Processing system for use with a user interface display

[0001] This application is concurrently filed together with commonly owned related application, Ser. No. 10/100,867 filed 19 Mar. 2002.

[0002] The present application is a non-provisional application based on the provisional application Ser. No. 60/278,280 entitled “A User Interface Display Configuration and Data Processing System” and filed Mar. 23, 2001.

[0003] This invention pertains generally to a computer user interface display configuration and processing system, and more specifically to a nonprocedural method for interfacing data in a data buffer to event driven user interface controls.

[0004] The description of forms to be displayed on internet browsers via the World Wide Web is typically accomplished by the use of Dynamic Hypertext Markup Language (DHTML). DHTML is a product of the Microsoft Corporation of Redmond, Wash. DHTML incorporates the concept of the Form as a means of presenting Graphical User Interface (GUI) controls for manipulation by the user. DHTML documents contain tags representing user interface devices, such as input text boxes etc., known as controls. The controls may initially be blank, or information may be placed in the controls when the form is initially displayed. The initial data for any of the controls is usually transmitted with the DHTML document along with the control tag. An example of a DHTML statement having an inputbox control with the data “19344” embedded in the tag might be written as:
<INPUT ID=ctlZipCode>19344</INPUT>

[0005] The user, in turn, may edit or enter new information into these controls. When the user is done, the contents of the GUI controls, is collected by the Web browser and forwarded as part of the Universal Resource Locator (URL) to the Web server in the form of control name/control value pairs.

[0006] Often, such forms are used to display and collect information related to a database. The database may contain tables, each table containing a collection of records, each record containing a collection of fields. When used in this manner, it is necessary to associate a control on the form with a corresponding piece of data in the database. Data binding refers to a software subroutine which associates data contained within a database with other information that is relevant to that data, including the data's location within the form, i.e. the control used to display and/or edit that data.

[0007] Currently, data binding relies on an architecture having four components, namely a data source object (DSO), data consumers, a binding agent and a table repetition agent. In order to bind database data to control elements present on an HTML page, a DSO, which represents the database, must exist on that page. To identify desired data within the database, the DSO requires selection information in some form. For example, the DSO may require an Open Database Connectivity (OBDC) string and a Structured Query Language (SQL) statement or only a Universal Resource Locator (URL). SQL is discussed in SQL—The Complete Reference by James R. Groff and Paul N. Weinberg, McGraw-Hill Professional Publishing (1999) ISBN 007-211-8458.

[0008] The data in the DSO may also be defined using Extensible Markup Language (XML). XML is discussed in Essential XML: Beyond Markup by Don Box, Aaron Skonnard and John Lam, Addison-Wesley Publishing Co. (2000) ISBN 020-170-9147. The XML DSO is a read only data provider. In order to use an XML DSO one must add a Java applet element to the HTML form page. Java is a product of Sun Microsystems, Inc of Palo. Alto, Calif. A param tag within the HTML form page specifies the location of the applet element. The XML DSO retrieves XML database data from a specified location, possibly from an external server computer via a network connection, parses the data, and provides the data to the bound control elements on the page. In this manner, the data consuming elements (the control elements) are isolated from the details of XML.

[0009] The Microsoft DHTML data binding software utilizes the Microsoft ActiveX Data Object (ADO) programming model which is discussed in Understanding ActiveX and OLE by David Chappell, published by the Microsoft Press, Redmond, Wash., ISBN 1-572-31216-5. The architecture of the existing state of the art, as practiced by Microsoft, is depicted in FIG. 1. The ADO programming model is a recordset model. The ADO recordset contains two components, namely a collection of Fields and a collection of Properties. Each record within an ADO recordset has a collection of Fields. The Fields collection is the default collection for an ADO recordset object. Each Field has a Name, Value and Count property. The Count property indicates the number of Fields in the collection. In the ADO recordset Properties collection each property has a Name, Type and Value. The ADO programming model permits only serial addressing of a record set, that is, only one record set at a time is accessible. One result of using the ADO protocol is that there is no practical method of binding one group of data control elements to more than one recordset object.

