REGISTER  |  SIGN IN
United States Patent6006242
Poole , ; et al.December 21, 1999
Title
Apparatus and method for dynamically creating a document
Abstract
An apparatus and method for dynamically constructing electronic and printable documents and forms. An entity reference is read from a document instance and compared to entity identifiers provided in a catalog containing a plurality of entity identifiers. Each of the entity identifiers in the catalog is associated with an entity resolution process. An inference engine or other entity resolving processor is invoked to effectuate the resolution process associated with a matching entity identifier. The inference engine or entity resolving processor resolves the entity reference to a resolved entity, such as a component of text or graphics to be included in a document. Linking between the document, entity reference, and resolved entity provides for detailed auditing of the entity resolution process. A resolved entity may contain one or more embedded entity references which are similarly resolved. The dynamic document construction methodology may be implemented using a distributed networking approach, or on a stand-alone computer system. A significant advantage of the present invention concerns the re-usability of textual, graphical, and other components, thereby providing for the construction of any arbitrary document type having any arbitrary number of presentation formats. In one embodiment, the inference engine used to resolve entity references is converted to an executable form to enhance portability. A document or form constructed in accordance with the present invention may be published in printed or electronic form, such as in the form of a World Wide Web (Web) page.
Inventors:Poole; Donald L. (St. Cloud, MN), Wyman; Richard K.  (St. Cloud, MN)
Assignee:Bankers Systems, Inc. (St. Cloud, MN)
Appl. No.:628874
Filed:April 5, 1996
Current U.S. Class:715/531 
Current International Class:G06F 17/27 (20060101)
Field of Search:707/505-508,530-531
U.S. Patent Documents
4866634September 1989Reboh et al.
4876665October 1989Iwai et al.
4905163February 1990Garber et al.
4964060October 1990Hartsog
4970658November 1990Durbin et al.
4984174January 1991Yasunobu et al.
5043891August 1991Goldstein et al.
5133045July 1992Gaither et al.
5133075July 1992Risch
5159669October 1992Trigg et al.
5161211November 1992Taguchi et al.
5218539June 1993Elphick et al.
5222236June 1993Potash et al.
5323313June 1994Davis et al.
5423041June 1995Burke et al.
5446653August 1995Miller et al.
5448729September 1995Murdock
5485544January 1996Nonaka et al.
5524257June 1996Koike et al.
5598511January 1997Petrinjak et al.
5608857March 1997Ikeo et al.
5671429September 1997Tanaka
5778398July 1998Nagashima et al.
Foreign Patent Documents
WO 95/19010Jul., 1995WO
Primary Examiner: Fetting; Anton
Attorney, Agent or Firm:Merchant & Gould P.C.
Claims

What is claimed is:
1. A method of electronically constructing a document, comprising the steps of:
reading a text component reference;
comparing the text component reference to a plurality of text component identifiers provided in a catalog, each of the text component identifiers in the catalog being associated with a resolution strategy;
implementing, in response to successfully comparing the text component reference with a matching text component identifier in the catalog, a resolution strategy associated with the matching text component identifier in the catalog using an inference engine to resolve the text component reference to a corresponding resolved text component;
outputting the resolved text component for subsequent incorporation at a location in the document;
generating a plurality of distinct document versions of the resolved text components using an artificial intelligence expert system to create each of the plurality of document versions based on a decision generated by the expert system's set of predetermined formatting rules; and
selecting one of the plurality of the document versions to be used as the document.

2. The method of claim 1, wherein:
the text component reference includes an embedded text component reference;
the implementing step includes the step of implementing, in response to successfully comparing the embedded text component reference with a matching text component identifier in the catalog, a resolution strategy associated with the matching text component identifier using an inference engine to resolve the embedded text component reference to a corresponding resolved embedded text component; and
outputting the resolved text component and the embedded text component for subsequent incorporation at the location in the document.

3. The method of claim 1, including the step of formatting the resolved text component in accordance with one of a plurality of formatting types.

4. The method of claim 1, including the step of entering the text component identifier and associated resolution strategy into the catalog.

5. The method of claim 1, wherein the outputting step includes the step of outputting the resolved text component for incorporation into the catalog using the inference engine.

6. The method of claim 1, wherein:
the catalog includes a plurality of text component identifiers, at least one of the plurality of text component identifiers being associated with a sequence of resolution strategies; and
the implementing step includes the step of implementing a first resolution strategy of the sequence of corresponding resolution strategies to resolve the text component reference to the resolved text component.

7. The method of claim 6, including the step of ordering the plurality of text component identifiers so as to control the sequence of resolution strategies in the catalog.

8. The method of claim 1, including the further steps of:
formatting the document in one of a plurality of format styles; and
outputting the formatted document in one of a printed and electronic form.

9. The method of claim 1, including the further steps of:
auditing the implementing step; and
storing auditing information in a file.

10. A method of electronically constructing a document, comprising the steps of:
reading an entity reference;
matching the entity reference with an entity identifier provided in a catalog containing a plurality of entity identifiers, each of the plurality of entity identifiers in the catalog being associated with an entity resolution process;
invoking one of a plurality of entity resolving processors to effectuate a resolution process associated with a matching entity identifier in the catalog;
resolving the entity reference to a resolved entity using the invoked entity resolving processor;
linking the resolved entity to a location in the document;
outputting the document in one of a printed and electronic form;
generating a plurality of distinct document versions of the resolved text components using an artificial intelligence expert system to create each of the plurality of document versions based on a decision generated by the expert system's set of predetermined formatting rules; and
selecting one of the plurality of the document versions to be used as the document.

11. The method of claim 10, wherein:
the plurality of entity resolving processors includes an inference engine;
the resolving step includes the step of accessing a database containing text components using the inference engine; and
the linking step includes the step of linking the text component accessed from the database to the document location.

12. The method of claim 10, wherein:
the plurality of entity resolving processors includes an iterator; and
the resolving step includes the seeps of:
replicating content of the document using the iterator; and
resolving the entity reference to the resolved entity using the replicated document content.

13. The method of claim 12, wherein the replicated content includes content specified by one of a client of the iterator and a user.

14. The method of claim 10, wherein:
the plurality of entity resolving processors includes an entity resolving processor that implements an artificial intelligence procedure; and
the resolving step includes the step of resolving the entity reference to the resolved entity using the artificial intelligence procedure.

15. The method of claim 10, wherein:
the plurality of entity resolving processors includes an inquiry resolving processor that requests an input from a user; and
the resolving step includes the step of resolving the entity reference to the resolved entity using the input from the user.

16. The method of claim 10, wherein the resolving step includes the step of resolving the entity reference to the resolved entity using a transformation rule.

17. The method of claim 10, including the further steps of:
validating the document against a document type definition; and
parsing the entity reference after the step of reading the entity reference.

18. The method of claim 10, wherein:
the matching step includes the step of matching the entity reference with an entity identifier of a previously resolved entity provided in an entity cache prior to the step of matching the entity reference with the entity identifier provided in the catalog; and
the resolving step includes the step of resolving the entity reference to the resolved entity using the previously resolved entity provided in the entity cache so as to bypass the step of invoking one of the plurality of entity resolving processors.

19. The method of claim 10, wherein the linking step includes the step of linking the resolved entity to the location of the document and to an information source accessed by the invoked entity resolving processor.

20. The method of claim 19, including the step of auditing links between the resolved entity, the document, and the information source.

21. A system for electronically constructing a document, comprising:
a central processor coupled to a display and a user input device;
a local memory, coupled to the central processor, configured to store a plurality of catalogs each containing entity resolution strategies and corresponding entity identifiers;
a plurality of entity resolution processors implementable by the central processor;
comparing means for comparing an entity reference read from the user input device with a matching entity identifier contained in the plurality of catalogs stored in the local memory; and
a communications interface coupled to a communications channel and a remote information storage device;
wherein the central processor implements one of the plurality of entity resolution processors indicated by an entity resolution strategy of the matching entity identifier so as to resolve the entity reference read from the user input device to a resolved entity by using information stored in the remote information storage device, the resolved entity subsequently being incorporated into each of a plurality of document versions based on a decision generated by the expert system's set of predetermined formatting rules, selecting one of the plurality of the document versions to be used as the document.

22. The system of claim 21, wherein the local memory is configured to store an audit log containing entity resolution information.

23. The system of claim 22, comprising means for browsing the audit log using the user input device and the display.

24. The system of claim 21, comprising means for replicating a portion of the document and incorporating the replicated portion into the document.

25. The system of claim 21, comprising formatting means for formatting the document for output to one of a printable form and an electronically displayable form.

26. A computer program product, comprising:
a computer usable medium having computer readable program code embodied therein for causing a computer to construct a document by a method including the steps of:
reading an entity reference;
comparing the entity reference to a plurality of entity identifiers provided in a catalog, each of the entity identifiers in the catalog being associated with a resolution strategy;
invoking, in response to successfully comparing the entity reference with a matching entity identifier in the catalog, one of a plurality of entity resolving processors associated with the matching entity identifier;
resolving the entity reference to a component using the invoked entity resolving processor to effectuate the resolution strategy associated with the matching entity identifier;
outputting the component for subsequent incorporation in the document;
generating a plurality of distinct document versions of the resolved text components using an artificial intelligence expert system to create each of the plurality of document versions based on a decision generated by the expert system's set of predetermined formatting rules; and
selecting one of the plurality of the document versions to be used as the document.
Description
MICROFICHE APPENDIX

This application contains a "microfiche appendix" including: Microfiche Appendix I, Document Services Object Listing (2 sheets, 99 frames); Microfiche Appendix II, BFO Processor Server Object Listing (2 sheets, 79 frames; Microfiche Appendix III, BFO Process Server Object Listing For Merge Services (1 sheet, 18 frames); and Microfiche Appendix IV, AutoSelection Server Object Listing(1 sheet, 34 frames).

FIELD OF THE INVENTION

The present invention relates generally to document production, and more particularly, to an apparatus and method for dynamically constructing an electronic document for subsequent publication in pre-printed or electronic form.

BACKGROUND OF THE INVENTION

Document production systems that insert customer-specific content into a pre-fabricated form are known. During a first phase of a conventional document production procedure, a computer-based editor is typically employed to produce a boilerplate document or form. The boilerplate document is then reviewed and revised by a document developer, and stored in a storage device, such as a hard disk drive of a computer. The boilerplate document is then made available for use to an end-user, such as a bank or loan officer. During a second phase, the end-user retrieves the boilerplate document, such as a loan application form, from the storage device and enters customer-specific content into the form. The customer-specific content is then inserted into specified locations on the form, and the completed form is then printed.

