have therefore finished second task of being carried out by content analysis module 20.

For the 3rd task, in content analysis module, provide another algorithm to carry out this task.The fundamental purpose of this algorithm be based on content positional information and with they be grouped into group (cluster)).Structural sign has been represented this information.The structural sign of discerning and selecting is:

＜TABLE 〉,＜FORM 〉,＜FRAMESET 〉,＜DIV 〉,＜UL 〉,＜OL 〉,＜DL 〉,＜P 〉,＜PRE 〉,＜ADDRESS 〉,＜BLOCKQUOTE 〉,＜Hn 〉,＜HR 〉,＜CENTER 〉,＜MENU 〉,＜DIR 〉,＜TD〉and＜NOSCRIPT 〉;

Because they can visually lump together group of objects when being presented at object on the client browser, select them that object is hived off.

The operation of carrying out the algorithm of this task is simple, and comes analyzing web content also to select the object that is used to divide into groups based on the existence of above-mentioned any sign in object.

For the 4th task, promptly mate the display object display mode and they are combined with the formation group, as following description, provide pattern matching algorithm.

Webpage can be thought and is made up of a plurality of content pieces (content chunk).These pieces are multimedia object set relevant with specific region of being concerned about or task.If basic object is defined as the object that comprises single multimedia unit (for example image or text text), and composite object is defined as the set of the object (basic object or composite object) of carrying out certain function together, and then piece self is senior composite object.When webpage being divided into a plurality of less page, for intelligibility, importantly the content piece is not decomposed.Therefore, before adjusting this content, the multimedia object that constitutes webpage need be combined into potential piece.

On the 6th the international document analysis and identification meeting (ICDAR2001) of calendar year 2001, at Yang, Y.and Zhang, among the H.J. " HTML Page Analysis Based on Visual Cues ", the poplar of Microsoft Research and open a kind of system that is used to locate these content pieces of having described.Following paragraphs outline similar system.These two systems all use HTML to indicate to carry out multimedia object are grouped into possible composite object at first, and the application model coupling is to find other possible combination afterwards.Difference between the described system is, is used for the range index and the algorithm that is used for pattern match of the similarity of definite various objects.

The initial combination of object

Before execution is grouped into possible composite object at first with multimedia object, html document is resolved to the xHTML tree to remove the HTML sign and to form the structure that is easy to handle.The xHTML tree comprises the HTML sign at node place and the multimedia object at leaf place.

Next step relates to structure group tree, and its middle period comprises multimedia object and node table is given instructions in reply and closed object (being potential content piece therefore), up to the top node of the whole webpage of expression.By at first in content with disconnect naturally relevant HTML sign (mainly be piece level sign, for example＜table,＜td,＜form,＜center and＜h) predetermined set on direct insertions＜g sign, xHTML set to change in groups sets.Secondly, with community set (for example, the width of the quantity of character or image and height in the text string) defined label (token) set together, each mark is used for a kind of multimedia object.The 3rd, start working from the multimedia object in tree middle period, mark is moved upward and remove except leaf and comprise＜g all nodes the node that indicates.When mark moved upward, it had accumulated the attribute relevant with node, if node has the child node more than, then all child nodes receive relative attribute.Ignore some for example＜tr format denotation, this is because they can't force any attribute on multimedia object, and different with the sign in the predetermined set, they do not represent new content piece usually.If＜g〉indicate that node has the child node more than, then in linear list, arrange sign to arrange identical left-to-right order with child node.

Marking (labelling) by the object (for example, table and unit) that will be associated with various level signs is potential group; This group tree has been combined with a plurality of compound objects (and content piece).This technology supposition (always incorrect)＜g〉indicate and do not divide any content piece.Yet, the content of form object is made mark can't between the content piece represented, distinguish by the repeated arrangement of similar multimedia object.Therefore, in case obtained group tree, then to belonging to each＜g〉list of the child node of node carries out pattern match.

Pattern match

First step during pattern match is handled is which mark in the list of determining in each child node is similar.It should be noted that each mark has one group of relative attribute.Each attribute comprises the right of type and value, and for example (font is 14pt) with (width, 100).Described value can be character string or integer.If attribute type does not have the value that is associated with it in essence, then this value is set at null character string, for example (black matrix).When comparing mark,, then distribute specific null attribute () to guarantee this community set non-NULL to it if specific markers does not have and its associated attributes.

In order to compare two sign α and β: (T with following attribute _i ^α, V _i ^α), i=1 ..., N ^αAnd (T _j ^β, V _j ^β), j=1 ..., N ^β, adopt following similarity index

$S(\mathrm{\α},\mathrm{\β})=\{\underset{i=1}{\overset{N^{\mathrm{\α}}}{\mathrm{\Σ}}}(T_i^{\mathrm{\α}},V_i^{\mathrm{\α}})\·(T^{\mathrm{\β}},V^{\mathrm{\β}})+\underset{j=1}{\overset{N}{\mathrm{\Σ}}}(T_j^{\mathrm{\α}},V_j^{\mathrm{\β}})\·(T^{\mathrm{\α}},V^{\mathrm{\α}})\}/(N^{\mathrm{\α}}+N^{\mathrm{\β}})$

Wherein, if i) have 1≤j≤N ^β, make $T_i^{\mathrm{\α}}=T_j^{\mathrm{\β}}$ And $V_i^{\mathrm{\α}}=V_j^{\mathrm{\β}}$ Then

$(T_i^{\mathrm{\α}},V_i^{\mathrm{\α}})\·(T^{\mathrm{\β}},V^{\mathrm{\β}})=1$

If ii) there is 1≤j≤N ^βMake $T_i^{\mathrm{\α}}=T_j^{\mathrm{\β}}$ And V _i ^αAnd V _j ^βAll be integer, then

$(T_i^{\mathrm{\α}},V_i^{\mathrm{\α}})\·(T^{\mathrm{\β}},V^{\mathrm{\β}})=\min (V_i^{\mathrm{\α}},V_j^{\mathrm{\β}})/\max (V_i^{\mathrm{\α}},V_j^{\mathrm{\β}})$

Iii) otherwise $(T_i^{\mathrm{\α}},V_i^{\mathrm{\α}})\·(T^{\mathrm{\β}},V^{\mathrm{\β}})=0$

Realize the comparison of list by dynamic time warping (dynamic programming algorithm), see table 1, wherein do not allow to aim at path (alignment path) and depart from diagonal line, and also comprise punishment non-diagonal motion more than the position of giving determined number (proportional) with the length of minimum sign.If greater than threshold value, think that then these two list are similar along the similarity index summation of aiming at the path, this be because they or identical or they length and composition aspect very little change is only arranged.