[0010] In the context of forms, a group of data control elements is a tabular grouping of data control elements that may be aligned either vertically and/or horizontally. A control element is defined as an element for display in a user interface display image, such as a form, questionnaire, web page, browser, spreadsheet, document or any other display, of data displayed for or entered by the user. The control element prompts a user to either make a selection or enter data. A desired presentation grouping of data, such as object oriented hierarchically arranged data, on a form can require data from several underlying recordset objects. The Microsoft ADO record set model does not integrate well with an object oriented hierarchical data buffer structure insofar as an HTML document builder must write code to manipulate multiple dataset objects, also called view objects, in order to bind the desired data to corresponding controls in the HTML form used for presentation formatting.

[0011] When an existing DHTML document is sent by a server to a client machine the document may already have embedded data. The client's web browser subsequently receives and displays whatever data is already contained in the server generated document. The user can then alter this displayed data. A form which has some of its data altered by the client user is returned to the server in its entirety when the user is finished. That is, both the data and the underlying form is retransmitted to the server. The repeated sending of redundant information between the server and client computers regarding a largely static form creates unnecessary network traffic.

[0012] Numerous examples of data binding protocols exist. U.S. Pat. No. 6,014,677, entitled DOCUMENT MANAGEMENT DEVICE AND METHOD FOR MANAGING DOCUMENTS BY UTILIZING ADDITIVE INFORMATION, issued to Hayashi et al. discloses a binding information creating device which associates a document with subsequent evaluation data based on earlier information contained within the document. A tag template is defined by an onscreen editor, and a tag template database is created to associate coinciding tags with the same document.

[0013] U.S. Pat. No. 5,940,075, entitled METHOD FOR EXTENDING THE HYPERTEXT MARKUP LANGUAGE (HTML) TO SUPPORT ENTERPRISE APPLICATION DATA BINDING, issued to Mutschler, III et al. discloses a web server program and associated database for storing description language of a form to be displayed. The server is coupled to a host having a CPU executing a legacy application containing the form. The server opens the form and associates data names with data values received from the host and sends them to the client.

[0014] U.S. Pat. No. 6,023,271, entitled FRAMEWORK FOR BINDING DATA VIEWERS/DATA MANIPULATION WITH ONE TO MANY OBJECTS THROUGH INTROSPECTION, issued to Quaeler-Bock et al., discloses a data structure that enables a client application to bind a set of GUI components to the attributes of at least one Business Object (BO).

[0015] U.S. Pat. No. 5,555,365, entitled METHOD AND SYSTEM FOR OPTIMIZING STATIC AND DYNAMIC BINDING OF PRESENTATION OBJECTS WITH THE OBJECT DATA THEY REPRESENT, issued to Selby et al. discloses the creation of a table that specifies relationships between GUI objects and the application object. Each time the application object is initialized the table is used to associate GUI objects to objects within the application object.

[0016] U.S. Pat. No. 5,430,836, entitled APPLICATION CONTROL MODEL FOR COMMON USER INTERFACE ACCESS, issued to Wolf et al., discloses an Application Control Module (ACM) that is executable by the applications. The ACM includes functional elements for initializing the data in the application, drawing or presenting a display screen defined by the data, running or processing user input events in accordance with the operation defined by the data, and closing the application.

[0017] U.S. Pat. No. 5,832,532, entitled MODEL INDEPENDENT AND INTERACTIVE REPORT GENERATION SYSTEM AND METHOD OF OPERATION, issued to Kennedy et al. discloses an interactive report generation system that includes a compiler, an evaluator, a renderer, and model interface functions. The compiler receives report, layout, and worksheet definitions, accesses model interface functions, and generates a report template. The evaluator receives the report template, accesses model interface functions, accesses a user model, and generates a report instance. A renderer receives the report instance in order to display and allow interaction with that report instance.

[0018] A presentation grouping of data on a form can originate from several underlying dataset objects. It is desirable that any desired control grouping should able to be achieved without the need to create additional view objects for each specific display arrangement. This would facilitate the use of business objects as they are defined in the application server without requiring the transformation of the data into an alternate buffer structure, or additional programming of view or dataset objects.

[0019] In accordance with principles of the present invention, a processing system, for use with a user interface display including image elements for prompting user data entry, includes a binding database. The binding database includes elements for linking different first and second data item identifiers with corresponding different first and second prompt elements of a user interface display image. The first data item identifier identifies a set of allowable data items for entry by a user via the first prompt element. A database parser automatically examines the binding database to detect a change in binding database elements. A processor updates the set of allowable data items in response to any detected binding database change.