An improved insurance document production system is disclosed in U.S. Pat. No. 5,446,653 (hereinafter referred to as "Miller"). The system disclosed in Miller improves upon the above-described procedure by employing a conventional relational database scheme to test customer-specific input information against a table of rule sets which, in turn, are directly linked to various boilerplate clauses. A rule set is assigned to each insurance policy clause and each endorsement clause. The insurance and endorsement clauses and rule sets are stored in a memory coupled to the main processor. Each rule set includes at least one rule that must be satisfied in order to include the associated clause in the document. After entering customer-specific parameters into the computer, such as desired insurance coverages and policy holder's state of residence, each and every rule in each and every rule set is evaluated to determine whether a particular clause is to be included in the document. In order to print a document, a printer database containing a redundant copy of each insurance and endorsement policy clause is utilized to supply the appropriate clauses when printing the final document.

Although the system disclosed in Miller provides for some degree of improvement when automating an insurance document production procedure, there remains a keenly felt need for a system that provides the capability to construct the structure and content of the entire form or document, and one that is not limited to simply inserting limited boilerplate content, such as the policy and endorsement clauses of Miller, into a limited number of locations on an otherwise modified version of a pre-fabricated boilerplate document. There exists a further need for a document production system that provides for a high degree of content re-use, such that a base of content can be used to construct any number and type of document. Moreover, there exists a need for a flexible inferencing capability that dynamically determines content to be included in a document, wherein direct linkage between content and content determining rules is obviated. The present invention fulfills these and other needs.

SUMMARY OF THE INVENTION

The present invention relates to an apparatus and method for dynamically constructing electronic and printable documents and forms. In accordance with an embodiment of the present invention, an entity reference is read from a document instance and compared to entity identifiers provided in a catalog containing a plurality of entity identifiers. Each of the entity identifiers in the catalog is associated with an entity resolution process. An inference engine or other entity resolving processor is invoked to effectuate the resolution process associated with a matching entity identifier. The inference engine or entity resolving processor resolves the entity reference to a resolved entity, such as a component of text or graphics to be included in a final document. A resolved entity may contain one or more embedded entity references which are similarly resolved.

In one embodiment, the entity reference resolution process provides for the production of a stream of resolved entities of components. Any number of documents of varying structure, content, and format may be constructed by referencing one or more components of the component stream. A significant advantage of the present invention concerns the re-usability of textual, graphical, and other document components or fragments, thereby providing for the construction of any arbitrary document type having any arbitrary number of presentation formats. Linking between the document, entity reference, and resolved entity provides for detailed auditing of the entity resolution process.

The dynamic document construction methodology of the present invention may be implemented using a distributed networking approach or a stand-alone computer system. In one embodiment, the inference engine used to resolve entity references is converted to an executable form to enhance portability. A document or form constructed in accordance with the present invention may be published in printed or electronic form, such as in the form of a World Wide Web (Web) page.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a depiction of one embodiment of the document production methodology of the present invention for producing a stream of validated components from which documents of varying type and style may be generated;

FIG. 2 is an illustration of a process by which entity references are resolved to corresponding validated components of the component stream shown in FIG. 1;

FIGS. 3 and 4 illustrate a method of dynamically constructing a document in accordance with ore embodiment of the present invention;

FIG. 5 is a depiction of an embodiment of the document production methodology of the present invention showing various links established between a document, entity references, and document components derived from various information sources;

FIG. 6 illustrates various levels of business rule complexities in connection with the resolution of entity references;

FIG. 7 is a service model representation of one embodiment of a dynamic document construction methodology;

FIG. 8 is a detailed depiction of various services provided in the business service layer of the service model illustrated in FIG. 7;

FIG. 9A is a diagram showing the major elements of one embodiment of a dynamic document construction apparatus;

FIG. 9B is a more detailed illustration of the diagram of FIG. 9A;

FIG. 9C is a block diagram of another embodiment of a dynamic document construction apparatus;

FIG. 10 is a detailed block diagram of a portion of the apparatus illustrated in FIG. 9C;

FIG. 11 illustrates linking between documents, regulations, and entities which provides for enhanced entity resolution auditing and compliance tracking;

FIG. 12 is an illustration of the main user interface window of an SGML Viewer/Editor;

FIG. 13 is an illustration of the main window for interfacing with a Knowledge Base Interface Controller (KBIC);

FIG. 14 is an illustration of the main user interface window of a Document Services Client (DSClient), which is a tool that simulates an application during document and rule development;

FIGS. 15-17 illustrate various windows associated with a tool for developing the layout of an SGML base document;

FIG. 18 illustrates the main window of an Audit Browser that enables a user to interrogate entities and links between entities and documents;

FIGS. 19-26 illustrate various aspects of an inference engine used to resolve entity references in accordance with one embodiment of the present invention; and

FIG. 27 is a system block diagram of a computer-based document construction system suitable for constructing a document in accordance with the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS

DEFINITIONS

The term "SGML" refers to Standard Generalized Markup Language, which is an international standard for structured documents (ISO 8879). Defined formally by the standard as a language for document representation that formalizes markup and frees it of system and processing dependencies, SGML is a language for describing the structure of documents (or information) and for describing a tagging scheme to delineate that structure.

The term "DTD" is an acronym for "Document Type Definition" which is a formal SGML document declaration that begins with the keyword DOCTYPE, followed by the name of the base element (also called the top or containing element) of that document, which, in turn, is followed by the declaration subset which tells a system exactly what markup to expect.

A "tag" is commonly referred to as descriptive markup. A tag is a code that is inserted in a document to identify the information contained in a logical element of the document. Tags delimit (mark, surround, or enclose) an element in an SGML file. An SGML document generally contains both tags and text. Each tag contains three types of information: a start delimiter (usually "<" for an element start-tag and "</" for an element end-tag); the name of an element, and an end delimiter (usually ">"). Examples include: <tag>This is a tagged element.</tag>; and <paragraph>. . . Includes many sentences . . . </paragraph>.

An "SGML Parser" is defined by the SGML standard as a program (or portion of a program or a combination of programs) that recognizes markup in SGML conforming documents. In terms of programming language processors, an SGML parser performs the functions of both a lexical analyzer and a parser with respect to SGML documents. That is, a parser validates DTDs against the rules of SGML and also checks a tagged document against the rules of its DTD. A parser that is able to read a DTD and to check whether markup errors exist and is able to report any existing errors is called a "Validating SGML Parser."

A "catalog" refers to an entity conforming to the SGMLOpen specification for catalogs. A catalog is a dictionary containing a number of entry pairs. Each entry pair in the catalog describes a resolution process or strategy for a specific entity reference.

A "catalog file" is an operating system (OS) file containing a catalog.

The term "catalog precedence" refers to a sequence of one or more catalogs. The catalogs in a sequence may be loaded from files or may exist only in memory. The catalogs are typically evaluated in order from front to back. There are methods described herein to push a catalog on the front of the precedence or on the back.

An "entity" is an item that can be referenced as a unit. Entities provide an easy and useful way of incorporating often repeated characters, strings of characters, characters not available on the keyboard or non-textual data such as images within a document or DTD. Entities also provide a means of incorporating separate units of information (graphics or external files, for example) that are stored externally into an SGML document at the time of processing.

An "Entity Manager" is a software service that resolves entity references. The Entity Manager controls Catalog Precedences and Catalogs, and is a peer of the SGML parser and the application.

An "instance" or "document instance" is a term used herein to refer to a sequential collection of data and markup characters organized to begin with an optional SGML declaration, followed by a document type definition (DTD), followed by the document instance conforming to the DTD.

A "component" or "text component" refers to one or more characters or words coded in ASCII. A component may be considered a text fragment usually no shorter than a sentence, and may include SGML markup.

A "Knowledge Base" is a term that refers to a collection of documents, document components, document type definitions, catalogs, rules, and links.

A "Storage Manager" is an access point to a storage domain. A primary objective of SGML-based systems is independence of information from a particular storage. The link between information and a specific storage domain is performed by a Storage Manager. A SystemID contains or implies the name of at least one Storage Manager. A Storage Manager is typically implemented as a Dynamic Link Library (DLL) that is linked to an Entity Manager.

Referring now to the Figures, and more particularly to FIG. 1, there is illustrated one embodiment of the document production methodology of the present invention. In accordance with this embodiment, a document developer initiates a document production session by defining the requirements of the document, as indicated at step 32. A document developer, for example, typically specifies the content that is to be included in a document in order to meet the objectives of the parties to a transaction, and to meet certain business, legal, and/or governmental rules and regulations. Each of the constituent portions of the document is associated with an entity reference which is selected by the document developer, as is indicated at step 34. At step 36, each of the entity reference associated with the document is resolved. A stream 40 of resolved entities or components is produced at step 38 at the conclusion of, or, alternatively, during the entity reference resolution process of step 36.

It is important to note that the entity reference resolution process of step 36 ensures that all business, legal, and governmental requirements applicable to a particular entity reference are duly satisfied. The resolution process thus produces content fragments having integrity by virtue of being compliant with one or more business, legal, or governmental requirements. At step 42, one or more of the components or fragments of the stream 40 are made available for constructing one or more documents having a desired structure and format style.

A significant advantage of the document construction methodology illustrated in FIG. 1 concerns the ability to integrate components selected from the stream 40 of components into SGML documents of varying types and styles. Document-X 44, for example, is shown as having been constructed using resolved and validated components A, B, C, and N in accordance with a first document structure and format style. Document-Y 46 and Document-Z 48 are shown as having been constructed using the same components A, B, C, and N to produce documents having differing structures and format styles. It is noted that other documents can be constructed using one or more of the components A, B, C, and N. It can be seen that any number of documents can be produced in accordance with desired structural and stylistic requirements, and published in printed or electronic form.

