Method for modeling, storing and transferring data in neutral form

Abstract

The present invention simplifies the data modeling process and enables its full dynamic versioning by employing a non-hierarchical non-integrated structure to the organization of information. This is achieved by expressing data modeling, storage and transfer in a particular non-hierarchical, non-integrated neutral form. The neutral form of the present invention enables complete parallel processing of both data storage and data transfer operations. It also enables the direct integration of separate but related data models and their data without remodeling or reloading. Finally, the present invention enables direct transfer of neutral form information in a manner that includes all of the properties required to independently understand and interpret each transferred data value.

Claims

We claim:

1. A method of modeling a set of information for storage in neutral form in a computer based environment, comprising the steps of:

a) modeling the set of information into instance data sets, each composed of an instance cluster comprised of data instance nodes, each data instance node in an instance cluster containing an assigned distinguishing structural tag comprising:

(1) a data reference

(2) a data type

(3) a data organization; and

b) each structural tag having defined components for each data value in each instance cluster.

2. The method of claim 1, further including the step of:

storing in the computer the assigned distinguishing structural tag for each of the data instance nodes.

3. The method of claim 1, wherein the data instance nodes in the cluster are organized into groups and subgroups for display purposes.

4. The method of claim 1, further including the steps of:

a) dividing the data instance nodes into a plurality of sub-clusters; and

b) designating as the basis a prime/distinct (P/D) data reference set of a cluster to appear in any data entry form for any sub-cluster that does not naturally contain it;

wherein data entry occurs separately at a sub-cluster level without loss of relationship to its proper instance cluster.

5. The method of claim 1 further including the steps of:

a) forming a new cluster and assigning it a name;

b) identifying certain nodes within the existing cluster and moving them to the new cluster;

c) excluding those members of the basis prime/distinct (P/D) data reference set during said step of identifying data instance nodes;

d) copying the nodes of the excluded basis prime/distinct (P/D) data reference set in the existing cluster into the new cluster; and

e) changing the designation of the copied prime/distinct (P/D) data reference set in the new cluster to the distinct form of the basis prime/distinct (P/D) data reference set if it had previously been acting as the prime form of basis prime/distinct (P/D) data reference set;

wherein data entry occurs multiple times in the new cluster as separate cluster instance data sets without redundant data entry in the remaining portion of the existing cluster.

6. The method of claim 5, wherein said step of forming and assigning a new cluster is done on an event time basis.

7. The method of claim 5, further including the step of repeating such steps of claim 5 when the newly formed cluster itself contains a one-to-many relationship.

8. The method of claim 1, wherein existing components of structural tag elements stored in the environment are used wherever their definitions and properties match those required for a new data value.

9. The method of claim 1, wherein the definition of a new or an existing component of a structural tag element which is stored in the environment includes an indication of its equivalency to another existing component of the structural tag element stored in the environment.

10. The method of claim 1 wherein each data reference in the set of information has a granularity which does not exceed that of the basis prime/distinct (P/D) data reference set.

11. A memory storing a modeled set of information in neutral form for processing in a computer, comprising:

a) the set of information being modeled into instance data sets, each composed of an instance cluster comprised of data instance nodes, each data instance node in an instance cluster containing an assigned distinguishing structural tag comprising:

(1) a data reference

(2) a data type

(3) a data organization; and

b) each structural tag having defined components for each data value in each instance cluster.

12. A data processing system, comprising:

a) a memory;

b) a central processing unit executing an application program on an organized set of information stored as a database in said memory;

c) said database stored in neutral form in said memory as a set of information being modeled into instance data sets, each composed of an instance cluster comprised of data instance nodes, each data instance node in an instance cluster containing an assigned distinguishing structural tag comprising:

(1) a data reference

(2) a data type

(3) a data organization; and

d) each structural tag having defined components for each data value in each instance cluster.

13. A method of identifying a pool of data values formed from dynamic liking of two different instance clusters stored in a computer memory, comprising the steps of:

a) selecting a cluster;

b) determining the collection of prime/distinct data reference sets in the selected cluster;

c) identifying another cluster which is on the same time basis and contains the same prime/distinct data reference set;

d) for such another cluster, identifying those instance clusters in which the set of data values for the linking prime/distinct data reference set matches that of a selected cluster instance;

e) for each identified instance cluster, combining the set of data values for that instance cluster with the matching set from the selected cluster to form a pool;

f) removing from the pool any data value that is not part of the instance segment of the universe of the instance defined by the prime/distinct data set through which the dynamic link is occurring; and

g) removing from the pool any respective instance cluster data sets for which the data values do not match any other prime/distinct data value sets that are common to both instance clusters.