Public?Boolean?Compare(Arraylist?A，Arraylist?B)

{

float?M[][]＝new?float[A.size()+1][B.size()+1]；

Float Allow=0.55//acceptable every mark average gain

Float P=0.3//to the punishment of non-diagonal angle conversion

for(x＝1；x＜＝A.si?ze()；x++)

{

for(y＝1；y＜＝B.size()；y++)

If ((x-y)＜=2 ﹠amp; ﹠amp; (x-y)＞=-2) if // be the diagonal angle with 2

M[x][y]＝Max(M[x-1][y-1]，M[x-1][y]-P，M[x][y-1]-p)

+S(A.get(x-1)，B.get(y-1))；

}

}

if(M[A.size()][B.si?ze()]＞Min(A.size()，B.size())*Allow)

return?true；

else?return?false；

}

Table 1 is used for the JAVA code of comparison list

For pattern (pattern) is detected, at first make up the lower triangular matrix that does not contain diagonal element, which child node of detailed survey (list) is similar each other.Then by checking the remarkable marking mode of all possible mode discovery, promptly covered the repetitive sequence of similar node of the child node of maximum quantity.Significantly marking mode is represented each beginning of new group.

In order to prevent not significantly marking mode, use a plurality of constraints, that is:

This pattern must be at least two child nodes on length; With

This pattern must repeat twice at least; And

The example of model identical (instance) should not overlap.

Because these remarkable marks have been represented the beginning of group (or content piece), therefore by increasing in group that subsequently child node is guaranteed not overlap simultaneously and rational similarity between group, organizing self can be expanded.

Four tasks being carried out by content analysis module 20 have been discussed above.After these tasks, described system will make up the XML tree based on the web page contents that retrieves.This tree comprises various additional informations about Web content, and its form is suitable for sending to adjusting module now.The information that provides by this processing about:

I) inseparable group, its adjustment period between should not be separated (that is, above-mentioned " piece ");

The ii) group and the function of single object, whether its expression they can be left in the basket or remove;

The iii) total display pixel and the characteristic of content source are when determining whether/being used by adjusting module when how content being divided into less page or leaf; And

The iv) prototype structure of content source and style information.Here, structure is meant the layout of content.The code how the structural information content is arranged and located.Style be meant the width of content object (for example, text or image), highly, color, font attribute (for example, font (font-face) and size) etc.

The form of the XML tree that is generated by this processing is such, and it comprises that now where identification allow being used to be presented at the identifier that the content on the independent webpage separates (be known as＜g〉sign).Disconnecting naturally in as previously mentioned, content-based (for example, such as＜table 〉,＜td 〉,＜form〉the piece level sign that waits) and, inserted a plurality of＜g based on the suitable content group that should show together〉indicate.The hierarchical tree form is meant these＜g〉sign formed a plurality of nodes in the place that expression can separate tree, and so minimum＜g〉sign is (promptly, the most close leaf＜g sign) the such node of expression, under these nodes, do not allow content is divided (that is, the content that can not propagate can not divide into groups) on different web pages.

The additional aspect of XML tree is that it comprises second kind of identifier.These identifiers are the labels that are associated with each content object (that is, single multimedia unit, for example image or text text) or composite object (carrying out the object set of certain function together).Those similar objects represent that with the identical identifier label similar object for example is same type and has similar style attribute and the tabulation of other characteristic (for example, the quantity of picture size or text character).During adjust handling (back descriptions), these labels are used for guaranteeing processing unanimity (for example, reducing all similar picture/text frames with identical ratio) to these objects when being presented on the webpage when the processing analogical object.

As described, content analysis module 20 sends analysis result (XML tree) to adjusting module 16 now.Adjusting module 16 is then from profile server 26 all obtainable client capabilities device profiles (they are pregenerated in this operator scheme) of retrieval.Adjusting module 16 triggers circulation then, and this circular flow one algorithm is to generate the different editions of Web content based on obtainable profile.The round-robin quantity of carrying out depends on the quantity of obtainable profile.Basically, generate the adjustment version of content at each client capabilities profile.

Shown that in the process flow diagram of Fig. 5 and Fig. 6 operation is to adjust the algorithm of content on the XML tree.Fig. 5 represents to be used to check whether described content is fit to the algorithm of the current profile scope of display device.This algorithm all checks at every turn by the circulation in a plurality of stages whether described content is fit to page or leaf (that is, display device), and if be not suitable for, then it carries out a plurality of conversion to reduce the size of content step by step.The algorithm of Fig. 5 only forms one of them stage of algorithm shown in Figure 6, and it is further operated content is divided into not same page (if desired).

From the algorithm of Fig. 6, its objective is and guarantee then it to be divided into a plurality of less pages or leaves on the single page of desired display device if content is too big for showing, make page or leaf go up blank amount of space minimum simultaneously.In order to divide, described algorithm uses in the XML tree construction＜g〉sign.This algorithm is by from top＜g〉node begin by the node node of sign (have＜g 〉) described tree, move and operate.At each node place, whether the content in the subtree of described algorithm computation below node be fit to show, suitably multimedia object (for example, decrease font size size/picture size) used conversion.If at current＜g〉content of node below and be not suitable for showing, then described algorithm is to moving down into son＜g〉node, and calculate (if desired, they further being cut apart) again in them each and whole webpage is output as a plurality of less webpages up to.

In more detail, with reference to Fig. 6, the step of being carried out by described algorithm is as follows.During handling, use two interim storage parts (storehouse " Q " and " T "), with interim memory node.Use the 3rd storage part (array " Trans ") to store and the relevant data of the conversion of waiting to be applied to various objects in the webpage (for example font size or picture size dwindles).