[0041] FIG. 2 is a block diagram which depicts the processing system 1 of the present invention. The system 1 is implemented in a personal computer which contains software and hardware (not shown) permitting a connection to the internet or other network populated by other computers. Operation of the system begins with a user viewing information displayed on a monitor or other user interface display device 2. Various types of image elements may be visible on the screen 2. The image elements are typically part of a document, form, questionnaire, spreadsheet, web page or browser. Referring to FIG. 3, the display 2 depicts, for example, an HTML form 3 in which the image elements include various prompt elements, e.g. 4, 5, 6, 7, 14, 15, 16 and 17, arranged in a tabular orientation as they would appear on the user display interface 2. As illustrated, these prompt elements may be aligned vertically and/or horizontally. The form 3 includes a prompt element 4 in which a street address may be entered, a prompt element 5 in which a zip code may be entered, a prompt element 6 in which a state may be selected from a drop down or pick list type of menu, and a prompt element 7 in which a city may be entered, and so forth. This information forms a physical or mailing address. Data may be entered into the form 3 by the user and/or may be obtained from the Server machine when the form initially loads onto the computer.

[0042] The prompt elements 4, 5, 7, 14, 15 and 16 are a type of prompt element known as a text box, which can display text for a user, and accept textual data from a user. The text box type prompt element is associated with a data item from which it receives its initial data (if any), and into which the data entered by the user is inserted. The prompt elements 6 and 17 are a type of prompt element known as a combo-box. In a combo box, a list of acceptable data entries is presented to the user when the user activates the arrow button at the right side of the prompt element. For example, for prompt element 6, a list of U.S. states is presented to the user when the arrow button is activated. The user can then select one of the acceptable data entries (e.g. one state) from that list. The selected data entry is then displayed in the display area making up the remainder (left hand side) of the prompt element. As before, the combo-box type prompt element may also be preloaded with initial data. The combo-box type prompt element, therefore, requires a source of data to provide the acceptable choices, and a data item from which it receive its initial data (if any), and into which the selected choice is inserted.

[0043] The HTML code defining form 3 is stored within the client personal computer, while the data populating the entry fields in the form is ultimately received from and subsequently transferred to other computers, such as a server machine, via the internet connection. FIGS. 10A, 10B, 10C and 10D, when joined together, are an example of an HTML code listing which defines such a form that may be stored on a personal computer. Within the HTML code is JavaScript code (lines 130-131) that will condition the personal computer to obtain data items residing on a Server machine at the time the form loads. Further JavaScript code (lines 132-133) conditions the personal computer to load these data items into an XML document embedded within the HTML document.

[0044] When user input data processor 8 (FIG. 2) receives user data intended for form 3 (FIG. 3) from, for example, a keyboard or mouse, the data is processed according to information, stored in a binding database 11, associated with the user input data. For example, referring to FIGS. 4, 6 and 7, the binding database 11 includes a binding table 13, and a first (40) and a second (41) data constraint table. The user input data processor 8 forwards the user data to binding database 11 where it is compared to information in the binding table 13 (FIG. 4) contained within the binding database 11. The binding table 13 in the binding database 11 associates data from the external server with corresponding prompt elements in the form 3 in a manner to be described in more detail below. The received user data is also compared to information in data constraint tables 40 and 41 to determine if the received value of the user input data meets data constraints (i.e. is it within proper range and/or does it coordinate properly with other data) in a manner also to be described in more detail below. If the user input data is within the appropriate constraints, it is forwarded to the processor 10 for subsequent display by the user interface display 2 and possible storage in the associated external data item. Conversely, if the data requires alteration and/or further processing before it may be properly entered in form 3, the processor 10 is conditioned to either perform the proper processing, or to inform the user of the problem via the user interface display 2 and solicit further information from the user via the user input data processor 8 to resolve the problem.