14. The method of claim 13, further including the step of:

repeating said step of identifying for other clusters in the computer memory.

15. The method of claim 13, further including the step of:

repeating said step of identifying for other matching prime/distinct data reference sets.

16. The method of claim 13 wherein data values to be removed from the pool are those having a granularity lower than that of the linking prime/distinct data value set that is the basis for the pool.

17. The method of claim 13 wherein the instance segment defined by the linking prime/distinct data value set is a nested instance segment and further that a particular instance data value that is exclusive to the instance segment within which the nest exists is to be including in the pool, and further comprising the steps of:

a) adding the data values that are exclusive to the instance segment within which the nesting is occurring; and

b) removing from the pool any respective instance cluster data sets now so expanded for which there in not a match with the specified exclusive instance data value.

Description

BACKGROUND OF INVENTION

1. Field of Invention

The present invention relates to a method of modeling, storing and transferring information in a neutral form in a computer. The present invention reduces the complexity and effort presently involved in the modeling and storage of information while creating a storage format which enables complete parallel processing of data. Furthermore it enables not only direct integration of different models and their stored data without remodeling or reloading but dynamic evolution of those models and their data after some earlier implementation.

2. Description of Prior Art

Since their introduction about 40 years ago, computers have increasingly been used as mechanisms for the storage of information. FIG. 1 summarizes the prior art for the computer storage, retrieval and transfer of information. Despite exponentially rising demand and all of the effort that has been expended during the past several decades to develop and apply methods for data modeling storage and transfer, the bulk of organized information is still handled outside of databases. This has been due to the complexity, inflexibility and cost of modeling, storage and transfer of organized data that is inherent in prior art techniques.

Whether they employed data models or not; prior art techniques have, so far as is known been unable to achieve any practical success in data transfer without a user being intimately familiar with both the structure and the definitions of the data involved, and also with the particular application programs and interface languages employed for its storage. The result has been an inherent complexity in data modeling storage and transfer, putting data organization and access beyond the reach of many potential information workers who generate and seek to exchange, interpret and use their data.

Users have been forced to apply the skills of information technology specialists for data organization and information management where they can afford or justify the costs. Otherwise, users have been forced to limit arbitrarily the scale of their work where they cannot afford such specialists. At the current complexity and escalating costs of prior art methods, only a small portion of the generated data that exists is processed into database form.

Un-modeled Data Files

Methods available in the prior art for the storage and transfer of un-modeled collections of information have required a great deal of a-priori knowledge on the part of the user, both of the structure and definitions of the data and the application programs involved. Spreadsheets have been one common method employed for such storage. Its usefulness is restricted to collections of data of limited complexity and size. Storage and transfer has typically been required to be a complete rather than a partial spreadsheet file. Un-modeled information has been typically stored directly in an application program format, e.g., an Excel.RTM. or Lotus.RTM. spreadsheet file.

Although the information can be downloaded to ASCII or other primitive standard formats for transfer purposes, it is done with significant additional loss of any already limited data definition. Larger and more complex collections of information typically have been stored in the format of the specialized applications programs within which they are generated, e.g., accounting programs, simulation modeling programs, or engineering calculation programs. Here too, retrieval and transfer typically have been limited to complete rather than partial files.

Often, such information has been merely organized into streams of ASCII or comparable format. This has imposed severe limitations on the storage of structure and content of the data as opposed to merely the data values. Transfer of any and all such organized information, including spreadsheets thus has relied heavily upon a user's knowledge in advance of the structure and content of the information. Typically, a user has operated the same application interface program that generated the data to reduce the degree of difficulty involved in the many interpretive aspects of its transfer. The result is a severely restricted exchange of information both within and between data user organizations.

Modeled Data--Database Files

As shown in FIG. 1, data modeling has represented an increase in the degree of organization from a collection of un-modeled organized information. The clear advantage has been that portions of a collection of information can be transferred rather than only the entire file. However, this advantage has been achieved with a major increase in the complexity and cost of the data management.