Begin in the described adjustment algorithm of step S6.1.At first, this algorithm takes out top＜g from tree〉node and it is sent into storehouse Q (step S6.2).This algorithm is also guaranteed transit shed T be cleared (step S6.3).This algorithm moves on to step 6.4 then, this step call the function f its_page that below will describe in more detail (T is Trans), to check among the current T of being stored in＜g〉whether node be fit to the client computer display, and suitably multimedia object is carried out conversion.Yet, because this moment, storehouse T was empty, thus fits_page (T Trans) will return " true (TRUE) ", move on to step S6.5 thereby handle, to use function are_siblings (T) to check that all nodes among the T whether have identical father node in this step.Because T also is empty, therefore return this moment " very ", move on to step S6.6 thereby handle, check among the storehouse Q whether also leave any node in this step.Because node (top＜g〉node current in Q) is arranged in Q, therefore handle moving on to step S6.7, in this step node is moved on to storehouse T from the top of storehouse Q, thereby cycle of treatment is got back to step S6.4.

At this moment, top＜g of tree〉node is in storehouse T, so step S6.4 uses function f its_page, and (T Trans) checks this node (promptly, the Web content of its all subtrees of below) whether is fit to client apparatus and shows, comprise suitably object is used conversion.This step relates to the algorithm shown in Fig. 5.In this stage, described algorithm checks whether whole web page contents is fit to display device (because the node among the current T of being stored in is a top node).If this content is enough little and suitable, then owing in storehouse Q, do not have node, then described algorithm to stop to step S6.6 by step S6.5, step S6.15 is jumped in processing, exporting whole tree in this step (uses to function render (T, Trans) call), then finishes.

Yet when whole web page contents and when being not suitable for display device, step S6.4 moves on to step S6.8 for not thereby handle, and removes top node and it is stored as node N from storehouse T in this step.At step S6.9, in storehouse T, still do not have node then, therefore handle moving on to step S6.10, check in this step whether node N is leaf.In this stage, answer and be not (because node N is top＜g〉node), therefore handle and move on to step S6.11, at this step all direct son＜g with node N node is put into storehouse Q and goes up to handle.Whether at step S6.12, checking has any node that is used to handle in storehouse Q then.In this stage, all direct son＜g of top XML tree node〉node is all in storehouse Q, so cycle of treatment gets back to step S6.3, removes transit shed T in this step.

Handle through step S6.4, S6.5, S6.6 and S6.7 then and move, wherein first direct child node is moved on to storehouse T from storehouse Q.This part of Web content (that is the content in the subtree below this child node) may show (that is, step S6.4 returns " very ") to being fit to client computer for a short time.In this case, handle and pass through step S6.4, S6.5, S6.6 and S6.7 once more, so that second child node is increased on the storehouse T from storehouse Q.This processing will repeat indefinitely, and whether each all this built-up sections of the scope of examination (that is, be increased on the storehouse T all nodes) still are adapted in the client computer display.If at any point, source contents (promptly, the content of representing by the node listing among the T work as forward part) be not suitable for display (promptly, step S6.4 is for denying), then handle and move on to step S6.8, remove the final node that is increased to storehouse T (that is, the highest node on the storehouse) in this step from storehouse T, and it is stored as node N.Step S6.9 checks whether also to leave node on storehouse T then, if, then handle and move on to step S6.13, this step with node N turn back to its from storage part (that is, storehouse Q), be provided with aftertreatment.Use then to function render (T, calling Trans), with a part (that is the content of representing by the present node among the storehouse the T) output of XML tree as one in the adjusted less page or leaf.Described algorithm continues step S6.12, to determine whether also leave node to be processed in storehouse Q, if then move on to step S6.3.

After step shown in Figure 6, described algorithm will repeat work in whole tree, and whether the described content of conversion can make its suitable display (that is, at function f its_page (T, Trans) in), if be not suitable for, then further divide described content to watch.This is to realize by processing below in algorithm: when the content of each node below is not suitable for display device, just at step S6.11 direct son＜g with it respectively〉node is increased in the storage part (formation among the storehouse Q), is provided with aftertreatment.This is used for content is divided into more and more littler part, thereby can check respectively whether subtree is the suitable dimension that is used to be presented on the client apparatus to watch described content.For any content part that be fit to show, attempt to make itself and the other parts combinations (that is, step S6.7) of content then, so that the amount minimum of demonstration empty space.Yet, only be concerned about when being the brotgher of node (sibling) and (, in tree, have identical father＜g〉node) just carry out this combination of the different piece of content.Like this, described adjustment algorithm in whole tree from top＜g node is to bottom＜g the work downwards (thereunder not allowing further content to divide) of node leaf.

By means of the hierarchical tree structure of the Web content in the XML tree, and by the sequence of steps that described adjustment algorithm is carried out, when displaying contents, can keep content from the top to bottom, order from left to right.In addition, guarantee on single page, only to show whole with relevant composite object (same packets rank).

In addition, described algorithm is guaranteed the consistance that the analogical object of all less pages or leaves shows.The storage part (array that for example, is called as Trans) of its conversion tabulation by use keeping being applied to object and the object tag (that is aforementioned second kind of identification tag of expression analogical object) that is associated with this object realize.For each content part, described algorithm checks whether array Trans changes similar object to watch, thereby guarantees to be applied to the conversion unanimity of these objects.Each function f its_page () successfully uses new conversion so that its suitable webpage size is just dynamically upgraded array Trans during handling to content part.

Should be understood that in the whole description of above-mentioned adjustment algorithm, two storage parts that are used for interim memory node during handling have been described, storehouse " Q " and " T ".For the purpose of the function of described algorithm, therefore described storage part can consider to be used to keep the relevant portion with those node associated content.Yet storage part self can be in fact by the storage address tabulation of simple realization for the position of content in the another part that points to storer, and this also is suitable for.