[0045] FIG. 4 illustrates a portion of a binding table 13 containing a plurality of entries, illustrated as rows. Each row specifies the linking of a data item with a prompt element. Within the binding table 13 each prompt element 14, 15, 16, and 17 is linked with at least one specified data item received from the server. The first column in binding table 13 identifies a prompt element on the form 3, and the third column identifies the data item from the external server to be linked with that prompt element via one or more data item identifiers 19, 20, 21, 22 and 23, each of which resolves to a set of data items. The second column identifies the type of the prompt element identified in the first column. In most cases there is only one data item identifier linked to each prompt element, and the data item identifier resolves to a set consisting of only one data item. In general, however, the data item can be a simple object property (such as the value of the object), an entire object (containing all of its properties) or an array (repeating group) of objects.

[0046] In a preferred embodiment of the present invention Microsoft XSL pattern syntax (previously known as XQL syntax) is used to represent the data item identifier. Microsoft documentation likens XSL Pattern syntax to SQL for use with XML, hence the original name of XQL (XML Query Language). XSL defines a “pattern” syntax which identifies nodes within an XML document. This capability provides the equivalent of an SQL WHERE clause. XSL is further discussed in Professional XSL by Kurt Cagle, Michael Corning et al., Wrox Press, Inc. (2001) ISBN 186-100-3579.

[0047] The XML data buffer tags (shown without any data entered) associated with the image elements or controls 4, 5, 6 and 7 appearing in form 3 would appear as follows:

[0048]

	<Patient>
	<Address>
	<AddressLine1/>
	<AddressLine2/>
	<City/>
	<State/>
	<ZipCode/>
	</Address>...
	<Address>...
	<Address>...
	<Patient>

[0049] In the preceding example the data item identifier /Patient/Address/ZipCode resolves to a data item having one object property: the value of the zip code. The data item identifier /Patient/Address resolves to a data item that is an entire object containing several objects with their own properties (multiple lines of address information, plus City, State and ZipCode), and the data expression /Patient/Address[State=PA], for example, resolves to a data element which is an array composed of a repeating group of objects, i.e. those patients who live in Pennsylvania.

[0050] Thus, a data item identifier resolves to a data item or to some portion of data within the embedded XML document containing external data. In XML this reference is to a document node, that is, one branch of the XML document as illustrated, for example in FIG. 11. A data item identifier can represent the data source for a prompt element, the data target (destination) for a prompt element or both.

[0051] For example, a combo box needs a data item identifier which will resolve to a list of allowable choices and it also needs a data item identifier which will resolve to a location in the external data where the choice made by the user will be placed:

[0052]

	DataSource:	ListOfDoctors
	DataTarget:	Patient/AttendingDoctor

[0053] Referring specifically to rows 30 and 31 of the binding table 13 (of FIG. 4), these rows both relate to the VIP status combo box 17, identified as prompt element ‘ctlVIP’ in the first column, and of type ‘ComboBox’ in the second column. The data item associated with the prompt element ‘ctlVIP’ 17 in row 30 is the source of the allowable values of the VIP status. In the illustrated embodiment, the allowable values are ‘True’ and ‘False’. However, there is no limit to the number of allowable choices, and for the present example, the allowable values could indicate different levels of VIP status, such as: ‘gold level’, ‘silver level’, ‘bronze level’ and ‘none’. This source data item is identified in cell 22 as existing at the location in the embedded XML document: ‘/AllowableValues/VIPStatuses’ which contains a list of the allowable data items. As described above, this list would be retrieved as an array containing the allowable data. The data item linked with the prompt element ‘ctlVIP’ in row 31 is the destination data item for data entered by the user into the combo-box 17. This destination data item is identified in cell 23 as being at the location ‘/Patient/VIPStatus’.

[0054] Data item identifiers in the binding table 13 can be changed at runtime to allow for dynamic views, that is, to display subsets of available data in the buffer which change based on data entered into other prompt elements.

[0055] The data item identifier associated with a prompt element can be stored with that prompt element (i.e., in the XML document defining the form containing that prompt element), with the data (i.e., in the embedded XML document) or separately. In a preferred embodiment of the present invention the most straightforward and flexible way to store the data item identifier information is separately in a data binding table 13, as illustrated in FIG. 4.

[0056] In FIG. 10B, the HTML code at line 127 begins the data binding process. The HTML code conditions the client computer to retrieve the external data from the server (line 130-131) via a server XML document, and loads the received external data into an XML document (line 132-133) which is embedded at line 106 into the form. FIGS. 11A and 11B depict the XML code listing for an exemplary XML document such as the XML document referenced at line 106.