Data modeling for a collection of information involves a combination of (a) organization of the data values; (b) some system or technique for definition of each data value; and (c) a system for structuring the relationships among various data values for storage in a manner which was capable of supporting accurate retrieval and transfer of targeted subsets of the information. Many modeling methods have been proposed, and of those only a limited number have proven useful enough to receive widespread use. Examples of proposed methods include: entity relationship (ER); Nijssen's information analysis method (NIAM); IDEFIX, a graphical language; EXPRESS, a product information model; and object oriented model (OOM). Discussions of these methods are available in many publications. Chapter 2, pp. 12-30 of Schenck "Information Modeling: The EXPRESS Way" is representative. Only ER and OOM are, so far as is known, presently in widespread use.

All of these prior art modeling methods have imposed a-priori relationships and some form of hierarchy both for the organization and the storage of data. The particulars of such a hierarchy have varied with the data modeling method and with the individual modeler. In each case, however, the hierarchy selected has been incorporated into a structure of the information for storage and has been different and specific to each major modeling-based database and its associated software products.

Thus the storage of information in relational and object oriented databases has been both program dependent and a function of the particular form of relationships and hierarchy chosen during modeling. As a result, the retrieval and transfer of database information have required a significant knowledge of both the hierarchy imposed on the data and the particular data storage structure that has been employed by the database product employed. These have been major limitations of the current technology which have become particularly evident to users involved in the retrieval and transfer of information.

Efforts have been made to reduce the severity of these problems through the development of a standard database software program interface language--standard query language (SQL). This effort has met with some success, in part because of the rapidly changing features of individual database software programs and the data itself.

Both the known data modeling techniques and the standard query language interface scheme (SQL) also have involved a complex set of rules and conventions whose understanding and practice are well beyond the reach of typical data users. See for example Date, "A Guide to the SQL Standard" Appendix A, pp 137-152. As a result, specialists in database management have been required for data modeling, which has given rise to high total database management costs and complicated communication problems between data users and data modeling specialists. The combination of these limitations has resulted in limited application of data modeling for the storage and transfer of organized data.

Each of the prior art data modeling and storage techniques has involved extensive integrated organization of the data and its structure, which has severely limited the capacity for parallel processing in both data storage and in data retrieval and transfer. Major relational and object oriented database vendors have developed complex optimization routines to guide the selection of pathways through these integrated structures. These have provided limited mapped segments that can be stored and retrieved in parallel. These segments have tended to be short relative to the overall retrieval process, prompting numerous mergings and analysis of intermediate results in order to determine the next correct stage of the storage or retrieval.

All of this effort has consumed processor capacity, reducing the amount of computer capacity available for direct data storage or retrieval work. The more parallel paths that were created, the more complex were the paralleling, merging and analysis efforts. The net effect was rapidly rising complexity and a declining percentage of useful storage or retrieval work produced from each parallel storage or retrieval path that was added.

As a result of such inherent diminishing returns, the prior art was able to achieve only limited gains in storage or retrieval performance through parallel processing. The limited gains that were possible came only with major increases in the complexity of overall database storage, retrieval and transfer operations.

Modeled Data--Neutral Form Files

The storage, retrieval and transfer limitations imposed by database program dependent data hierarchies and structures has been particularly evident in large, complex databases involving diverse interactions. Such complexities were common to the engineering and manufacturing life cycle where there has been a great need to share data across applications, across vendor platforms and between contractors, suppliers and customers. As a result, efforts have been made to define a neutral form for data modeling, storage and transfer which would be independent of both the application program from which it was taken and of any application to which it would be applied. After considerable years of effort, the International Standards Organization (ISO) published ISO 10303, Product Data Representation and Exchange. This set a standard for both a neutral form (ISO STEP Neutral Form) and a modeling language (EXPRESS) through which data was to be organized for incorporation into the neutral form for storage and transfer.

Severe practical limitations have so far led to minimal implementation of ISO 10303. ISO STEP and EXPRESS do nothing to reduce the complexity inherent in prior art data modeling and storage. In fact, they have seemingly increased the complexity by adding a layer of requirements on top of those which have already existed. The overhead associated with constructing an ISO STEP neutral form was high, typically between 10 and 20 times the size of the raw data file. Furthermore, there was no technique for exchanging EXPRESS modeling information together with a neutral form file so that the neutral form information could be immediately interpreted upon receipt of the transfer. While eliminating direct application program dependence, the construction and interpretation of the neutral form file has also still been hierarchically dependent on the EXPRESS modeling form of hierarchical representation.