The block diagram of FIG. 2 illustrates one embodiment of an entity reference resolution process. A document developer interacts with a user interface 20 to select entity references representative of content to be included in a document. An unresolved entity reference 24 defined in a document instance 22, such as entity reference &A shown in FIG. 1, is initially compared against a Catalog 26 containing pairs of entity identifiers and associated resolution strategies. A comparison is made between the name of the entity reference to be resolved and the entity identifiers contained in the Catalog 26. Upon a successful match, the resolution strategy associated with the matched entity identifier in the Catalog 28 is effectuated by use of an Inference Engine 28. The Inference Engine 28 resolves the entity so as to return a resolved entity 30, also termed a component. Other unresolved entity references 24 contained in the document instance 22 are similarly resolved and typically organized as a linearized stream 40 of resolved entity references or components. Upon resolving all of the entity references contained in a document instance 22, a resolved document instance is thereby produced. A document of a desired structure and format style may then be produced as a printed or electronic document or form.

In one embodiment, a document developer constructs a document or form, such as Document-X 44 depicted in FIG. 1, by use of an SGML (Standard Generalized Markup Language ISO 8879) editing system. Those skilled in the art understand that SGML is a language for describing the structure of documents or information contained therein, and for describing a tagging scheme to delineate that structure within the text of the document. It is to be understood that other markup languages and document production technologies may be employed to effectuate the document construction methodology of the present invention. For purposes of explanation, and not of limitation, the apparatus and method of the present invention will be described herein generally within the context of the SGML standard, including various ancillary SGML capabilities. It is also to be understood that the principles, features, and advantages of the present invention will be described herein generally within the context of an object-oriented programming implementation, although other computer languages and techniques may be employed.

In FIGS. 3 and 4, there is illustrated in flow diagram form various steps involved in creating a document in accordance with the present invention. As is indicated at step 101, knowledge is entered into the Knowledge Base in the form of documents, document components, document type definitions, catalogs, rules, links, and other information needed to construct any number of document and form types. The constituent information contained in the Knowledge Base may include entire documents, portions of documents, phrases, sentences, words, and characters, including non-alphanumeric characters. The content of the Knowledge Base provides the resources needed to construct any type of document or form using the constituent information contained therein.

Knowledge is preferably entered into the Knowledge Base by a domain expert who is experienced in some field of domain of human knowledge. At step 103, the knowledge is entered into the Knowledge Base in units of text or text fragments referred to herein as components. At step 105, the rules that dictate the access and utilization of components are also entered into the Knowledge Base. A document developer and/or user develops a document instance typically by use of a text editor at step 107. The document developer selects one or more entity references to develop a document instance at steps 109 and 111. At step 113, each of the entity references associated with the document instance are resolved. The document components corresponding to the resolved entity references are made available as a stream of resolved entities or components which may be incorporated into a final document. At step 114, one or more documents containing any number of resolved entities or components may be constructed and formatted in accordance with a specific presentation style and published in printed or electronic form.

In FIG. 4, there is illustrated in greater detail one embodiment of the dynamic document construction capability of the present invention. At step 115, the document construction routine is initiated. At steps 117, 119, and 121, the document developer authors a document instance and associates entity references with the document instance in the manner described hereinabove. During the dynamic document construction procedure, an entity reference is read from the document at step 123. One or more catalogs are searched at step 125 in order to match the entity reference with a corresponding entity identifier stored in a catalog. It is noted that more than one entity identifier and corresponding resolution strategy may be stored in one or more of the catalogs. It is desirable that the resolution strategy of the first matching entity identifier in a catalog be executed.

If, as is tested at step 127, the entity reference is matched to a corresponding entity identifier in the catalog, the resolution strategy associated with the matching entity identifier is implemented at step 129, such as by invocation of an Inference Engine. At step 131, the Inference Engine or other resource resolves the entity reference. In some cases, an entity reference that has been resolved may include one or more entity references which require resolving. In such a case, as is tested at step 133, any remaining unresolved entity references that are nested within the resolved entity reference are resolved at step 131.

Upon resolving the first entity reference read from the document, as well as any nested entity references embedded therein, the resolved entity is returned and made available for incorporation into a document in the form of a corresponding document component at step 137. Each of the resolution steps discussed hereinabove is preferably audited, as indicated at step 139, and audit information is stored in an Audit Log for subsequent access. At step 141, the resolved entity may be formatted in a particular manner, published, or otherwise utilized in the document construction routine. It is noted that the format style of the document, as well as any of the document components corresponding to the resolved entity references, is typically determined after completing the resolution process, but may alternatively be determined during the resolution process. Upon completing the resolution process for the first entity reference read from the document, as is tested at step 143, all remaining entity references associated with the document are resolved by repeating the steps beginning with step 123. The document construction routine is terminated at step 145.

Turning now to FIG. 5, there is illustrated in block diagram form a depiction illustrating the methodology by which a document is dynamically constructed in accordance with one embodiment of the present invention. A document, such as document instance-1 62 or document instance-2 64, may be defined from text and graphical components accessed from a Knowledge Base 31. As previously mentioned, the Knowledge Base 31 further includes various document type definitions (DTDs), catalogs, rules, and links. In the embodiment illustrated in FIG. 5, the Knowledge Base 31 is defined to include document components A through F, which, in turn, are linked to a business or governmental regulation source, such as regulation-Y 80 or regulation-Z 90. Also illustrated is a Catalog 26 within which is stored pairs of entity identifiers (component identifiers) and corresponding entity reference resolution strategies. Entity 1 stored in the Catalog 26, for example, is associated with an entity resolution strategy that is implemented by a Store Manager named INFENG, which is a short form of the name Inference Engine. The document instance-1 62 is defined to include entity references &1, &2, and &4. During the document construction procedure, the entity reference &1 is read from the document instance-1 62 and compared against the entries of the Catalog 26. A match is determined between the entity reference &1 and the ENTITY 1 identifier stored in the Catalog 26. The reference to INFENG in the associated resolution strategy indicates that entity reference &1 is to be resolved by employment of an Inference Engine 28. The Inference Engine 28 resolves entity reference &1 to document component-A 66 which is linked to paragraph-1 94 of regulation Z-90. The content of regulation-Z 90 may then be incorporated into a final document 65 by referencing document component-A 66.

By way of further example, the entity reference &2 of the document instance-1 62 is resolved by comparing entity reference &2 with the entity identifiers stored in the Catalog 26. Matching entity identifier ENTITY 2 indicates that entity reference &2 is to be resolved by implementing the Store Manager named FILE 54. The Store Manager FILE 54 resolves entity reference &2 by returning document component-B 68 which is linked to a file containing section-1 92 of regulation-Z 90 as. The content of section-1 92 of regulation-Z 90 may then be incorporated into a final document 65 by referencing document component-B 68.

As will be discussed in greater detail hereinbelow, a particular entity reference of a document instance can be resolved by one or more Storage Managers of varying types. The Store Manager identified as ASK SERVICE 58, for example, may request a particular input from a user by presenting a visual or audible question to the user during the resolution process. The input received from the user, which may be validated, is received and incorporated as the resolution of entity reference &4 of document instance-1 62. Further, the Store Manager identified as SCHEME 60 resolves an entity reference by use of artificial intelligence techniques, such as by use of an expert system. The entity reference &3 contained in document instance-2 64 is resolved by means of an ITERATOR Store Manager 56 which, as described more fully hereinbelow, provides for dynamic construction or replication of a portion of a document, such as additional signature blocks needed to accommodate a co-signer, and incorporation of the constructed portion into the document at run-time. It will be appreciated that information resources other than the Store Managers identified in FIG. 5 may be employed to resolve entity references associated with one or more document instances. Upon resolving all of the entity references contained in a document instance, a prepared document 65 may then be produced. The prepared document 65 may be published as a printed form, an electronic form, or a Web or Internet form, for example.

An important aspect of the present invention concerns an Audit Log 100 within which is stored audit information concerning each document 65 prepared by use of the dynamic document construction methodology of the present invention. The information contained in the Audit Log 100 includes the entity references resolved in the document, the means by which each entity reference is resolved, and the document component corresponding to the resolved entity reference. In the illustrative example shown in FIG. 5, it can be seen that the document instance-1 62 is linked to regulation-Y 80 and regulation-Z 90. When resolving the entity references &1, &2, and &4 of the document instance-1 62, the implicated Storage Managers Inference Engine
28 (INFENG), FILE 54, and ASK SERVICE 58 returned document components A 66, B 68, and D 72, respectively. As such, paragraph-1 94 of regulation-Z 90 is linked to the document instance-1 62 by the document component-A 66. Section-1 92 of regulation-Z 90
is linked to the document instance-1 62 by the document component-B 68.

Further, it can be seen that regulation-Z 90 is linked to the document instance-1 62 by the document component-D 72. The Audit Log 100 will thus contain information concerning all components, information sources, and other linking information concerning the resolution of entity references associated with the resolving of a particular document instance. It can also be seen that all of the document components and document instances that incorporate in whole or in part a particular information source, such as regulation-Y 80, can also be determined. The links established between the various document components, document instances, and information sources can be navigated by use of the information contained in the Audit Log 100. As such, a prepared document 65, which is typically identified by a product ID for a particular customer, can be fully interrogated as to the creation of, or revision to, any of the resolved entities and components constituting the prepared document.

In FIGS. 7 and 8, there is provided a service model view of one embodiment of the present invention. The service model illustrated in FIGS. 7 and 8 is a useful way of viewing applications as a set of features or services that are used to fulfill user requests. The User Service Layer 120 includes various reference applications and utilities that provide for the presentation of information and functionality, navigation, protection of user interface consistency and integrity. A Business Service Layer 122 includes services that provide for shared business policies, generation of business information from data, and protection of business integrity. A Data Services Layer 124 provides for the definition of data, storage and retrieval of data, and protection of data integrity. The interfaces between the User, Business, and Data Service Layers 120, 122, and 124 are preferably published as APIs (Application Programming Interfaces), thus enabling external users of the system to extend or replace functionality by developing their own services.

The services shown in FIGS. 7 and 8 are preferably networked together and operate cooperatively to support one or more business processes. In accordance with this model, one or more applications which can concurrently utilize the various services of the model can be viewed as a collection of User, Business, and Data services that meet the needs of the business process or processes it supports. Because the services shown in FIGS. 7 and 8 are preferably designed for general use and follow published interface guidelines, they can be re-used and shared among multiple applications. Selected services from each of the User, Business, and Data Service Layers 120, 122, and 124 can be packaged in various forms to meet particular system and end-user requirements.