[0057] Data items in the XML document are bound to prompt elements via JavaScript subroutines depicted in FIGS. 12, 13, 14 and 15. That is, these JavaScript subroutines perform the transfer of data between the data items in the embedded XML document and the associated prompt elements in the form. In order to accomplish this data binding process, the location of the data in the embedded XML document is linked with the prompt element where the data ultimately will be displayed and edited via binding table 13, as described above. The entries in the binding table 13 are defined in the HTML form (FIG. 10A) at lines 120-126. The entries illustrated in FIG. 10 represent only a portion of the binding table entries: those illustrated in FIG. 4. Lines 122-126 insert data into the binding table 13 to map the data items in the embedded XML document to the appropriate prompt elements via an XQL pattern or path that resolves to a node or nodes in the XML document.

[0058] The routine listed in FIG. 16 operates in conjunction with the code in the HTML document (FIG. 10) to populate the embedded XML document. This routine inserts data items from the XML document from the external source into the appropriate node in the embedded XML document to which the XQL resolves.

[0059] In FIG. 4 the binding table 13 links the XML data item from the embedded XML document, obtained at line 130 of the HTML form (FIG. 10), to the specific prompt elements that appear in the HTML document at lines 122-126. This scheme offers the advantage that more than one prompt element 14, 15, 16, etc. can be mapped in the same binding table 13. In addition, the binding table 13 can be quickly scanned to discern which prompt elements are bound to which data item, a necessary operation in order to refresh the display 2 for prompt elements whose data has changed. To aid in performing this necessary function, the JavaScript subroutine depicted in FIG. 18 determines which prompt elements are bound to a specific branch in the XML document.

[0060] Every prompt element (14, 15, 16, and 17) that is linked with a data item (19, 20, 21, 22 and 23) in the binding database 11 has at least one entry in binding table 13. Each prompt element is further associated with a generic binding routine based on the type (24, 25, 26, 27 and 28) of the prompt element. This generic binding routine performs the actual copying of the data between the embedded XML document and the linked prompt element. For example, the generic binding routine associated with prompt element types 24, 25 (InputBox) would be InputBox_XML_UI, where UI refers to the user interface (FIG. 13A, line 320). This generic binding routine transfers textual data between a specified text box prompt element and the data item in the embedded XML document to which it is linked.

[0061] Other examples of binding routines applicable to each type of control scenario are shown in FIGS. 12-15. FIG. 12 is an example of the routine which controls the binding the XML data to all of the linked prompt elements in the form. FIG. 13 is a routine which controls the binding of linked XML data to only a specified prompt element. FIG. 14 depicts the binding routine, described above, that will bind the XML data to a relatively simple control such as an input box, while FIG. 15 shows the binding routines for a relatively complex control such as a data list or combobox.

[0062] The rows 29, 30, 31 and 32 of binding table 13 (of FIG. 4) are an example of how prompt elements within one presentation group form can be linked with data items located within different object and/or record types. In this case, the four rows 29-32 are associated with three different record types, namely Person/Name 19, 20; Person/Education 21 and Patient 23.

[0063] Data items are automatically read from the embedded XML document and written to each prompt element (4, 5, 6, 7) in the form 3 when the form 3 is initially loaded. As seen in FIG. 5, the loading of form 3 into the display interface 2 begins with Form Load Event 33. At step 34, the generic Buffer_to_<Control Type> routine is launched for each specific control type appearing in the Binding Table 13. For example, a ComboBox type prompt element would cause initiation (FIG. 13, line 322) of the routine XML_to_ComboBox (FIGS. 15A and 15B). Similarly, an InputBox type prompt element would cause initiation (FIG. 13, line 320) of the routine XML_to_InputBox (FIG. 14).

[0064] If the data item to which the prompt element is bound is not found in the embedded XML document, the associated controls are hidden. This feature allows a security layer to remove sensitive data from the embedded XML document or to prevent such data from ever reaching the buffer initially. Further, hiding of the unbound controls prevents the user interface from failing. Furthermore, hiding the control and the prompt for the control prevents the user from realizing that a potentially secure field even exists. When the control is hidden the control still occupies the same physical space on the display 2 so that other controls on the form 3 are not affected.