ISO 10303 could possibly offer potential for the effective use of parallel processing of ISO compliant data. The ISO STEP neutral form took the form of groups of the data and their related information, each of which was organized into a separate and independent record in the neutral form file. Because there was no dependency between groups of neutral form records, each could be processed completely in parallel.

U.S. Pat. No. 4,864,497 attempted to address the issue of practical neutral or common data structures for the storage and access of information by multiple application programs. However, the approach taken preserved data hierarchy and integration as the premise for data modeling. It also adopted a convention for neutral file definition that did not apparently reduce the complexity or rigidity present in prior art data structure technology.

Accounting for Relationships Among Data

Prior art methods for addressing and accounting for relationships among data have been inherently complex, rigid, and highly application program dependent. As shown in FIG. 2, a hierarchy and integration of structures consistently have been employed in prior art data structure technology for the storage, retrieval and transfer of information having any significant complexity. At sufficiently low levels of complexity, non-integrated, non-hierarchical methods have been employed, but they have been limited to application dependent program formats or low level formats such as ASCII, the limitations of which have already been discussed above.

Hierarchy and integration in data structure, including all known forms of relational and object oriented technology, has involved the creation of a top-down logical network to describe all of the inter-relationships and precedence orders associated with the flow of information to be data modeled. The information described by this network has then been used to construct data models which have defined integrated structures for data storage. Since there was a direct relationship between the logical network, its corresponding data model, and integrated structure for the storage of actual data, changes in the logical network necessarily changed the data model, and in turn the integrated storage of the data.

Since the networks involved were typically complex, single changes in the network could precipitate extensive changes in data modeling and storage. Any errors in logic introduced during a change, either in the logic network or data model itself or in the protocols used for storage, could result in faulty storage which produced errors in data retrieval and transfer. Any changes in the data model or data storage protocols made after actual data storage had begun created significant problems for future access to such data.

Prior art technology has sought first to define the largest possible universe within which to construct its information network and second to preserve the resulting information network and its related data model and data structure for as long as possible once it has been established. It has been forced to do so because of various factors. These have included: the complexity and interlocking relationships between the hierarchy and integrated structure of the data model; the corresponding complex structure of data storage and the difficulty in changing data hierarchy and structure once it has been established; the inability to first create and then to merge a series of separate more localized information networks, data models and data structures into one final composite or universal set. See for example Teorey, "Database Modeling & Design: The Fundamental Principles," Chapter 3.

In effect, large amounts of time, often man-months to man-years, have had to be spent gathering information to define a single internally consistent universal information network. This has then been translated into a single rigid fixed element data model and in turn into a rigid, fixed element data storage structure. Once the resulting database was placed in service, there was an almost unyielding resistance to change it. Since the work environment was actually changing daily, the result was that a data model and data storage structure were well out-of-date the day they went into service. However, they remained in service unchanged for extended periods of time because of the great complexity and cost associated with changing them.

U.S. Pat. No. 5,303,367 represented an apparent effort to overcome at least a portion of the severe limitations in the prior art technology. However, the concept of hierarchical and integrated networking, modeling and storage of information was retained completely. The complexity underlying the limitations of prior art data structure technology were thus maintained. This complexity was increased by introducing the concept of a networking structure inversion, which had to be carried out each time a new element of structure was added to the system. In addition, a format for data storage that provided all of the information necessary for use in retrieval and transfer without a-priori knowledge of the data was apparently lacking.

Recently, the concept of data warehouses has become a topic of interest. Data warehouses are one of the mediums for the co-mingling of data from different databases. Prior art data structure technology did not permit, as far as is known, the co-mingling of different data models or their associated data without creation of a new model that encompassed each of the different models and their content. Also, substantial effort or "manual cleansing" of data structures and data information was typically required for legacy or new databases in order to incorporate the information into another database. Included in cleansing operations were the rationalization of a wide variety of different but equivalent data formats or configurations. This was done so that all of the values reported in the consolidated field name or set of field names conformed to the same format and configuration. These cleansing operations represented one of the largest expenditures of time and money in the creation of data warehouses.