Referring in greater detail to FIGS. 7-14, it can be seen that each of the User, Business, and Data Service Layers 120, 122, and 124 include a number of services having features that will now be summarized in brief. With regard to the User Service Layer 120, knowledge components may be entered into the Knowledge Base 31 by a domain expert through use of an SGML Viewer/Editor, shown generally as user interface 20, a Knowledge Base Interface Controller (KBIC) 176, or Import Service 178. Knowledge is stored and managed by the Document Management System (DMS) 174. The SGML Viewer/Editor 20 is a front end to the SGML Document Management System (DMS) 174. It may be employed to assist authors in creating and testing document components. It is particularly well-suited to authoring and navigating links between components, such as between a unit of text targeted for a document and a unit of text that is part of or represents a regulation. The main window 220 of SGML Viewer/Editor 20 is illustrated in FIG. 12.

The SGML Viewer/Editor 20 is also used as a communications medium between document engineers and application developers. It allows document authors and engineers to record information about document entities and application developers to record information about the resolution of those entities. It provides the resources to store tester responses in objects called scenarios. Scenarios can be modified, subclassed, and executed from the SGML Viewer/Editor 20. It will allow the user to construct queries that navigate links in any direction. For example, the user could determine all regulations impacting a given document, or the user could determine all components impacted by a given regulation. Further, all documents using those components could be identified, as well as all user products based on those documents.

The Knowledge Base Interface Controller (KBIC) 176 is an interface to the rule validation and storage facilities. The main window 230 of Knowledge Base Interface Controller 176 is shown in FIG. 13. This tool is used to select a rule from the Knowledge Base 31, inspect its properties, edit and test it. It interfaces with Knowledge Services and the DMS 174.

A Document Services Client (DSClient) is a tool that simulates an application during document and rule development. It is a client interface to all document resolution and formatting services. The main window 240 of DSClient is shown in FIG.
14. Through the DSClient interface, a document developer can create an environment of SGML catalogs, specify a set of SGML document instances, name the output files, and perform resolution, formatting, preview, and print functions on a document instance. The DSClient allows developers to simulate all parts of the document production process, from install through configuration to document generation.

An Audit Browser 280, shown in FIG. 18, provides access to information in Audit Logs 100 provided through the Auditing Service 156. Individual Audit Logs 100 and other objects can be extracted or deleted, and annotations can be added. The Audit Browser 280 presents a tree view of a selected Audit Log. The entities in the source document instance are represented as first level nodes in the tree. Since resolution of any entity can involve the resolution of many other entities, the firing of rules, and the resolution of Marked Section Control Entities, discussed hereinbelow, the audit records of those resolutions are made available by expanding the appropriate node.

A Business Forms Object (BFO) Processor provides an electronic forms development environment that permits a document author to create and program electronic forms. The BFO Processor can read existing files of varying formats and establish document, page, and field attributes for a BFO. It is able to process large sets of electronic form files in unattended or batch mode. For example, it can be set to create BFO files of all of the files in a particular subdirectory. The BFO Processor can be used to edit BFO properties, allowing the user to place and name electronic form fields. The BFO Processor can produce SGML control instances which, as described in detail hereinbelow, are a special kind of document that is used to prepare the merge data for an electronic form. The main window of the BFO Processor 250 is shown in FIG. 15.

The BFO Processor 250 is an OLE custom control that makes electronic form methods, properties, and events available to Visual Basic.RTM. and other OCX clients. It provides a merge facility to place client data on an electronic form. This facility allows the client to modify field attributes at run-time. It also provides for word wrapping of text across multiple lines and will accommodate text overflow beyond field boundaries. Client access to a preview display for electronic forms is also provided. The user can select pages for printing, select a printer, and print the document without leaving the preview display. Forms-based data collection is made available during view/print such that selected fields are input-capable. Additional features of the Electronic Forms Custom Control include field export and form validation functions, providing client access to the data entered on-screen, and previewing of raw PCL (Hewlett-Packard Printer Command Language) files. In field edit mode, the user can manually create, place, and describe fields for merge data or on-screen data collection. The user can set attributes of fields so that certain fields are input-capable while other fields are not. The user or client can select and place a bitmap or Windows metafile for use as a logo.

A Stylesheet Interactive Development Environment (IDE) is used to create and validate document formatting rules. The Stylesheet IDE operates in conjunction with other document services to test the formatting rules. In one embodiment, the Stylesheet IDE is a visual interface for creating Scheme code. Scheme is an ISO standard artificial intelligence language in which formatting rules will be expressed. The Stylesheet IDE is intended to be used in conjunction with commercial text editors, and one or more editing and testing tools, such as the SGML Viewer/Editor 20, KBIC 176, and/or DSClient.

An Electronic Bulletin is an interactive tool that can be used by a customer to search their Knowledge Base for documents or changes the customer may have made to meet certain criteria. If the criteria are met, the Electronic Bulletin can prepare a report on the actions the customer should take to remain in compliance with regulations and policies that impact the document. Since the customer can modify their Knowledge Base in several ways, such as by altering or adding text, catalog entries, or by establishing institutional policy to include or ignore certain Knowledge Base components, the Electronic Bulletin allows the customer to analyze whether such modifications impact document compliance with regard to regulatory and institutional policy rules and mandates. In one embodiment, the Electronic Bulletin is constructed so that the customer can acquire it interactively, such as by use of the Internet.

A Document Formatter 170 is a service that applies formatting rules to resolved SGML documents. Formatting rules are developed in sets called formatting stylesheets. The technology is based on Scheme, and implements a subset of the ISO Standard called DSSSL (Document Style Semantics and Specification Language, ISO DIS 10179). The DSSSL ISO Standard addresses the need for detailed, typographically sophisticated specification of layout and composition in a manner that is independent of particular formatting systems or processes, as well as independent of the display medium. As a document instance is parsed, the formatting rules from the selected stylesheet are attached to the document, creating a Scheme program. When parsing is complete, the Scheme program is executed. The outcome of the Scheme program is the desired transformation. For example, an RTF (Rich Text Format) formatting stylesheet would result in a Rich Text Format file that could be imported into a word processor for preview and printing. The Document Formatter 170 can create RTF sequences that enable the formatted document to be processed in a form useable by commercial word processors. RTF Services are used in conjunction with the Document Formatter 170 to provide access to field editing. The Document Formatter 170 can also produce transformed SGML documents, such that an original SGML document is transformed into another SGML document.

In one embodiment, RTF Services are a set of templates and macros for Microsoft.RTM. WordMS-Word is used to preview RTF files. The RTF output produced by the Document Formatter 170 references a pre-established template. When MS-Word is invoked to import and display the RTF file, macros in the template protect the document from arbitrary modification by the user. Document editing in MS-Word is permitted. Data objects that are candidates for editing are formatted as RTF fields. Those fields could represent transaction data or could also be whole paragraphs or sections of the document. When those fields are edited, various scenarios are possible, including: the field macro in the underlying template invokes a function published by the client via an API; the field macro invokes an application, which presents a dialog to capture the user changes; and the user is allowed to edit the data and an audit identifier is printed on the document, thereby encoding the nature of the edit.

Another RTF Service is an MS-Word macro that will align all identified fields on a page to an imaginary six lines per inch grid. This function is used in the dynamic creation of static forms, since end-users often require that preprinted forms be adaptable to "typewriter-like fill-in." There are two aspects to the service. First, when the RTF is created by the Document Formatter 170, the fields have to be identified with uniquely named bookmarks. Second, after the RTF is imported into MS-Word, a macro will find and align the bookmarks.

Document Properties are a set of dialogs, such as that shown in FIG. 17, that allow clients to provide user access to such document properties as margins, base font and point size, page orientation, and paper size. The common dialogs can notify the user if certain changes would not be compliant. For example, there may be a legal requirement for a certain sized top margin on the first page of a document. Both the legal requirement and the specific value are determined via the Entity Manager
152. If the user requests a smaller-than-legal top margin, a warning message is displayed, but the selection is allowed. These dialogs may be used as part of a print sequence, or may be used during a configuration activity to establish common settings for documents of a given type.

Transformation Services 134 are used to convert documents from one format to another. The services represented in FIG. 8 are provided as a non-exhaustive list of possible conversion services. Import Services, for example, are services used to import text from `foreign` sources into the Knowledge Base 31, and enable authors to develop knowledge remotely, then import and synchronize that knowledge with the Knowledge Base 31. The DOC to FIF Service is an MS-Word macro that can create a JetForm Field Interchange Format (FIF) file that fully describes a document that has been imported into MS-Word from a Rich Text Format (RTF) file. The FIF file can then be imported into JetForm Design and compiled into a JetForm MDF file. The Logo Transformation Service is a facility to create bitmaps of any arbitrary resolution from EPS files. This utility is often necessary because EPS files are vector images that can be rendered at any resolution. The Eform Transformation Service operates in conjunction with the BFO Processor to create electronic forms in arbitrary formats, such as the .UFF format from Consolidated Business Forms and the JetForm .MDF format.

Entity Services 136 are those services involved in the resolution of entity references. The Entity Insertion service creates entity resolution entries in SGMLOpen-compliant catalogs. The Entity Manager service and its associated Storage Managers uses the information in those catalogs to determine what text or other value should replace any arbitrary entity reference. The Entity Manager 152 is a facility for entity resolution based on the SGMLOpen specification for catalogs. The APIs of the Entity Manager 152 allow clients to modify or override error handling, create and register new Storage Managers 154, and create and register Catalogs 26. The client application can access Entity Manager 152 directly or through the SGML Parser/Validator 150. Client applications can access Catalogs 26 via the Entity Insertion service. Other services must generally be accessed through the Entity Manager interface. The Entity Manager 152 is implemented following the SGMLOpen and HyTime Corrigendum model for catalogs and storage managers. The model allows an arbitrary number of catalogs to be loaded. An entity reference is resolved by searching the catalogs in order for a matching entry. The first entry found is used. The entry specifies the Storage Manager 154 used to resolve the entity reference, along with information to be passed to the Storage Manager 154.

A high-level diagram of one embodiment of the dynamic document construction apparatus of the present invention is shown in FIGS. 9A and 9B. Within the context of an object-oriented programming implementation, Document Services 153 encapsulates the APIs between a client Application 151, the SGML Parser 150, and the Entity Manager 152, which includes or interacts with one or more Catalogs 28 and Storage Managers 154. The Document Services objects, which are described in detail in Appendix I, publish classes for manipulation of an SGML environment, for resolution of entity references and/or documents, and for previewing and printing documents. Also described in Appendix I are the classes associated with the SGML Parser 150, the Entity Manager 152, the Catalogs 28, and the Storage Managers 154.