[0065] Following the scan at step 34 for entries in the binding table 13, step 35 processes the buffer initialization constraints appearing in table 40. Certain data constraints are static and can exist as part of the form 3. For example certain prompt elements may be linked to data items that are always required (e.g. ‘Patient/Name/familyName’ 19 and ‘/Patient/VIPStatus’ 23). Also, certain pick lists (e.g. list of allowable VIP status values 6) can also contain values that do not often change.

[0066] In FIG. 6, table 40 depicts an example of data constraints 36 and 37 which are evaluated at the time of form initialization (INIT). When the form 3 initially loads, the constraints listed in “Action” column 38 are performed for any prompt element that maps to the data item identifiers appearing in column 39. In the example shown, the prompt element 17 (ctlVIP) will be marked as required and its associated control pick list will be populated with the values “True” and “False”. INIT constraints 36, 37 are processed when the linked data items are initially loaded into the prompt elements. Since INIT constraints 36, 37 do not change after the form is initially opened, these constraints are never again processed.

[0067] Prompt elements are activated if their linked data item is relevant and inactivated if their linked data item is not relevant. Furthermore, prompt elements are highlighted if their linked data is marked as required. The routine depicted in FIG. 17 marks as required any prompt elements that are bound to an XML data item marked as being required. The routine shown in FIG. 19 highlights prompt elements marked as required, while FIG. 20 highlights the label (if any) of prompt elements which are marked as required. Similar routines (1) mark prompt elements as not relevant if their linked data item is irrelevant, and inactivate both the prompt element and the label (if any) of any prompt element marked irrelevant; and (2) mark prompt elements as relevant if their linked data item is relevant, and activate both the prompt element and the label (if any) of any prompt element marked relevant. One skilled in the art will understand how to generate code to provide these functions.

[0068] Another class of data constraint is extremely dynamic. In this case, the required data constraint is not known until runtime. For example, the allowable values in a pick list might change frequently enough that embedding such values into the form 3 is inappropriate. This includes constraints that may cause a data item to become e.g. either relevant, irrelevant or required as other data items (and the prompt elements with which they are linked) on the form are changed. For example, the data item last_PAP_Test_Date (not illustrated in FIG. 3) is required if the patient gender (also not illustrated in FIG. 3) is FEMALE. Conversely, if the patient gender is MALE then last_PAP_Test_Date is irrelevant and should not be entered. This class of data constraint needs to be evaluated after each data item that is linked with the underlying condition is changed.

[0069] For example, as soon as the patient gender is changed to MALE, the data item last_PAP_Test_Date becomes irrelevant and each prompt element that is bound to the last_PAP_Test_Date data item is inactivated. Conversely, as soon as the patient gender is changed to FEMALE, the data item last_PAP_Test_Date becomes relevant and required and each prompt element that is bound to the last_PAP_Test_Date data item is activated and highlighted. In FIG. 7, table 41 depicts an example of data constraints which are dynamically evaluated as data is changed on a form. This type of data constraint is marked in the first column to be evaluated at ONCHANGE time and the data items participating in the condition 45 are listed in column 43. Whenever data is changed on the form, the table 41 is scanned to determine if the changed data item matches any data items listed in column 43. If so, the changed data item must be checked against an associated condition, listed in column 45. If the associated condition is met, the associated actions listed in column 44 are performed.

[0070] For example, if /Patient/Sex is the data item that is changed, then both of the entries illustrated in table 41 of FIG. 7 will be triggered, causing their conditions in column 45 to be evaluated. If the condition of column 45 produces a TRUE result, then the corresponding action of column 44 will be performed. Referring specifically to the top row, if the /Patient/Sex changes to “F” (Female) then any prompt element bound to last_PAP_Test_Date (column 42) will be set to “Required”. Referring to the second row, if the /Patient/Sex changes to “M” (Male) any prompt element bound to last_PAP_Test_Date will be inactivated. Similar processing may be performed with respect to a combo-box to change the allowable entries based on a change in data in some other prompt element.