Provide the exemplary embodiment of the adjustment algorithm (pseudo-java code) that is used to adjust webpage below:

splitpage(top_node)

{

Stack Q; //Q is the storehouse that keeps pending node

Stack T; //T is the storehouse that just keeps at processing node

Node N; //N is interim node

Arraylist Trans; //Trans keeps object/group tag application according to object in the array of the conversion of object

Q.push (top_node) // with the top＜g in the XML tree〉node is increased to storehouse Q

While(Q.size()＞0)

{

While (! T.empty ()) T.top (); // empty storehouse T

while(fits_page(T，Trans)&&?are_siblings(T))

{

If (Q.size ()==0) goto end; // XML the tree of finishing

N=Q.pop (); // take out top node from storehouse Q

T.push (N); // and it is increased to storehouse T

}

N=T.pop (); // remove top node and it is stored as node N from storehouse T

Among If (T.size ()＞0) // inspection storehouse T whether any node is arranged

{

Q.push (N); // node N is turned back to storehouse Q be provided with the back use

Trans=render (T, Trans); The a part of XML page or leaf of // output

}

else

{

If (is_leaf (N)) Trans=render (N, Trans); If // present node N is a leaf, // a part of XML page or leaf then exported

else

{

for?each?child?C?of?N?in?reverse?order

{ Q.push (C); } // with all direct son＜g of N〉node is put into and is used on the storehouse Q handling

}

}

}；

End:render (T, Trans); The decline of // output XML page or leaf

}

Wherein SplitPage () uses surface function down:

is_leaf(N)

If interim node N is a leaf, then this function returns " very ".

are_siblings(T)

If father＜g that the nodes sharing among the storehouse T is identical〉node (perhaps T comprises individual node or is sky), then this function returns " very ".

fits_page(T，Trans)

This function determines whether the node (that is the content of all subtrees below these nodes) among the storehouse T is small enough to be fit to display device.At first, if storehouse T comprise have with Trans in the node of identical identifier label, the then conversion that those objects are associated.If this moment, described content was fit to page or leaf, then (T Trans) returns " very " to fits_page.

Yet, if the also uncomfortable hinge of described content, can be to a plurality of different conversion of described content application, these are indicated among Fig. 5.If any in these additional object conversion successfully makes described content be fit to display device, then increase described conversion to array Trans with suitable object tag, and fits_page (T Trans) returns " very ".

Alternatively, if even after having used all possible additional transformation, source contents also and be not suitable for display device, then (T Trans) returns " vacation " to fits_page.

If storehouse T is empty, then this function also returns " very ".

render(N，Trans)，render(T，Trans)

This function is used to export the part (that is the content of being represented by one or more node of XML tree) of described content as one of less page or leaf through adjusting.This function has input parameter and the array Trans of node N, perhaps storehouse T and array Trans.Because Trans comprises the identifier tags of the object that so far has been provided and those conversion and the parameter that is associated with it, if therefore N or T comprise label and the identical object of label among the Trans, then to those objects application correlating transforms.Make suitably conversion of all the other objects (that is, making the display device that is fit to expectation to make the white space minimum simultaneously) then, and increase described label and conversion to Trans.At last, this function returns Trans.

As previously mentioned, call function fits_page (T, Trans) during, if described content also (for example is not suitable for the profile scope, the display device of expectation), then can below with reference to Fig. 5 they be discussed in more detail to a plurality of different conversion of described content application.Realize the following inspection of algorithm execution of this function: whether total pixel and the character of i) checking source contents can be fit to the profile scope;

Check ii) after having removed white space and line whether source contents can be fit to the profile scope; And

Check iii) whether display object is moved, adjusts size, breviary (summarising) and change characteristic can make source contents be fit to profile.

These inspections realize by eight possible conversion.Use these conversion in order, but after using each conversion,, estimate to determine whether can be presented on the described device until the content after the conversion of this point by the client capabilities of reference pin to this expectation device.If determine and to show, then no longer use further conversion, otherwise use all eight conversion.

Available conversion is as follows in order:

First conversion: font is dwindled.Here original font being transformed to " Verdana " is the less font size of type family.

Second conversion: image dwindles.Purpose is to make image object dwindle 10%, and carries out recurrence (recursion) and reach optimum dimension or 50% up to it.

The 3rd conversion: control object is dwindled.Purpose be if the result greater than the optimum dimension of object, then the ratio based on default screen size and client apparatus comes reduced objects.

The 4th conversion: remove in the space.Purpose is those unnecessary spaces of removing between the paragraph.

The 5th conversion: line is removed.Its purpose is identical with the 4th conversion.

The 6th conversion: decorative picture is removed.Purpose is based on the size of object and removes the image with decoration attribute.

The 7th conversion: remove the decoration text.Purpose is the redundant text of removing as decorating (if they are specific characters).

The 8th conversion: image is replaced.If be useful on the alternative text of image, then described algorithm will relatively substitute text size and image self.Short will be selected as adjusting the result.

Fig. 5 represents mapping algorithm in more detail, and has shown adaptable eight different conversion particularly.With reference to Fig. 5, the process that provides thus starts from step 5.1, at two counters of this step initialization.More specifically, the first counter i is initialized as i=1, the second counter r is initialized as r=0.

Then, handle proceeding to step 5.2, estimate to determine whether Web content is fit to the demonstration of desired display device in this step.Enjoy comparison by the client apparatus display performance characteristic that provides in the client capabilities profile with the characteristic of described content and profile server 26 and carry out described evaluation.In order to generate specific version, in step 5.2, always carry out described evaluation, at this profile, by the version of current example generation through adjusting of adjustment algorithm at single passenger plane family performance profile through adjusting.

If the evaluation in step 5.2 shows that Web content can be fit to the demonstration of desired display device, do not need any conversion, then mapping algorithm returns " very " in step 5.3 end and for function f its_page ().

On the contrary,, then handle and proceed to step 5.4, estimate to determine whether counter i equals 1 in this step if negate in the evaluation of step 5.2.What should look back here is, when described algorithm at first when step 5.1 begins, i is initialized as 1 with counter, so the evaluation of step 5.4 is sure, proceeds to step 5.6 thereby handle.Here, first conversion of beginning font scaled version, its form of 5.8 and 5.10 of taking steps.

In step 5.8, the font size of all texts in the Web content to be adjusted is made as 1, but can selects other value in other embodiments.Then, in step 5.10, the font printed words of all texts in the Web content are set at " Verdana ".The result of these steps is the sizes that reduced any text object in the Web content significantly.After these steps, handle and proceed to step 5.12, make counter i add 1 in this step, and aftertreatment turn back to the evaluation at step 5.2 place, wherein estimate to determine whether be fit to the demonstration of desired display device this moment through the Web content of conversion.If the evaluation result at step 5.2 place is sure, promptly this moment, Web content can be presented on the desired display device, then handled to proceed to step 5.3 and end.On the contrary, further conversion is if desired then handled and is proceeded to step 5.4, estimates counter i in this step and whether equals 1.Here,, thereby return negative result, handle the evaluation that proceeds to step 5.14 place thus, estimate counter i in this step and whether equal 2 because counter i has increased and equal 2 this moment in step 5.12.Here sure value will be returned, step 5.16 will be proceeded to thereby handle.