SUMMARY OF INVENTION

With the present invention and in the preferred method for its implementation, multiple collections of data items are organized and stored in a computer-based environment as sets of information in a neutral, non-hierarchical form. Optionally, application of the neutral and non-hierarchical form can be limited to the organization of the collection, allowing for the use of alternative means for the storage of the collection of itself.

According to the present invention, neutral form means that each data item in each collection has specified and associated with it sufficient information, expressed in a particular universal and invariant form, to define itself, to distinguish itself from all other data items, and to recognize any other data item to which it is related. Such information provides a later method for recognizing and sequentially assembling selected related data items, both that exist within the original data item collection and that traverse two or more data item collections that have been organized and stored together. According to the present invention, non-hierarchical form means that the data items in a collection are organized with no priority among themselves or among those of any other collection they are stored with. Non-hierarchical form further means that organization of data items in a collection is achieved without knowledge of or the establishment of relationships with data items either in previously organized and stored collections or with data items in collections that will be organized and stored together at a later time. Other methods heretofore have required the organization and structure of all such relationships to be recognized, understood and defined at the time of storage.

According to the method of the present invention, all data items from any number of collections, once neutrally and non-hierarchically organized and stored together, coexist in parallel together, at once completely distinguishable and separately accessible, yet at the same time totally aware of and sequentially relatable to all other data items to which they are related. One can thus find any sequentially related subset of data items from among the total assembly of those present in storage without having knowledge of or having taken any steps to establish the sequential relationships with any such data items during the organization and storage of data items.

According to the present invention, the universe of a particular scope of information to be modeled in a computer can be represented as a collection consisting of any number of distinct scope segments. Each of these scope segments (referred to later as instance segments) can be modeled separately as individual sets of information.

By formulating such individual scope segment models and their corresponding sets of information in a particular and neutral form, individual segments from the same universal scope of information are automatically and dynamically linked. Through their dynamic linking these independent scope segment models and their corresponding sets of information function as the equivalent of a single model and set of information for the universe of a particular scope of information. The neutral expression of the information also provides an effective medium for the transfer of information.

The method of the present invention that accomplishes these advantages is comprised of three components: (1) the organization of information into instances; (2) expression of instances of information in neutral form; and (3) universal typing and neutral form expression of data items in an instance together with the dictionary information on all properties of those data items for their transfer.

An individual instance is characterized as one segment of the universe of that instance. Individual segments of the same universe of an instance are intrinsically and dynamically related. A computer environment is characterized as a universe of instances, typically spanning more than one instance universe. Certain portions of instances from different instance universes are intrinsically dynamically linked if they share certain information in common. The neutral form expression of instance information is employed to store individual data items in a set of information in a manner which at once enables both the complete isolation and therefore parallel processing of individual instance data sets and the recognition and exploitation of intrinsic dynamic links between different instance data sets for the retrieval and transfer of targeted subsets of diverse sets of information stored in the same computer environment.

The neutral form is achieved by assigning to each data item in an instance data set a generalized structural tag comprised of three elements: (1) a data type; (2) a data reference; and (3) a data organization. The meanings and properties of each of these elements for a particular data item are stored in a dictionary system whose frame of reference is the computer environment within which the data item in neutral form is stored.

All structural tags employed within a specific computer environment are defined to be internally consistent and unambiguous. The properties of the structural tag and the discrete data values stored in neutral form are used to target and retrieve instance data sets from an environment. Universal data typing is defined to enable simultaneous storage of the data items from both retrieved instance data sets and from the associated records in the environment's dictionary system.

Using this universal data typing, dictionary data items employed in the structural tags of retrieved data items are themselves expressed in neutral form. The combination of retrieved instance data items and their associated dictionary data items, now expressed in the same neutral form, are combined into a single file for transfer. This transfer file is completely self contained, including not only the data items of interest, but all of the information required to understand and to interpret each such data item.