In order to provide a better understanding of the advantages and features of the present invention, a more detailed discussion of the methodology by which entities are defined, resolved, and utilized when dynamically constructing static and dynamic documents is provided. For purposes of clarity, the term entity is used interchangeably with the term component or text component. An entity may be a string of characters or a whole file of text. An entity may also comprise non-ASCII characters and graphics. In order to include an entity or text component in a document, a construction termed an entity reference is used. For example, the following entity declaration:

______________________________________ <!ENTITY BSI "Bankers Systems, Inc."> ______________________________________

defines an entity having a name BSI and a value defined as the character string "Bankers Systems, Inc." This representation is considered an instance of an entity declaration, which declares an internal entity. The following declaration, by contrast, declares an external entity:

______________________________________ <!ENTITY ChapTwo SYSTEM "sgmlmkup.txt">. ______________________________________

This declaration defines a system entity having a name ChapTwo and a value defined as the text associated with the system identifier. In this example, the system identifier is the name of an operating system file, sgmlmkup.txt, and the replacement text of the entity is the contents of the file.

After an entity has been declared, it may be referenced anywhere within a document. This is accomplished by supplying the entity name prefixed with the ampersand (&) character and followed by the semicolon character. For example, the entity definition:

______________________________________ <para>(C) &BSI;</para> ______________________________________

resolves to:

______________________________________ <para>(C) Bankers Systems, Inc.</para>. ______________________________________

It may be desirable, for example, to construct a paragraph of a document such that a list of remedial actions available to a creditor may be extended as needed. As is shown in the following exemplary declaration, a nested entity reference can be embedded to facilitate document extension:

______________________________________ <Para>The creditor can collect this debt from you without first trying to collect from the borrower. The creditor can use the same collection methods against you that can be used against the Borrower, such as suing you, &AdditionalMethods; etc. If this debt is ever in default, that fact may become a part of your credit record.</Para>. ______________________________________

In the above example, the default list of actions that can be used against the cosigner could be established by defining the entity AdditionalMethods as being empty. In this case, the implicated sentence above would read ". . . such as suing you, etc." If it is desirable to extend the list of remedial actions, the entity AdditionalMethods may be defined to include some appropriate language, such as:

______________________________________ <!ENTITY AdditionalMethods "garnishing your wages, ">. ______________________________________

Using the above entity definition, the text concerning the extended list of remedial actions would read "such as suing you, garnishing your wages, etc." It can be appreciated that entities should be contextually valid and balanced. A balanced entity will typically contain a start tag and a corresponding end tag.

Entities can be used in many ways. For purposes of explanation, it may be convenient to classify entities according to three basic types: Content Entities, Data Entities, and Marked Section Control Entities. A Content Entity is one that represents document language. Typical uses of Content Entities include: language re-use, such that many documents can reference the same entity; alternate text control, such that a user's right to substitute language will be controlled at the Content Entity level; and compliance tracking of regulatory content, such as by tracking links to a government regulations database that will be stored as attributes of the content entity. It is noted that a unit of language that is linked to a specific regulation should be defined as a content entity.

A Data Entity name represents a specific use of data. It is important to note that data entity names are not database field names. A Data Entity name may be subject to specific presentation limitations that may not apply to a field, (e.g., customer-name on a particular document may be limited to 25 characters even though other presentations of customer-name may allow more characters).

A Marked Section Control Entity is an entity whose resolved value can only be "INCLUDE" or "IGNORE." It is occasionally convenient to mark some portion of a text for special treatment by an SGML parser or other language processor. Certain portions of legal boilerplate, for example, may need to be included or omitted systematically, depending on the state in which the document is intended to be valid. For example, the statement "Liability is limited to $50,000" may need to be included in Delaware, but excluded in Maryland. If a Marked Section Control Entity resolves to INCLUDE, the section it controls appears in the document, otherwise, the section does not appear in the document. Marked Section Control Entities are resolved before the document can be completely assembled. It is also possible that the resolution of Marked Section Control Entities will reveal other (nested) Marked Section Control Entities that must also be resolved.

The ability to efficiently and accurately resolve entities of differing types and complexity in a predictable, controlled manner is an important feature of the present invention. Upon initial inspection, entities might appear to the skilled artisan to be nothing more than "include files," such as those used in program source files to re-use bits of source code. However, the entity resolution methodology of the present invention advantageously exploits the SGML parsing model so as to provide enhanced control of the entity resolution process. It will be appreciated that this enhanced entity resolution control capability is equally applicable to parsing schemes other than that defined by the SGML standard. It is understood that a particular entity may be defined multiple times and in multiple locations. The SGML Parser will resolve an entity reference by use of the first definition it finds. This characteristic of the SGML Parser model makes it possible to control the resolution of an entity by, for example, inserting a new definition upstream of a default definition.

Referring now to FIGS. 9C and 10, entity references are resolved by the Entity Manager 152. The Entity Manager 152 is preferably a generalized entity resolver which can be utilized by other system clients that wish to resolve an entity name to a string, text, or other type of component. For example, the Entity Manager 152 may be used by the Inference Engine 28 or by a calculations service. When the SGML Parser 150 encounters an entity reference, such as the unresolved entity reference 24
indicated in FIG. 9C, the SGML Parser 150 transmits the entity name to Entity Manager 152. The Entity Manager 152 initially searches its Entity Cache 168 of resolved entity names and, if found, returns the corresponding string or file contents to the SGML Parser 150. In one embodiment, the Entity Manager 152 is implemented in a 32-bit DLL. In accordance with this embodiment, the service of the entity manager will be the only connection between the SGML Parser and the application. However, the services of the Entity Manager may also be used by any interested system client.

If the entity name is not found in the Entity Cache 168, the Entity Manager 152 searches any of the Catalogs 26 that have been registered with it. In the context of an object-oriented programming implementation, an entity catalog is an object that maps an entity's external identifier or name to a file name or string value. The significant difference between the Entity Cache 168 and a Catalog 26 is that multiple catalogs can be loaded and catalogs may be read from a storage disk, while the Entity Cache 168 is constructed in memory and only one can exist.

Any number of catalogs 26 may be defined and made available to the Entity Manager 152, including the following exemplary catalogs: a standard catalog that defines certain Marked Section Control Entities so as to manage the release of date-sensitive language; a catalog that defines entities such that static forms can be produced; an institution setup catalog that resolves data entities such as institution name and address; a branch setup catalog that resolves lender state, branch name, and address; a policy loan setup catalog for policy loans that may resolve any arbitrary set of entities as is appropriate for a particular policy; a customer catalog that provides for resolving of entities such as customer name and address common to many transactions between the institution and the customer; and a transaction catalog that contains the contents of the Entity Cache 168 after all documents have been assembled.

It is considered important to provide the enduser the ability to control the order in which the Entity Manager 152 evaluates the catalogs 26. A bank, for example, may wish to assert the primacy of the institution catalog over a branch catalog, thus preventing modification of the institution's specified alternate text by the branch location. As is illustrated in the following example, the precedence or primacy of one catalog over another may be controlled by the end-user.

The code provided below in Table 1 illustrates the parsing of a document with a minimum number of objects:

TABLE 1 ______________________________________ Public Function PrepareDocument(pDoc As String, pDocOut As String, pErrorFile As String) As Long Dim mySgmlMgr As Object Dim myParser As Object Dim myCatalog As Object `. . . Create the SGMLManager object Set mySgmlMgr = CreateObject("BSI.SgmlManager") `. . . Load your Storage Manager(s) If Not mySgmlMgr.LoadStorageManager("asksm.dll") Then MsgBox "Unable to load asksm.dll Storage Manager." PrepareDocument = False Set mySgmlMgr = Nothing Exit Function End If `. . . Load your catalog(s) Set myCatalog = mySgmlMgr.PushBack("bsi.cat") `. . . Create a parser Set myParser = mySgmlMgr.CreateParser(Nothing) `. . . Parse the document PrepareDocument = .sub.-- myParser.ParseDocument("E", pDoc, pDocOut, PErrorFile) `. . . Destroy the objects Set myCatalog = Nothing Set myParser = Nothing Set mySgmlMgr = Nothing End Function ______________________________________

In order to utilize multiple catalogs, a separate object must be created for each catalog to be loaded. An efficient means to create multiple catalog objects is to create an array of objects, as illustrated in the code of Table 2 below:

TABLE 2 ______________________________________ Dim myCatalogs( ) As Object Dim myFilename As String `. . . Read text file myFile for the `. . . names of the catalogs to load `. . . Read first record Line Input #myFile, myFilename Do While Len(myFilename) `. . . Allocate a place for the catalog ReDim Preserve myCatalogs(UBound(myCatalogs) + 1) `. . . Put the catalog at the back Set myCatalogs(UBound(myCatalogs)) .sub.-- = mySgmlMgr.PushBack(myFilename) `. . . Read nextrecord Line Input #myFile, myFilename Loop ______________________________________

It is preferable for purposes of enhancing control of the entity reference resolution process that entities are always resolved in a consistent manner. The Entity Manager 152 initiates its evaluation at the front-most catalog, such as Catalog 1
shown in FIG. 9C, and continues progressively toward the back-most catalog, such as Catalog N. The Entity Manager 152 searches for the first occurrence of an entity identifier in the sequence of catalogs that matches the name of the entity reference to be resolved. Thus, the Entity Manager 152 will implement the first resolution strategy it locates upon determining the occurrence of a matching condition. This progression through multiple catalogs by the Entity Manager 152, however, can be advantageously altered in a controlled manner by inserting an appropriate matching entry in the front-most catalog.

If it is uncertain which catalog instance is the front-most catalog, a CatalogPrecedence object can return a catalog object at an arbitrary position in the precedence. The following subroutine provided in Table 3 demonstrates this capability:

TABLE 3 ______________________________________ Public Sub InsertEntitylnFrontmost( .sub.-- pPrecedence As Object, .sub.-- pEntityName As String, .sub.-- pStorageMgrName As String, .sub.-- pSystemID As String) as integer Dim myFrontCatalog As Object `. . . Get the front catalog Set myFrontCatalog = pPrecedence.Index(0&) myFrontCatalog.AddEntity pEntityName, .sub.-- pStorageMgrName, pSystemID Set myFrontCatalog = Nothing End Sub ______________________________________

This subroutine can be modified to enable controlled insertion of a catalog entry in any arbitrary position by including the position number as a parameter. As such, the client of this subroutine must be informed as to the number of loaded catalogs, which can be determined by the Count method of the Catalog Precedence.