In step 5.16, the beginning image dwindles conversion.Dwindle in the conversion at this image, at first in step 5.18, the maximum possible that obtains image is dwindled.This is strict encoded radio, for example 50%.Then in step 5.20, estimate this maximum and dwindle value whether greater than ten times of the value of counter r.What should look back here is, is zero in step 5.1 with the value initialization of counter r, and therefore in first time during recurrence, the evaluation of step 5.20 will be returned sure value.Here handle and proceed to step 5.22, the image in the Web content is dwindled 10% in this step.Handle to proceed to step 5.24 then, make counter r add 1 in this step, and from this step to step 5.26, estimate r in this step and whether equal 5.In the recurrence first time, r only increases and value 1 only in step 5.24, so the evaluation of step 5.26 will be returned negative value.In this case, handle and directly turn back to step 5.2, estimate: whether be fit to the demonstration of desired display device through the content of conversion in this step.If be fit to, then handle in step 5.3 and finish, and if be not suitable for, then handle and proceed to step 5.14 through step 5.4, wherein because i is not increased once more, thus sure evaluation returned, thus applying step 5.18,5.20,5.22,5.24 and 5.26 image dwindle conversion once more.

As can be seen from Figure 5, can dwindle conversion by five application images, when using this conversion at every turn, for each recurrence, size of images further dwindles 10%, perhaps dwindle under ten times the situation of the currency that is not more than r, make picture size dwindle the maximum available value of dwindling image maximum available.Yet in either case, described conversion is all recursively used five times, up to counter r=5.Under counter r=5 situation, the evaluation of step 5.6 will be sure, proceed to step 5.12 and handle, and in this step counter i be increased.From step 5.12, handle and always turn back to step 5.2, estimate in this step: this moment, whether described content was fit to the demonstration of desired display device.If the conversion of having used is abundant, then sure value is returned in this evaluation, finishes in step 5.3 thereby handle.Then, if the conversion of having used and insufficient, and need further conversion, processing will and also be passed through step 5.15 (because this moment, i equaled 3) this moment through step 5.4 and proceed to step 5.30.

Here, estimate counter i and whether equal 3, if, then handle and proceed to step 5.32, begin control object and dwindle conversion by proceeding to step 5.34 in this step.

In step 5.34, the ratio of the default screen size of acquisition Web content and the actual screen size of desired display device.Based on this ratio, in step 5.36, based on described ratio by use the size that this ratio calculates each control object to default size.Then in step 5.38, whether the driven dimension of estimating each control object may size less than the minimum of each object, if be not less than, then handle and proceeds to step 5.42, can reduce the size of control object based on the ratio that is calculated in this step.Yet, if less than the permission minimum dimension of each control object, handling, the size of being calculated proceeds to step 5.40, allow minimum possibility size to reduce the size of control object in the Web content in this step based on described.Described permission is minimum may size to be predetermined in advance.

After step 5.40 or step 5.42, handle proceeding to step 5.12, in this step counter i is increased, proceed to step 5.2 afterwards, whether be fit to the demonstration of desired display device this moment through the content of conversion in this step evaluation.

The evaluation result of supposing step 5.2 negates, and counter i is the value of equaling 4 at this moment, therefore handles through step 5.4, step 5.14 and step 5.30 proceeding to step 5.44, equals 4 evaluation in this step I and returns sure value.Consequently make to handle to proceed to step 5.46, begin the space in this step and remove conversion.

This conversion relates to the object flag of watching in the Web content, and removes the object that has special sign and/or satisfy other specified conditions.Therefore, in step 5.48, remove have sign＜BR and be to have sign＜TD and＜DIV first child node and those objects of child node at last of object.Then, in step 5.50, also remove and have sign＜BR〉object and be to have sign＜Table those objects of the brotgher of node of object, then in step 5.52, any continuous white space in the Web content display object is reduced to single space, therefore in step 5.54, any continuous interruption in the Web content display object is reduced to one.At last, in step 5.56, with any＜table〉the unit filler (cell padding) of object and unit interval value be reduced to zero.The result that conversion is removed in the space is reduced to bare minimum with the white space in the Web content.

After step 5.56, handle and proceed to step 5.12, make counter i be increased to 5 in this step.Carry out the evaluation at step 5.2 place then, to determine whether to be fit to through the conversion content at this moment the demonstration of desired display device, if then handle in step 5.3 and finish.Yet if not, handle and will proceed to step 5.58 through step 5.4, step 5.14, step 5.30 and step 5.44, the result who equals 5 evaluation in this step I is sure.Afterwards in step 5.60, use line and remove conversion, its be used for removing in step 5.62 have＜HR all display object of sign.This has identical functions with the 4th conversion of using in the past (that is, reducing white space).

After step 5.62, handle and again proceed to step 5.12, increase at this step counter i.And then carry out the evaluation of step 5.2, and and suppose the result that it produce to be negated, then handling will be through step 5.4,5.14,5.30,5.44 and 5.58 and proceed to step 5.64.Because i has been increased to 6, therefore the evaluation in step 5.64 will produce value certainly.

Therefore, after step 5.64, handle proceeding to step 5.66, begin to decorate text in this step and remove conversion.This carries out at step 5.68 place, removes its function in this step from Web content and is the text of decorative purpose by content analysis module 20 detections.

After step 5.68, handle proceeding to step 5.12, increase counter i in this step, then proceed to step 5.2 and sentence whether evaluation is fit to the desired display device this moment through the conversion content demonstration.Suppose to be not suitable for, then handle and pass through step 5.4,5.14,5.30,5.44 respectively, 5.58 and 5.64 evaluation proceeds to step 5.70, here because the value of this hour counter i is 7, so returns definite results.This causes processing to proceed to step 5.72, begins decorative picture in this step and removes conversion.In step 5.74, remove those images or object that its function is detected to decorating by content analysis module 20.Therefore, removed the image that can not help the actual semantic content of Web content.