The method of the present invention not only overcomes the limitations of the prior art but it also introduces significant capabilities heretofore not available. It establishes a method for organizing and storing sets of information which achieves:

relationships between an unlimited number of different and individually modeled instance segments of the same universe of an instance as well between certain instance segments of different instance universes without a-priori knowledge of their existence, their structure or their content and without remodeling or reloading of the data for individual instance segments;

practical form for expression of individual instance data items in the instance data sets that comprise an instance segment which

contains all of the information required to understand and interpret each data item; and

yet is independent both of the source and the ultimate application of the information;

independence of the individual instance data sets of an instance segment which enables complete parallel processing of information for storage as well as for retrieval and transfer;

practical non-hierarchical and non-integrated format for the structuring of instance data sets that comprise an instance segment which

is independent of the size and complexity of a set of information;

yet accounts for all relevant relationships among data items comprising each instance data set; and

allows dynamic versioning without loss of connection to information based on a previous version of data structure;

reconciliation of independently defined instance structures and instance data sets, enabling the direct co-mingling and integration of unlimited numbers of different sets of information;

reconciliation of different legacy instance data set structures and data values with reduced needs for manual cleansing and reorganization;

parallel processor organization and storage of individual instance data sets across diverse sets of information encompassing different instance universes and different instance segments of any particular instance universe;

parallel processor retrieval of all or any portion of an unlimited number of different but related sets of information without any requirement for a-priori knowledge of the formats, structures, or contents or relationships among the various sets of information;

transfer of any retrieved information complete with all the properties and relationships required to understand and interpret the information without any requirement for a-priori knowledge of its format, structure, or content.

BRIEF DESCRIPTION OF DRAWINGS

FIG. 1 is a schematic diagram illustrating various types of prior art techniques and their characteristics.

FIG. 2 is a schematic diagram of alternative forms of structuring of data relationships.

FIG. 3 is a schematic diagram of data organization according to the present invention.

FIG. 4 is another schematic diagram of data organization according to the present invention.

FIG. 5 is a schematic diagram of possible relationships between data according the present invention.

FIG. 6 is a schematic diagram of neutral form expression of data organized according to the present invention.

FIG. 7 is a further schematic diagram of neutral form data organization.

FIG. 8 is a schematic diagram of organization of data dictionaries of meanings and properties for data organization according to the present invention.

FIG. 9 is a schematic diagram of universal data typing with the present invention.

FIG. 10 is a schematic diagram in summary form of a data storage, retrieval and transfer process according to the present invention.

FIGS. 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25 and 26 are more detailed illustrative diagrams of data modeling portions of the diagram of FIG. 10.

FIGS. 27, 28, 29, 30, 31, 32, 33, 34, 35, 36 and 37 are more detailed diagrams illustrative of data storage and retrieval portions of the diagram of FIG. 10.

FIG. 38 is a schematic flowsheet diagram of implementation of the present invention in a computer system.

FIGS. 39, 40, 41, 42, 43 and 44 are more detailed diagrams of the flowsheet of FIG. 38.

FIGS. 45, 46, 47, 48, 49, 50 and 51 are diagrams supporting examples of the preferred embodiment of the present invention.

DESCRIPTION OF PREFERRED EMBODIMENT

1.0 Introduction

The present invention provides a neutral and non-hierarchical means in a computer based environment for organizing a collection of data items into sets of information for exchange, storage, integration and retrieval that simultaneously:

(a) avoids any requirement for in-advance knowledge and codification

of other collections of data to which said collection is to be related for integration and

of the specific relationships with other collections of data which are accessible for integration; and

(b) insures without anticipation the direct integration and expression of all valid relationships that exist with any other collections of data that have been organized and stored by the same means.

According to the method of the present invention, a target collection of data to be so organized is first segregated into consistent groups, one for each invention-defined instance universe contained in the raw collection of data items. Each resulting group of consistent data items is then treated as a non-hierarchical collection or cluster of nodes, i.e., one node per data item, each node having no particular physical ordering relative to another and the collection of nodes having no a-priori knowledge of or physical ties to any other collection of nodes. According to the method of the present invention, nodes in such a consistent collection or cluster are then each assigned a neutral form tag. This tag is comprised of the set of properties required to fully identify, define and distinguish each node from all others in that collection or cluster. The tag is referred to as a neutral form tag because:

(a) the structure of its components are invariant with data item and collection,

(b) the contents of each component are rigorously defined and stored in a comprehensive dictionary system,

(c) the contents of each dictionary entry are reconciled against the set of dictionary entries that has accumulated from the creation and assignment of tags for all consistent data collections or clusters that previously have been organized and stored in the same computer-based environment, and

(d) the tags and their dictionary information remain associated with a collection or cluster of data items when they are transferred from one computer-based environment to another.