An Entity Browser, shown in FIGS. 11 and 12, is a tool that permits a user to navigate an Entity Dictionary 166. The Entity Browser provides search and sequence functions and is supported and managed by the DMS 174. It also enables a user to access and modify the attributes of an entity, as well as the text of Content Entities. For entities that have rules expressed in an Inference Engine 28, the Entity Browser connects to the editor of the Inference Engine 28. The Entity Browser is a user's interface to entities when authoring alternate or additional text. It is also used when the user is defining entities for user-supplied electronic forms. Custom Interfaces may be supplied for one or more Entity Dictionaries. Such custom interfaces typically contain the definitions of entities that are specific to certain data processors or user interfaces. Control of the entity resolution process is further enhanced by permitting users, such as financial institutions and their branches, to author alternate text that contains entity references in addition to those initially provided to the users. The institution, for example, may wish to specify that the institution's Entity Dictionary 166 be searched before the branch's Entity Dictionary 166, or vice versa.

Each entity is referenced to one or more business rules that describe the transformations required to respond to an entity resolution request. FIG. 6 shows the main classes of transformations or business rule types. A non-transformation, indicated by line 106, is an action that simply moves a string from some source to the Entity Cache 168. A simple transformation, as indicated by line 108, is an action in which data is altered in some way between the source and the Entity Cache 168. For example, a data value of `1` may be transformed to "INCLUDE," while all other data values are transformed to "IGNORE." By way of further example, a floating point number may be formatted according to some particular specification. A complex transformation as indicated by line 110, is, by definition, any transformation that is not simple. A complex transforation may involve concatenating the values of several entities. It may involve some interaction with a Calculations Service, an Inference Engine 28, or any other business object.

In view of the potential number and diversity of Entity Dictionaries, it is important to define a common data model for entities. It may be desirable, therefore, to ascribe an entity the attributes defined in Table 4 below:

TABLE 4 ______________________________________ ATTRIBUTE ATTRIBUTE DEFINITION ______________________________________ ENTITY TYPE: GENERAL (ALL ENTITIES) Entity Name The name used in the entity definition Public Identifier A name corresponding to the entity name used in formal SGML document interchange and written according to the SGML standard Owner BSI (Manufacturer) or USR (User) Entity Type Data (DTA), Language (LNG), Marked Section Control (MSC) Resolution The default mechanism for resolving Process.sub.-- Formal this entity if it is not in the Entity Cache or the catalogs. Values can be: the Inference Engine (IE) 28 - the entity name is the rule identifier; the Entity Dictionary 166 (ED) - the value in the Blank Form Value field is used to resolve this entity Resolution The probable part of the application Process.sub.-- Short that will supply the resolution. Examples are "Institution Information", "Borrower Information", "Transaction Info - General" Description A narrative describing the entity Loan Purpose An entity could be identified as specific to one or more loan purposes, including consumer, commercial, or agricultural. Author The person who created the entity. Date Created Date the entity was entered in the Entity Dictionary 166 Date Last Modified Date the entity was last changed. Modified By Person made the last change to the entity Indexing Minimum and maximum index potentials Constraints Maximum Length Maximum length of the string that resolves this entity. ENTITY TYPE: CONTENT AND MARKED SECTION CONTROL ENTITIES User override Can this entity be replaced by the user? SGML File Name If applicable, the name of the file containing the SGML text of the entity. ______________________________________

As mentioned previously, the Entity Browser tool shown in FIG. 12 provides a document developer or an enduser an interface to modify entity definitions and textual content. When invoked from an application, for example, the Entity Browser allows the user to specify or insert their own text for any entity flagged as customer modifiable. In addition, the user can insert or modify entity names using any name selected from a set of entity names provided by the application developer. Further, the user can exploit a User Defined Fields mechanism in order to define new entity names. User defined entity names are preferably prefixed with "U." designation. Collection of data values for User Defined Fields is handled by the User Defined Fields subsystem.

An important feature concerns the transformation of an SGML document instance into a static form or document. A static form or document is understood to have a stable or pre-established structure after being constructed into which user information may be subsequently inserted. The structure of a static document, however, is generally not alterable by the end-user. In accordance with the present invention, a static document is constructed dynamically and subsequently published in a pre-printed or electronic static format. In contrast to a static form, a dynamic form has a structure that may be altered to accommodate transaction data, generally to meet a previously unperceived requirement, such as the selection of certain disclosures due to the transaction interest rate and repayment terms. A key distinction is that a static form is one that remains unchanged for many transactions, while a dynamic form is one that is unique for each transaction. The document construction methodology of the present invention provides for the dynamic construction and alteration of both static and dynamic documents and forms.

In the construction of either a static or dynamic document, all entity references incorporated in the document must be resolved. It is noted that the resolution of Marked Section Control Entity references contained in a static form or document may differ from the resolution of the same Marked Section Control Entity references incorporated in a dynamic document. Most Data Entities will resolve to white or blank space in a static form. The process that transforms the SGML document will produce the PCL (Printer Control Language) files for each page and a text format file that describes the locations of the fields on the form, preferably by describing the Cartesian X and Y location coordinates of the fields on the form. The control parameters for the transformation process are captured along with the transaction catalog.

When the appropriate entities are resolved such that the static SGML form is constructed, this set of entity resolutions is obtained from two catalogs. The entities necessary to generate the language of the constructed document are obtained from an E-form Control Catalog. The remaining entities, those that represent the white or blank space on the static form, are obtained from a separate E-form Data Catalog. An entity resolution in the E-form Data Catalog, for example, might be 25 blank spaces, or 40 underlines. The E-form Data Catalog is used to produce the pre-printed version of the static form.

A control instance is a separate SGCML instance that is generated by the BFO Processor from a BFO file. The control instance conforms to a DTD authored specifically to support e-form merge files. Tools and documentation to support end-user creation of a control instance may be provided to permit development of custom forms by the user. Such a tool would support field creation and modification, as well as browsing of the entities available for user selection.

The following code illustrates one embodiment of how a control instance might appear:

TABLE 5 ______________________________________ <!DOCTYPE EFORM SYSTEM [ ->-- Provide for subsetting the DTD <!ENTITY % BSI.TRAN.DTDsubset PUBLIC "-//BSI//ENTITIES Transaction declaration subset//EN"> %BSI.TRAN.DTDsubset; ]> <EFORM> <FM><FormFileName="myform.bfo"></FM <Body> <EformField FXYName="field1">&BSI.DC.CustName.25CharMax; </EformField> <EformField FXYName="field2">&BSI.DC.CustAddr1.25CharMax; </EformField> <EformField FXYName="field3">&BSI.DC.CustAddr2.25CharMax; </EformField> <EformField FXYName="field4">&BSI.DC.CustAddr3.25CharMax; </EformField> </Body> </EFORM> ______________________________________

When the control instance is assembled, it will resolve each Data Entity with the same process as a dynamic document. In the above example of Table 5, the entities to be resolved are shown below in Table 6:

TABLE 6 ______________________________________ BSI. TRAN.DTDsubset the transaction catalog. BSI.DC.CustName.25CharMax a data entity (customer name no longer than 25 characters.) BSI.DC.CustAddr1.25CharMax a data entity (customer address
1 no longer than 25 characters. BSI.DC.CustAddr2.25CharMax a data entity (customer address 2 no longer than 25 characters. BSI.DC.CustAddr3.25CharMax a data entity (customer address 3 no longer than 25 characters. ______________________________________

The transformation process for a control instance provides for the creation of a merge file that can be supplied to a merge tool. For the instance resolved in Table 5 above, the transformation engine determines that the electronic form is located in the file "myform.bfo." When examining this form file, the transformation engine determines that it is a BSI electronic form, for example, and therefore produces a merge file according to that specification. If the electronic form were determined to be a JetForm (.MDF) file, by way of further example, the process would locate the appropriate field position coordinates file and produce the proper command stream for effecting the JetForm MiniMerge process.

For purposes of illustrating the advantages of resolving an entity reference by cooperative operation between the Entity Manager 152 and an Inference Engine 28, the following example is provided. In order to resolve a particular entity reference, the Storage Manager 154 invokes the Inference Engine 28 in accordance with the resolution strategy associated with a matching catalog entity identifier. During the entity reference resolution process, the Inference Engine 28 investigates whether Regulation-Z applies. This inquiry by the Inference Engine 28 may be implemented in the following manner:

TABLE 7 ______________________________________ SomeRule select: case : (RegZApplies) { // some action } default: { T; } } RegZApplies { ASK ("&RegZApplies;" IsTorF(*), RegZApplies); } ______________________________________

It is noted that two rules are used, which provides several advantages. First, the actual implementation of the RegZApplies logic is "hidden" from the referencing rule (SomeRule). Whether the referencing rule performs some logic or represents a simple entity reference is of no import to the referencing rule, or to the author of the rule. The Inference Engine 28 simply needs to know whether or not Regulation-Z applies. This insulation of the logic implementation from the referencing rule advantageously provides for the re-use of rules. A second advantage is that the Inference Engine 28 does not need to know any particulars about the Entity Manager 152. It is desireable that the Inference Engine 28 not be "customized" for any particular possible client in order to maximize its effectiveness. The Inference Engine 28 should be defined so as to be flexible enough to communicate with a wide variety of clients. The string passed back in the ASK transaction may be anything that the caller will understand and be able to resolve.

With further reference to the example of Table 7 above, the Inference Engine 28 requests the Entity Manager 152 for a value for "&RegZApplies;". If RegZApplies has not yet been evaluated, the Entity Manager 152 preferably invokes the Inference Engine 28, or other Storage Manager registered to resolve RegZApplies. Once resolved, the Entity Manager 152 stores that value in the Entity Cache 168, thus making it available to the SGML Parser 150 and any other Entity Manager client.

Consider the case is which state-specific language may be selected for a given document through the use of Marked Section Control Entities. Initially, each document is authored such that the language for all states is ignored. The application simply defines the Marked Section Control Entity for the desired state to "INCLUDE" so as to cause a particular state's language to be incorporated in the document.