After step 5.74, handle again proceeding to step 5.12, in this step counter i is increased.Carry out at step 5.2 place afterwards and estimate: whether the content after conversion this moment is fit to the demonstration of desired display device, and suppose that it negates value that this evaluation is returned, then handle respectively through the evaluation of step 5.4, step 5.14, step 5.30, step 5.44, step 5.58, step 5.64 and step 5.70 and proceed to step 5.76, this step carry out about this moment i whether equal 8 evaluation.Because this evaluation will be returned definite results, therefore handle and proceed to step 5.78, begin image in this step and replace conversion.This starts from step 5.80, for each image display object, carries out whether having about image the evaluation of alternative text in this step.If, then handle and proceed to step 5.82, further estimate in this step, whether total Pixel Dimensions of the alternative text of evaluation map picture is less than image self.Only less than situation under, just image is replaced with alternative text.More many space if text occupies than image, then very clear usefulness substitutes text, and to replace image nonsensical.After the replacing it of step 5.84 place, handle proceeding to step 5.12.Similarly, negate value if the evaluation of step 5.80 or step 5.82 is returned, promptly image does not have alternative text, or alternative text and be not less than the conventional images size, and then processing similarly proceeding to step 5.12.Here be to be noted that before processing proceeds to step 5.12 and increases counter i, successively to the image transformation shown in each image applications Fig. 5.In addition, foregoing each conversion is used in the processing of a plurality of objects in the Web content, thereby before making the counter i increase that allows next conversion of application, each related object in Web content is used each conversion.

In step 5.12, counter i is increased once more, thus its present value 9 in this case.Therefore, when processing proceeds to the evaluation at step 5.2 place, satisfy the alternative condition of i this moment, and therefore must finish described mapping algorithm greater than 8 evaluation.

Therefore as can be seen, described mapping algorithm is used for using each conversion successively to the display object of Web content, and whether the Web content of estimating through conversion after using each conversion can be presented on the display of desired display device from top description.If then do not use other conversion, and described function returns " very " to fits_page ().

After finishing described adjustment processing, promptly after the adjustment algorithm of Fig. 6 finishes, generate the different editions of Web content at the specific client performance profile.Yet,, therefore must move described algorithm repeatedly to generate the Web content version through adjusting of each client capabilities profile because a plurality of client capabilities profiles are arranged.(that is, after generating all versions) after this, adjusting module be the version generation profile ID that generates based on the client capabilities profile, and will adjust version and ID is stored in the content caching.The logical relation of profile ID and the adjusted content of physics is the form of database structure cross reference link.These versions of adjusting content are ready for and are retrieved then, and send the final user to when request.

Therefore, in the aforesaid operations pattern, system according to the embodiment of the invention is used for generating these adjustment versions before the user asks a plurality of adjustment versions of primitive network content, each version is used to have known and concrete desired display device particular characteristics.

Yet as previously mentioned, described system can also operate with webserver identity.Below this operator scheme will be described.

Imagine that described system is as the webserver and and Internet connection.This server is asked from the http that final user's device 1 receives Web content.This http request at first is transmitted to client capabilities and finds module 12, and this module is used for determining the display characteristic set of final user's device 1, for example screen size, color depth, browser type, network connection etc.

Client capabilities discovery module comes the pick-up unit performance (to please refer to the Current Technologies for DeviceIndependent that quotes previously based on the existing standard that is for example proposed by M3I, Mark H Butler, HP Labs Technical Report HPL-2001-83 4April 2001, need its related content to understand the present invention fully, it is incorporated herein by reference).Now, most of Internet-browsers comprise final user's device information in the request that sends to the webserver at first, for example browser type and version, IP address, screen resolution etc.When the final user imports URL by web browser, final user's device will begin and server communication.In order to obtain final user's device information, client capabilities is found the simple Javascript of module 12 uses ^TMRetrieve the client capabilities information of sending, and send this information to server by Java servlet program from final user's browser.Following is Javascript ^TMThe sample of program, this program is by being called the Java of " clientprofile (client's profile) " ^Servlet obtains final user's device information and sends it to server:

<script?language="JavaScript">

function?getdeviceinfo(){

document.formclient.pageUpdata.value=document.lastModified;

document.formclient.availHeight.value=screen.availHeight;

document.formclient.availWidth.value=screen.svailWidth;

document.formcliont.bufforDopth.value=screen.bufferDopth;

document.formclient.colorDepth.value=screen.colorDepth;

document.formclient.fontSmoothingEnabled.value

screen.fontSmoothingEnabled;

document.formclient.height.value=screen.height;

document.formclient.width.value=screen.width;

document.formclient.updateInterval.value=screen.updateInterval;

document.formclient.javaEnabled.value=navigator.javaEnabled{};

document.formclient.appName.value=navigator.appName;

document.formclient.appVersion.value=navigator.appVersion;

document.formclient.cookieEnabled.value=navigator.cookieEnabled;

document.formclient.cpuClass.value=navigator.cpuClass;

document.formclient.mimeTypes.value=navigator.mimeTypes;

document.formclient.appCodeName.value=navigator.appCodeName;

document.formclient.platform.value=navigator.platform;

document.formclient.opsProfile.value=navigator.opsProfile;

document.formclient.plugins.value=navigator.plugins;

document.formclient.systemLanguage.value=navigator.systemLanguage;

document.formclient.userAgent.value=navigator.userAgent;

document.formclient.userLanguage.value=navigator.userLanguage;

document.formclient.userProfile.value=navigator.userProfile;

document.formclient.action="clientprofile";

document.formclient.submit();

}

</script>

Determined client's feature collection of final user's device 1, client capabilities finds that the feature collection that module 12 will be determined subsequently like this sends determination module 14 to, and this module is used for comparison final user equipment energy characteristic and the client capabilities profile set that is stored in profile server 26.If determination module 14 can be with the set of final user's equipment energy characteristic and one of client capabilities profile coupling, then determination module visit storage is through the content caching 10 of the different editions of adjustment content, and the profile ID of the client capabilities profile of use and final user's equipment energy characteristic coupling is as the index to it.Then from the version through adjusting of the Web content of the profile ID of the indexed client capabilities profile to coupling of content caching 10 retrieval, and by network it is supplied to final user's device 1.Therefore, in this operator scheme, described system can be complementary final user's device display characteristic and preset device feature collection, thereby the suitable version through adjusting that generates in advance of definite Web content is to send final user's device to.