[0071] As data is entered or changed in any prompt element, the new data is automatically inserted into its associated location in the embedded XML document containing the external data item. More specifically, data input events, such as change and click events, are routed to known generic event handling routines. The generic event handling routines for each prompt element move the data from the prompt element to the embedded XML document. An example of the appropriate instruction is:
<ControlType>_To_Buffer

[0072] When data is updated in this manner, the processor 10 detects that the previous version of the updated data item is different from the current version of that data item. The mechanism used to identify such data is to mark that data item as used or “dirty”. That is, whenever a user enters new data or updates existing data, that data is marked ‘dirty’. This can then be used to assist server processing.

[0073] Any data object in the XML document that was originally null (i.e. any XML document node which has no data initially associated with it) is tagged as a placeholder until data is placed into that data object in response to a change in the linked prompt element. Placeholders (new data objects) in the XML document are marked as “Insert” when new data is placed in them. More specifically, once a data object is marked dirty, then that data object is marked as a new object, that is, an object into which data is to be inserted. Any parent (an object containing the new object) placeholder objects are also marked “Insert”. Existing objects, i.e. any node already containing data, are marked as “Update” when new data is placed in them. In addition, previous data values can be saved in a list to support “undo” processing.

[0074] If a subset of the external data is updated from a source other than the prompt elements, e.g. from a data update from the internet, only those prompt elements that are affected by the changed data are rebound. That is, such partial database updates rebind only the relevant prompt elements. The XML nodes of the data changed are compared to the data item identifiers (19-23) in the binding table 13 and only those prompt elements whose linked data item identifiers fall within the changed data path are updated.

[0075] As the user interacts with the prompt elements (4, 5, 6, 7, 14, 15, 16 and 17) of the HTML form 3, the data in at least some of the prompt elements changes. The DHTML browser will initiate events that are directed to the prompt element event handlers, some of which are in the HTML document itself (FIG. 10B, line 137 to FIG. 10D line 190). Some of the event handlers are so generic that they reside in JavaScript files. The prompt element event handlers generally extract the changed data from the prompt element and update the linked node in the XML document. For example, in FIG. 10C (lines 154-164) when the data in the zip code control 5 is changed by the user, line 162 updates the XML document with the new zip code (InputBox_UI_TO_XML (control)).

[0076] Optional custom code may be associated with the generic event handlers. For example, additional program code may be invoked by creating event handling functions using a specific naming convention (<controlName>_<eventName>). The generic event handling routines will search at runtime for the existence of additional custom functions and, if found, invoke those functions instead of executing their generic functions. An example of custom logic that may be added before or after data updating is as follows:

[0077]

function ctlZipCode_onchange(aNumericEdit)
{
	// do custom validation here
	if passMyValidation(aNumericEdit) == false
	return; //abort buffer updating
	//otherwise let the system update the XML doc
	this.UI_TO_XML();
}
function ctlZipCode_onafterchange(aNumericEdit)
{
	//custom code here
	DoSomething(); //Verify City and State?
}

[0078] Referring to FIG. 8, the form data change processing model described above is illustrated. The form data change event handler 50 begins with waiting for a data change event to occur at step 51. At step 52 the subroutine searches for the presence of custom pre-change code before changing the data in the database 11. If such custom code is found to exist, step 49 executes the custom code. If no such custom code is found, the data is moved to the XML document at step 53 using the generic binding routines so that the data may eventually be sent back to the Server machine. After the data change is executed, another search is performed at step 46 for the existence of custom post-change code. If such custom code is found it is executed at step 47. If no such custom code is found, then in the processing at step 48, the ON_CHANGE constraints in the data constraints table 41 of binding database 11 are evaluated and if any are triggered, they are executed. The routine depicted in FIG. 9 is a JavaScript code listing which examines the data runtime for ON_CHANGE validations for the selected control.

[0079] The instruction loop depicted in FIG. 8 exists for each prompt element. As seen in FIG. 10, the HTML form data change instructions appear at lines 138-147 for the “street” data, at lines 148-153 for the “address” data, at lines 154-166 for “zip code”data and at lines 178-190 for “city” data.

[0080] User written functions 49 can implement custom (more restrictive) validation logic which can cause the data binding to be cancelled (data is not written into the database 11) by simply coding a “return” statement. User written functions 47 can be implemented which are invoked after the data is written to the database 11 in order to cause or perform some post data change task such as data transmission.