Another significant advantage of the present invention concerns the capability to dynamically build SGML structure within a document. As is defined in the Definitions Section hereinabove, Iterators are scripts that can build SGML structures from the information published by business objects via catalogs. They can infer structure from entity names in the catalog and, with a given set of rules, construct an appropriate SGML entity. An exemplary iterator, in accordance with one embodiment, is described below in Table 8:

TABLE 8 ______________________________________ A SAMPLE ITERATOR ______________________________________ ;; Name element (define Name (list "m.sub.-- Name" "Name" "DATA")) ;; Description element (define Description (list "m.sub.-- Desc" "Description" "DATA")) ;; Collateral element (define Collateral (list "Collateral" "Collateral" "GROUP" (list Name Description))) ;; CollateralList element (define CollateralList (list "m.sub.-- CollateralList" "CollateralList" "CONTAINER" (list Collateral))) ;;the following statement executes the iterator (itr-item-container "TXN." CollateralList 1 "COUNT") ______________________________________

The following items are required prerequisites to defining and utilizing an iterator in accordance with this embodiment: The DTD for the target document; the Entity Dictionary 166 entry that represents the location in the target document; and the list of entity names to be published by the business object(s). Examples of the prerequisites are provided in Table 9 below:

TABLE 9 ______________________________________ The DTD section of interest is: <!ELEMENT CollateralList (Collateral+)> <!ELEMENT Collateral (Name,Description)> <!ELEMENT Name (#PCDATA)> <!ELEMENT Description (#PCDATA)> An entity reference has been put in the document where the CollateralList element construct must be entered. This may be defined as: <Para>The collateral items for this loan are:</Para> &BSI.DTA.CollateralList; <Para>The above items . . . An entry is created in the catalog that associates the entity BSI.DTA.CollateralList with an iterator. The specification of this entity may be defined as: ENTITY BSI.DTA.CollateralList "<SCHEME>CollateralList" A business object CollateralList in the transaction publishes according to the following rules: 1. A member COUNT will be published containing the number of collateral items. 2. For each item in the list, a string "ITEM.sub.-- #" will be published in front of the name where `#` is the index in the list. 3. A "TYPE" will be published for each ITEM.sub.-- # specifying what kind of item it is. In addition, CollateralList is a list of Collateral business objects. Thus, for each object of this type, the following will be published: Name Description ______________________________________ m.sub.-- Name The name of the collateral item. m.sub.-- Desc A description of the collateral item. m.sub.-- Value A dollar value attached to the collateral item. ______________________________________

From the above information, it can be assumed that a catalog similar to the following provided in Table 10 may be published by the business object:

TABLE 10 __________________________________________________________________________ TXN.m.sub.-- CollateralList.COUNT "<LITERAL asis>3" TXN.m.sub.-- CollateralList.ITEM.sub.-- 1.TYPE "<LITERAL asis>Collateral" TXN.m.sub.-- CollateralList.XTEM.sub.-- 1.Collateral.m.sub.-- Name "<LITERAL asis>10' boat" TXN.m.sub.-- CollateralList.ITEM.sub.-- 1.Collateral.m.sub.-- Desc "<LITEEAL asis>A standard boat" TXN.m.sub.-- CollateralList.ITEM.sub.--
1.Collateral.m.sub.-- Value "<LITERAL asis>$500.00" TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.TYPE "<LITERAL asis>Collateral" TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.Collateral.m.sub.-- Name "<LITERAL asis>'79 Monza" TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.Collateral.m.sub.-- Desc "<LITERAL asis>A bad car" TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.Collateral.m.sub.-- Value "<LITERAL asis>$100.00" TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.TYPE "<LITERAL asis>Collateral" TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.Collateral.m.sub.-- Name "<LITERAL asis>Home" TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.Collateral.m.sub.-- Desc "<LITERAL asis>Borrower's Home" TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.Collateral.m.sub.-- Value "<LITERAL asis>$8900.00" __________________________________________________________________________

When writing an iterator script, the following objectives should be considered. A CollateralList start and end tag must be wrapped around the whole construct. For each item, a Collateral start and end tag must be wrapped around the Collateral name and description. The m.sub.-- Name member's fully qualified entity must be wrapped with the Name tag inside the collateral element. The m.sub.-- Desc member's fully qualified entity must be wrapped with the Description tag inside the collateral element. The m.sub.-- Value is ignored for each element.

In order to accomplish these objectives, the iterator code should be written to include a description of the mapping of m.sub.-- Name to Name. This is accomplished by declaring a mapping from the unqualified m.sub.-- Name entity to the Name element that is of a type "DATA." This signifies that the unqualified entity m.sub.-- Name will, when qualified, hold the data which must be inserted in the Name tag. This portion of the iterator code may be similar to following:

______________________________________ (define Name (list "m.sub.-- Name" "Name" "DATA")). ______________________________________

The above code indicates that the unqualified m.sub.-- Name entity will be tagged with the element Name which is a data element. This will produce the following: <Name>&Something.m.sub.-- Name;</Name>; where `Something` is the qualification for m.sub.-- Name such that the appropriate entity is retrieved.

A description of the mapping of m.sub.-- Desc to Description may be specified in the following manner:

______________________________________ (define Description (list "m.sub.-- Desc" "Description" "DATA")). ______________________________________

The above code will produce the following:

______________________________________ < Description >&Something.m.sub.-- Desc;</ Description >. ______________________________________

A description of the mapping and structure of Collateral may be specified in the following manner:

______________________________________ (define Collateral (list "Collateral" "Collateral" "GROUP" (list name Description))). ______________________________________

In this description, the collateral element is being mapped from the entity name part `Collateral.` This is required because in the published catalog, every ITEM.sub.-- # was followed by the type "Collateral." Every member of "Collateral" will be prefaced by this name. Thus, to access "m.sub.-- Name," it is necessary to know that collateral is before it. In addition, the name "Collateral" was the name used in the "TYPE" entity for each item. This is important if a list object contains different kinds of information.

A description of the mapping and structure of CollateralList may be specified in the following manner:

______________________________________ (define CollateralList (list "m.sub.-- CollateralList" "CollateralList" "CONTAINER" (list Collateral))). ______________________________________

This description essentially indicates that the CollateralList element is mapped from the entity name part m.sub.-- CollateralList and is a container. A container may contain multiple different kinds of objects. In this case, it is stated that the CollateralList element may only contain Collateral elements, as was specified in the content model for CollateralList.

The following iterator definition is provided in Table 11 below:

TABLE 11 ______________________________________ ;; Name element (define Name (list "m.sub.-- Name" "Name" "DATA")) ;; Description element (define Description (list "m.sub.-- Desc" "Description" "DATA")) ;; Collateral element (define Collateral (list "Collateral" "Collateral" "GROUP" (list Name Description))) ;; CollateralList element (define CollateralList (list "m.sub.-- CollateralList" "CollateralList" "CONTAINER" (list Collateral))) ;;the following statement executes the iterator (itr-item-container "TXN." CollateralList 1 "COUNT") ______________________________________

The iterator defined in Table 11 above will produce the following using the catalog of Table 10 above:

TABLE 12 ______________________________________ <CollateralList> <Collateral> <Name>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 1.Collateral.m.sub.-- Name;</Name> <Description>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 1.Collateral.m.sub.- - Desc; </Description> </Collateral> <Collateral> <Name>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.Collateral.m.sub.-- Name;</Name> <Description>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 2.Collateral.m.sub.- - Desc; </Description> </Collateral> <Collateral> <Name>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.Collateral.m.sub.-- Name; </Name> <Description>&TXN.m.sub.-- CollateralList.ITEM.sub.-- 3.Collateral.m.sub.- - Desc; </Description> </Collateral> </CollateralList> ______________________________________

An iterator may be characterized as being one of three types: a data element description; a group element description; and a container element description. A data element is an element that has a model of #PCDATA, which is defined in the SGML Standard as "parsable character data." The content of the element is the value of some particular entity. The entity that should map to this element may have many different parents depending on where the entity was published. Thus, in the data element definition the entity is only specified by its terminating name. That is, the entity name to map from is the trailing name from the catalog whose parent may be many different things.

During iteration, a procedure called itr-data is called to iterate a data element. This procedure is provided with the parent name. From the parent name and the terminating name, a full entity name may be specified. The syntax of a data element definition is provided as follows:

TABLE 13 ______________________________________ data-element-def => (define name data-element-desc) data-element-desc => (list entity-name element-name "DATA") entity-name => literal element-name => literal name => a name following scheme naming conventions literal => a quoted string following scheme conventions An example of data element: ;; Name element (define Name (list "m.sub.-- Name" "Name" "DATA")) ______________________________________

A group element is an element whose content model is made up of a collection of sub-elements. The definition defines the order in which the sub-elements can appear and what iterator definition to use to create such sub-elements. In addition, it can specify whether there is an entity name fragment that should be appended to the parent so that the sub-elements have the appropriate parent name to build their entity references. The syntax of the group element definition is as follows:

TABLE 14 ______________________________________ group-element-def => (define name group-element-desc) group-element-desc => (list entity-name element-name "GROUP" group-content-desc) entity-name => literal .linevert split. null element-name => literal name => a name following scheme naming conventions literal => a quoted string following scheme conventions null => `( ) group-content-desc => (list definition-by-name+ ) definition-by-name => name Note: A definition-by-name must be previously defined as some kind of valid iterator definition. An example: ;; Name element (define Name (list "m.sub.-- Name" "Name" "DATA")) ;; Description element (define Description (list "m.sub.-- Desc" "Description" "DATA")) ;; Collateral element (define Collateral (list "Collateral" "Collateral" "GROUP" (list Name Description))) ______________________________________

A container element is an element that can contain more than one of a set of elements, including itself. That is, a container element can contain other container elements and itself. The definition contains a set (list) of elements that can be contained within the element. The syntax of the container element definition is as follows:

TABLE 15 ______________________________________ container-element-def => (define name container-element-desc) container-element-desc => (list entity-name element-name "CONTAINER" container-content-desc) entity-name => literal .linevert split. null element-name => literal name => a name following scheme naming conventions literal => a quoted string following scheme conventions container-content-desc => (list ( definition-by-name .linevert split. self-ref )+ ) definition-by-name => name self-ref => (list entity-name element-name "SELF") element-name "SELF") An example is: ;; Borrower element (define Borrower (list "m.sub.-- Party" "Borrower" "GROUP" (list Name Address))) ;; BorrowerGroup element (define BorrowerGroup (list "m.sub.-- BorrowerList" "BorrowerGroup" "CONTAINER" (list Borrower (list "m.sub.-- BorrowerList" "BorrowerGroup" "SELF")))) ______________________________________

Three different commands may be used to manipulate iterator scripts. The first command is "itr-data." This command iterates a data element definition and a given catalog, and generates a data element SGML component. An exemplary usage of this command is given by (itr-data parent definition). The second command is "itr-group." This command iterates a data element definition and a given catalog, and generates a group element SGML component. An exemplary usage of this command is given by (itr-group parent definition). The third command is "itr-item-container." This command iterates a container element definition and a given catalog from a start index to an end index, and generates a group element SGML component. The start and end index values can be given as a terminal entity name to be retrieved from the catalog. An exemplary usage of this command is given by:

TABLE 16 ______________________________________ (itr-item-container parent definition string string) ;; or (itr-item-container parent definition number string) ;; or (itr-item-container parent definition string number) ;; or (itr-item-container parent definition number number). ______________________________________

In order to illustrate an advantageous use of an iterator when constructing a document, the following example is provided. In this example, a borrower construct needs to be generated using an iterator conforming to the following rules: first, a BorrowerGroup start and end tag must be wrapped around the whole construct; second, a Borrower Group consists of one or more Borrower or BorrowerGroup elements; third, if the item is a borrower, a Borrower element will be tagged; fourth, if the item is a borrower list, a Borrower Group element will be tagged; fifth, the following DTD component of Table 17 will be used:

TABLE 17 ______________________________________ <!ELEMENT BorrowerGroup ((Borrower.linevert split.BorrowerGroup) .sup. +) > <!ELEMENT Borrower (ComplexName, .sup. PermAddress) > <!ELEMENT ComplexName (FirstName,LastName) > <!ELEMENT (FirstName.linevert split.LastName) (#PCDATA) <!ELEMENT PermAddress (Street+,City,State,Country) > <!ELEMENT (Street.linevert split.City) State.linevert split.Country) (#PCDATA) > ______________________________________

Finally, a catalog will be published similar to the following partial catalog of Table 18 below:

TABLE 18 __________________________________________________________________________ ENTITY TXN.m.sub.-- BorrowerList.COUNT "<LITERAL asis>3" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.TYPE "<LITERAL asis>Party" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- FirstName "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- LastName "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailCity "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailCountry "<LITERAL asis>USA" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailCounty "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailFromDate "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailNonUS "<LITERAL asis>0" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailState "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailStreet "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MailToDate "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- MiddleName "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermCity "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermCountry "<LITERAL asis>USA" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermCounty "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermFromDate "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermNonUS "<LITERAL asis>0" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermOwnRent <LITERAL asis>1" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermPostatlC ode "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermState "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermStreet1 "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermToDate "<LITERAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.TYPE "<LITERAL asis>Party" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- FirstName "<LITEEAL asis>" ENTITY TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- LastName "<LITERAL asis>" __________________________________________________________________________

In order to accomplish these objectives, the following iterator code is provided in Table 19 below:

TABLE 19 ______________________________________ An iterator definition for Borrower is given by: ;; ComplexName element: (define FirstName (list "m.sub.-- FirstName" "FirstName" "DATA")) (define LastName (list "m.sub.-- LastName" "LastName" "DATA")) (define ComplexName (list '( ) "ComplexName" "GROUP" (list FirstName LastName))) ;; PermAddress element: (define Street (list "m.sub.-- PermStreet" "Street" "DATA")) (define City (list "m.sub.-- PermCity" "City" "DATA")) (define State (list "m.sub.-- PermState" "State" "DATA")) (define Country (list "m.sub.-- PermCountry" "Country" "DATA")) (define PermAddress (list '( ) "PermAddress" "GROUP" (list Street City State Country))) ; Borrower Element (define Borrower (list "Party" "Borrower" "GROUP" (list ComplexName PermAddress))) An iterator definition for BorrowerGroup is given by: ;; BorrowerGroup Element (define BorrowerGroup (list "m.sub.-- BorrowerList" "BorrowerGroup" "CONTAINER" (list Borrower (list "m.sub.-- BorrowerList" "BorrowerGroup" "SELF")))) The iterator must be invoked by the following: ;; Iterate all borrowers (itr-item-container "TXN." BorrowerGroup 1 "COUNT") ______________________________________

The above defined iterator of Table 19 will produce the following using the catalog of Table 18 above:

TABLE 20 __________________________________________________________________________ <BorrowerGroup> <Borrower> <ComplexName> <FirstName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- FirstName;</FirstName> <LastName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- LastName;</LastName> </ComplexName> <PermAddress> <Street>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermStreet 1;</Street> <City>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermCity;</C ity> <State>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1 .Party.m.sub.-- PermState; </State> <Country>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 1.Party.m.sub.-- PermCount ry;</Country> </PermAddress> </Borrower> <Borrower> <ComplexName> <FirstName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- FirstName;</FirstName> <LastName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- LastName;</LastName> </ComplexName> <PermAddress> <Street>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- PermStreet 1;</Street> <City>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- PermCity;</C ity> <State>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- PermState;< /State> <Country>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 2.Party.m.sub.-- PermCount ry;</Country> </PermAddress> </Borrower> <BorrowerGroup> <Borrower> <ComplexName> <FirstName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList. ITEM.sub.-- 1.Party.m.sub.-- FirstName;</FirstName> <LastName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.I TEM.sub.-- 1.Party.m.sub.-- LastName;</LastName> </ComplexName> <PermAddress> <Street>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.ITE M 1.Party.m.sub.-- PermStreet1;</Street> <City>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.ITEM. sub.-- 1.Party.m.sub.-- PermCity;</City> <State>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.ITEM .sub.-- 1.Party.m.sub.-- PermState;</State> <Country)&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.IT EM.sub.--
1.Party.m.sub.-- PermCountry;</Country> </PermAddress> </Borrower) <Borrower> <ComplexName) <FirstName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList. ITEM.sub.-- 2.Party.m.sub.-- FirstName;</FirstName> <LastName>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.I TEM.sub.-- 2.Party.m.sub.-- LastName;</LastName> </ComplexName> <PermAddress> <Street>&TXN.m.sub.-- BorrowerList.ITEM 3.m.sub.-- BorrowerList.ITEM.sub.- - 2.Party.m.sub.-- PermStreet1;</Street> <City>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.ITEM. sub.-- 2.Party.m.sub.-- PermCity;</City> <State>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.ITEM .sub.-- 2.Party.m.sub.-- PermState;</State> <Country>&TXN.m.sub.-- BorrowerList.ITEM.sub.-- 3.m.sub.-- BorrowerList.IT EM.sub.-- 2.Party.m.sub.-- PermCountry;</Country> </PermAddress> </Borrower> </BorrowerGroup></BorrowerGroup> __________________________________________________________________________

It is noted that the above listing was generated from a catalog that continued from the partial catalog shown in Table 18 above. Since the entire catalog is very large, only a portion was provided in Table 18 to illustrate the advantageous use of iterators in accordance with one embodiment of the present invention.

The concept of an iterator will now be expressed in terms of its specific use as a Storage Manager 154 within the context of the embodiments illustrated in FIGS. 9C and 10. In the following illustrative example, the more important steps involved in constructing a document in accordance with these embodiments are described. These document construction steps include:

STEP 1--collecting transaction data (instantiating business objects);

STEP 2--publishing data (each business object writes itself to a catalog);

STEP 3--autoselection (transform the Comprehensive Document List to the Suggested Document List);

STEP 4--interacting with the user to create the Selected Document List from the Suggested Document List;

STEP 5--determining whether additional data must be collected and, if so, notifying the user;

STEP 6--generating documents (resolving document entities); and

STEP 7--formatting, previewing, and printing a resolved document.

In this illustrative example, it is assumed that an Iterator Storage Manager declaration is defined and is included in a Catalog 26 in accordance with the following syntax:

TABLE 21 ______________________________________ ENTITY entity-name "ITR:itr-file-name//parm-name-1 (start-value[ end-value])// . . . //parm-name-n(start-value[ end-value])" itr-file-name The content to be iterated. The content must be in a file. The file name can be entered directly or referenced as an entity. If an entity name is used, the content can be replaced/overridden by normal catalog precedence. parm-name-? The name of the parameter. This is a name that appears between pound marks (#) in itr-file-name. For each parm-name, a stack is created, such that a value can be set and manipulated. start-value The first value to be substituted for the associated parm-name. An entity name can be used. If end-value is omitted, a string or an entity that returns a string can be used. Otherwise, start-value must be an integer or an entity name that resolves to an integer. end-value The last value to be substituted for the associated parm-name. If end-value is omitted or if end-value is the same as start-value, itr-file-name is iterated once using start-value. If end-value is supplied, itr-file-name is iterated for each integer value between start-value and end-value incrementing by +1 each time. If end-value is supplied, and if end- value is less than start-value, itr-file-name is not iterated. An entity name that returns an integer can be used. ______________________________________

It is noted that the start and end values for parms can be qualified by parm-name delimited by pound marks in the same way that parameter names are qualified in itr-file-name. It is further noted that there is no mark between the start and end value. This allows the values be returned by a rule as a list.

The content to be iterated, itr-file-name, must contain SGML processing instructions for each parm-name in order to control the scope of iteration. In Table 22 below, there is provided sample catalog entries that use the itr-file-name language provided in Table 23.

TABLE 22 ______________________________________ ENTITY PartyInfo "ITR:&pPartyInfo;//PartyItem(1 &NumberParties;)//AddressNumber(1 &Party.sub.-- #PartyItem#.NumberofAddresses;)" ENTITY pPartyInfo "FILE:Borrow.itr". ______________________________________

TABLE 23 ______________________________________ <!-- File name: BORROW.ITR --> <?SCOPE PartyItem> <PartyInfo> <Name prefix="Name of Debtor">&Party.sub.-- #PartyItem#.Name;</Name> <?SCOPE AddressNumber> <Address>&Party.sub.-- #PartyItem#.Address.sub.-- #