As previously mentioned, described system also provides another operator scheme, and it makes up the operation that is provided by previous mode.Here, when final user's device 1 request Web content, as before, client capabilities finds that module 12 is used for determining its display characteristic, then it is sent to determination module 14.Determination module 14 is attempted the performance of final user's device 1 is complementary with the client capabilities that is stored in the profile server 26 then, if the coupling of discovery, the suitable version of retrieval network content from content caching 10 then through adjusting, and send it to final user's device 1 by network.Yet if can not mate, determination module 14 be by will finding that the details of final user's device 1 that characteristic that module 12 determines is relevant sends adjusting module 16 to client capabilities, and adjusting module 16 is operated.Adjusting module 16 generates the corresponding new client capabilities profile of performance with final user's device 1 subsequently, this profile is stored in the profile server 26, and identical mode begins its operation with the adjustment version that generates Web content in advance the time, aims at the new version through adjusting of the Web content that final user's device 1 adjusts with generation.That is, adjusting module 16 makes content analysis module 20 operate, and this content analysis module analyzing web content moves described adjustment algorithm to allow adjusting module, thereby generates the new adjustment version of the Web content that is exclusively used in final user's device 1.Then the Web content through adjusting is returned to determination module, this module is transmitted to final user's device 1 by network with it.In addition, if desired, also new Web content through adjusting is stored in the content caching 10, uses in the future for final user's device similar to final user's device 1.Therefore, in this another operator scheme, can generate the redaction of Web content, then if desired, it be stored to be used to serve user's request in the future in response to user's request dynamic ground through adjusting.

Except the aforesaid operations pattern, described system also provides customized module 24.It is a front end, only browses the various versions through adjusting of the Web content that is stored in the content caching in order to the permission web page authors, thereby if desired it is further refined or improves.Functional in view of this, will customized module 24 be discussed further.

Therefore, generally speaking, described system allows to generate them before the user asks the different editions of Web content, thereby can serve user's request by the display characteristic of final user's device and the version that generates in advance are complementary, and therefore allow to produce reaction apace with minimum calculating strength.In addition, if desired, can dynamically generate the specific final user device of the redaction of Web content, also store the Web content that dynamically generates then so that be used to serve the request in future of self similarity final user device afterwards through adjusting with the matching request Web content through adjusting.

Should understand, in the described here specific embodiment, specifically handle with reference to the adjustment of Fig. 5 and Fig. 6, described process relates to iterative processing, this processing is like this work, promptly the Web content (XML tree) by by analysis, divide this content and to its conversion (if desired) thereby to dwindle the suitable display device of its size.Yet, also it is contemplated that the embodiment of alternative, wherein described process is revised as, before dividing content, earlier whole contents is reduced to its minimum possibility size, then carry out such conversion, thus promptly once more magnification ratio be fit to display size with the size that increases content.When relaying the ratio of big content, be necessary then to repartition described content so that it is distributed between the page or leaf better.Therefore, this process also relates to repeatedly divides and conversion, but whether the size of determining content is fit to be presented at step on the device may relates to calculating and whether will show unacceptable large-area white space on device.

Therefore, described grouping/hive off pattern algorithm in groups, and the relation between definite these groups with webpage part group of objects.How described grouping and relation influence come the display web page content by the less page or leaf of the several contents on the less screen device.The method that a kind of definite group should be arranged in less page of which page or leaf of gathering of content adopts the pattern algorithm that divides into groups/hive off, and further depends on following preset parameter value:

The maximum acceptable number (total) of calculating pixel in group to be selected;

The white space accepted (can determine) for page or leaf by test through dividing;

The acceptable page or leaf vertical length (it can be the parameter that experience is determined) of dividing; And

The acceptable page or leaf horizontal length (it can be the parameter that experience is determined) of dividing.

By adopting circulation (promptly, iteration or recurrence) method will arrange with the web page contents on the device that is presented at the display area with first preliminary dimension and be divided into a plurality of content pages that are used to be presented on one or more other device, each described other device can both have different display areas, and the form that is used for displaying contents on each device is consistent relatively.Therefore, expression comprises the form that all demonstrations of the page or leaf of the content that is derived from original web page should be able to shared unanimity.

Unless context clearly needs, otherwise in whole instructions and claims, word " comprises " etc. and should be interpreted as with exclusive or limit is adversative is included; That is to say that meaning is " comprise, but be not limited to ".

Repeat the text of making a summary below with the part of book as an illustration:

The equipment and the method that are used to adjust web page contents have been described.The adjustment that is used to be presented at the web page contents on the less desired display device often need be divided into described content a plurality of less pages or leaves.Described equipment and method relate to such process, and its processing that will divide content is with mutually integrated so that should processing the best to the application of conversion (for example, decrease font size size, image etc.).On whole web page contents, systematically carry out described process, recursively described content is divided into more and more littler part, and alternately carry out various conversion with it simultaneously so that the visual white space amount minimum on the less page or leaf.In addition, preferred embodiment is also followed the tracks of the conversion of object having been used, thereby guarantees afterwards any analogical object to be used the consistance of these conversion.

Claims (27)

1, a kind ofly be used to adjust web page contents to be presented at the equipment on the desired display device, comprise adjustment component, be used for described content is divided into a plurality of less webpages to be presented at described device, this adjustment component is designed to:

(i) described content is divided into a plurality of content parts, and in the described content part at least one repeatedly repeating step (ii) to (vi);

Whether (ii) analyze described content is suitable for being presented on the described device with the size of determining described content part;

If (iii) the size of described content part is unsuitable for being presented on the described device, then use at least one content conversion to this content part;

(iv) analyze and whether be suitable for being presented on the described device with the size of determining described content part through conversion through the content of conversion; And

If (size of vi) described content part through conversion is unsuitable for being presented on the described device, then described content part further is divided into a plurality of content parts.

2, equipment according to claim 1 wherein, determines whether suitable analytical procedure comprises (ii) and (iv) whether definite described content is small enough to be presented on the described device to described size.

3, equipment according to claim 1 and 2, wherein, described adjustment component also is designed for:

Determine (iv) that in step content part through conversion is small enough to be presented under the situation on the described display device, then

This content part and another content part through conversion is made up to form the combined arrangement part.

4, equipment according to claim 3, wherein, described adjustment component also is designed for:

Analyze described content and whether be suitable for being presented on the described device, if the size of this combined arrangement part is too big for described device, then at least one content conversion of these combined arrangement certain applications with the size of determining the combined arrangement part.

The described equipment of claim 4 when 5, basis is subordinated to claim 3, wherein, described adjustment component comprises the storage part that is used for content part, and the described step that wherein makes up two content parts comprises from this storage part and select described another content part that described adjustment component also is designed to:

Analyze described content to determine whether be suitable for being presented on the described device through the combined arrangement size partly of conversion, if the combined arrangement size partly through conversion is too big for described device, then take described combined arrangement part apart so that described another content part is turned back in the described storage part.

6, equipment according to claim 5, wherein, described adjustment component also is designed to:

If the combined arrangement through conversion partly is small enough to be presented on the described device, then itself and second content are partly made up.

7, according to each described equipment in the aforementioned claim, also comprise analysis component, this analysis component is designed for described web page contents is changed into the hierarchical tree form, and this grade tree format comprises and is marked into a plurality of nodes that expression is suitable for described content is divided into the position of less webpage.

8, according to each described equipment in the aforementioned claim, wherein, described adjustment component also comprises the storage part that is used for content part, and wherein divides content and comprise to described storage part with the described step that forms less content part and increase a plurality of content parts.

9, according to each described equipment in the aforementioned claim, wherein, described adjustment component comprises:

The conversion storage part, the label that is used to store the record of the conversion that has been applied to content and has used the content type of described conversion.

The described equipment of claim 9 when 10, basis is subordinated to claim 3 wherein, also comprises as the combined arrangement step partly defined in the claim 3:

To be applied to described another content part according to the content conversion of described conversion record, thereby the content of the same type of indication is as one man used conversion in writing down with described conversion.

11, a kind ofly be used to adjust web page contents, comprise by carrying out following steps described content is divided into a plurality of less webpages to be presented on the described device to be presented at the method on the desired display device:

(i) described content is divided into a plurality of content parts, and in the described content part at least one repeatedly repeating step (ii) to (vi);

Whether (ii) analyze described content is suitable for being presented on the described device with the size of determining described content part;

If (iii) the size of described content part is unsuitable for being presented on the described device, then this content part is used at least one content conversion;

(iv) analyze through the content of conversion to determine whether be suitable for being presented on the described device through the size of the content part of conversion; And

If (vi) the size through the content part of conversion is unsuitable for being presented on the described device, then described content part further is divided into a plurality of content parts.

12, method according to claim 11 wherein, determines whether suitable analytical procedure comprises (ii) and (iv) whether definite described content is small enough to be presented on the described device to described size.

13, according to claim 11 or 12 described methods, also comprise step:

Determine (iv) that in step content part through conversion is small enough to be presented under the situation on the described display device, then

This content part and another content part through conversion is made up to form the combined arrangement part.

14, method according to claim 13 also comprises step:

Analyze described content and whether be suitable for being presented on the described device, if the size of described combined arrangement part is too big for described device, then at least one content conversion of described combined arrangement certain applications with the size of determining described combined arrangement part.

The described method of claim 14 when 15, basis is subordinated to claim 13, wherein, the step of two content parts of described combination comprises from a storage part selects described another content part, and comprises step:

Analyze described content to determine whether be suitable for being presented on the described device through the combined arrangement size partly of conversion, if the combined arrangement size partly through conversion is too big for described device, then takes described combined arrangement part apart and also described another content part is turned back in the described storage part.

16, method according to claim 15 also comprises step:

If the combined arrangement through conversion partly is small enough to be presented on the described device, then itself and second content are partly made up.

17, according to each described method in the claim 11 to 16, also comprise step:

Described web page contents is changed into the hierarchical tree form, and this grade tree format comprises and is marked into a plurality of nodes that expression is suitable for described content is divided into the position of less webpage.

18, according to each described method in the claim 11 to 17, wherein, the division content comprises to storage part with the described step that forms less content part increases a plurality of content parts.

19, according to each described method in the claim 11 to 18, wherein, also comprise:

Keep being applied to the record of conversion of content and the indication of having used the content type of described conversion.

20, the described method of claim 19 when being subordinated to claim 13, wherein, combination also comprises as the step of the content part defined in the claim 3:

To be applied to described another content part according to the content conversion of described conversion record, thereby the content of the same type of indication is as one man used conversion in writing down with described conversion.

21, a kind ofly be used to adjust the equipment that is presented at the web page contents on the device, the display of this device is little more a lot of than its ability display size of webpage, thereby the content of webpage need be divided into a plurality of pages or leaves to be presented on the described device, and this equipment comprises:

Parts are designed so that the processing of dividing content combines with using conversion, by recursively described content being divided into more and more littler part, using various conversion simultaneously and make less page or leaf go up visual white space amount minimum.

22, equipment according to claim 21, also comprise the parts that are designed to follow the tracks of the conversion that has been applied to each smaller portions, wherein this equipment also comprises such parts, and it is designed to guarantee consistance by the identical conversion of any similar certain applications to web page contents.

23, a kind ofly be used to adjust web page contents to be presented at the method on the device, the display of this device is little more a lot of than its ability display size of described webpage, thereby web page contents need be divided into a plurality of pages or leaves to be presented on the described device, this method may further comprise the steps:

Recursively described web page contents is divided into a plurality of smaller portions of first preliminary dimension;

Use conversion to described a plurality of smaller portions, the step of dividing content is combined with the step of using conversion, recursively described content is divided into more and more littler part, use various conversion simultaneously so that less page or leaf is gone up visual white space amount minimum.

24, method according to claim 23, further comprising the steps of:

Tracking be applied to various piece each conversion and

Storage about the information of all conversion of being applied to various piece and

The identical conversion of any similar certain applications to described web page contents.

25, a kind of computer program or package, it is designed to when being carried out by computer system, it/they make described system enforcement of rights require each described method in 11 to 20,23 or 24.

26, a kind of through modulated carrier signal, it comprise with described computer program of claim 25 or package at least one corresponding data.

27, a kind of computer-readable recording medium, it stores in computer program according to claim 25 or the package at least one.

Images