Finally, with the present invention, data items in two separate collections are related when nodes in the two clusters have certain neutral form tag properties in common. Through such commonalities in certain neutral tag information, complex data relationships are made available among individual collections of neutrally organized collections of data without the requirement for any anticipation of either related data collections or the specific basis for each such relationship during the organization and storage of each individual data collection.According to the present invention, it has been found that information can be modeled by organizing various related data values into groups called instances. Any universe of related data items, i.e., the universe of an instance, can be subdivided and organized into a series of instance segments, each having a collection of instance data sets, one for every occurrence of that instance segment. It has been further found that by modeling and storing the instance data sets of individual instance segments in a particular non-hierarchical, non-integrated neutral form, both the models and their data in neutral form can be combined without remodeling or reloading to form a larger portion of the universe of an instance. According to the present invention, this neutral form at once also gives rise to greatly simplified modeling techniques, parallel processing in both data storage and retrieval operations, and simplified transfer of all or portions of the stored data.

According to the method of the present invention, an instance segment is structured in non-hierarchical, non-integrated form as a cluster consisting of non-hierarchical data reference nodes, one for each data value in an instance data set. Each data value in an instance data set, also referred to as an instance cluster, is expressed in neutral form by assigning to it a generalized structural tag that is formed of three elements which, taken as a whole, contain all of the definitions and properties required to describe that data value. These definitions and properties are recorded in a dictionary system according to the method of the invention. One such dictionary system is associated with the contents of each neutral form file. Multiple instance segments in the same neutral file use the same dictionary system. The first element in the structural tag is data type, which specifies the physical form (length, format) of the data value. The second element is data reference, specifying the name and properties (definition, units, upper and lower limits, constraints, etc.) of the data value. The third element is data organization, specifying the other data references, their properties as a group, and their respective data values in relation to the tagged data value.

According to the method of the present invention, a neutral form file can contain instance segments from different instance universes. Furthermore, in the method of the present invention relationships exist between all or certain portions of cluster instances in a neutral form file to the extent that two cluster instances share certain data references and their respective data values in common. These relationships are deemed dynamic, because they happen automatically and as a function of the cluster instances that are present at any one time in the neutral form file.

2.0 Nomenclature

For ease of understanding, the terms used in the subsequent portions of the description of the present invention and their meanings are set forth alphabetically below:

    Term              Meaning
    Alias             One dictionary entry that is equivalent to another.
    Association       A name that is closely associated with a dictionary
                      entry name.
    Author            User of the method of the present invention
                      who creates entries in the dictionary system.
    Author ID         Identification number unique to the world of users
                      of the method of the present invention which insures
                      discrimination of authors in all data transfers
                      between two parties.
    Cluster           Non-hierarchical collection of data reference nodes
                      which define an instance segment.
    Cluster - Associated An autonomous cluster that has been formed from a
                      comprehensive initial cluster.
    Cluster - Compre- Single cluster formed during the initial creation of
    hensive Initial   the data model for a set of information which con-
                      tains all of the known data references associated
                      with that set of information.
    Cluster - Group   A heading within a cluster which groups, e.g.,
                      logically, a collection of data references.
    Cluster - Instance An occurrence of the data items for a cluster
    Cluster - Nested  Cluster inserted into another cluster in a one-to-one
                      relationship such that
                      the collection of instance data sets of the nested
                      cluster are accessible from within the other cluster;
                      and
                      one and only one of the nested instance data sets is
                      selected to function as a part of the instance data
                      set of the other cluster.
    Cluster - Node    One data reference in a collection comprising a
                      cluster.
    Cluster - Sub-cluster Collection of data references in a cluster which are
                      to appear together in a data input form.
    Cluster - Subgroup A subheading within a cluster which groups, e.g.,
                      logically, a collection of data references under
                      a cluster group heading.
    Data              One or more data values or data items
    Data Instance Node A node on an instance cluster corresponding to one
                      data value in an instance data set.
    Data Item         One data value in a set of information
    Data Organization Combination of the cluster instance structure and
                      the cluster instance data set properties associated
                      with any data value.
    Data Organization The set of data references and data values for one
    Instance Data Set cluster instance data set.
    Data Organization The collection of data references and properties
    Structure         comprising a cluster to which a data item belongs.
    Data Reference    Name, description and definition of certain
                      properties of a data value
    Data Reference -  Descriptive name for a data reference
    Name
    Data Reference    Discriminator of the context within which a data
    - Schem           reference name is to be applied.
    Data Reference Set A subset of data references in a cluster whose sets
    Prime or Distinct of data values are prime/distinct data value sets.
    Data Reference Set The prime/distinct data reference set in a cluster
    - Basi Prime/     whose data values are basis prime/distinct data value
    Distinct          sets.
    Data Reference Set Any prime/distinct data reference set in a cluster can
    Prime/Distinct to be expressed as the basis prime/distinct data ref-
    Basis Prime/Distinct erence set by removing those data references from
    Conversio         the cluster that have a high granularity relative to
                      the new basis prime or distinct data reference set.
    Data Reference Set- Data references comprising a subset of the data ref-
    Subordinate       erences in the universe of an instance., i.e., the data
                      references for one segment of the universe of an
                      instance.
    Data Type         Physical form property of a data value, e.g.,
                      numeric, alphanumeric, photo, drawing, sound,
                      video, time series, geo-spatial, etc.
    Data Value        One instance of a data reference.
    Basis Prime/Distinct A.sub.i = Subset, {p.sub.i . . . p.sub.n }A.sub.i, of
     each Instance Segment
    (P/D Data Value Set whose values define a unique set, {p.sub.i . . .
     p.sub.n }A, in
                      U.sub.A
    Basis Prime Data  A basis P/D data value set for UI.sub.Ai which resides
    Value Set         in only one universe of an instance UI.sub.Ai of universe
                      U.sub.A.
    Basis Distinct Data a basis P/D data value set for UI.sub.Ai which can
     reside
    Value Set         in more than one universe of an instance UI.sub.Aj of
                      universe U.sub.A.
    Prime/Distinct (P/D) A basis P/D data value set for an Instance Segment
    Dat Value Set     of UI.sub.Bi which resides in an Instance Segment of
     UI.sub.Ai
    Dictionary - Condi- Entries made into the dictionary system of an envi-
    tional Entries    ronment during the modeling of an instance, which
                      entries have not been reconciled with existing
                      dictionary entries.
    Dictionary - Envi- Collection of dictionaries which list and define all
    ronmen System     of the properties of the structural tags that mark
                      instance data sets stored in an environment.
    Domain            Collection of one or more clusters.
    Dynamic Link      Union of two or more instance segments of the
                      same or different universes for the expression of
                      relationships between query specified x.sub.i 's from two
                      different but related instance segments where any
                      P/D data value sets in common between the instance
                      segments match.
    Environment       The collection of those different universes, U.sub.A,
                      whose instance data sets are recorded in neutral
                      form which is defined by a single consistent dic-
                      tionary system.
    Event             An instance.
    Glossary          Collection of properties for entries in a dictionary.
    Granularity       The relative level of specificity reflected in one or
                      a set of data references as it relates to another data
                      reference (not to itself) or set of data references.
                      For example: social security number has a high
                      specificity while a date/time stamp has a low
                      specificity; Last name has a high granularity relative
                      to zip code.
    Granularity -     Ratio of the granularity of one data reference (or set
    Relative          of data references) to another.
    Granularity -     Relative granularity of greater than one.
    Relative High
    Granularity -     Relative granularity equal to or less than one.
    Relative Low
    Inputting         Entry of a single sub-cluster of data items or a
                      complete cluster instance data set into neutral form.
    Instance - An     An occurrence of a collection of related data values,
                      all expressed on the same time basis.
    Instance Data Set Set of data values for one occurrence of an instance
                      segment.
    Instance Segment  Unique subset of UI.sub.Ai data values, {x.sub.i . . .
     x.sub.n }.sub.Ai,
                      spanning a single event or instance i within which
                      is contained a basis P/D data value set.
    Instance Segment  A subset of an Instance Segment of UI.sub.Ai which
    Nested            constitutes an Instance Segment of UI.sub.Bi.
    Instance Structure Collection of data references comprising an instance
                      segment.
    Kid               Identifier for a data reference name
    Lid               Identifier for author of an element of a model.
    Loading           Multiple sets of inputting performed in bulk.
    Model             One cluster or two or more non-integrated but
                      dynamically linked clusters which define the rela-
                      tionships and properties of a set of information.
    Neutral Form      Data values for an instance data set, each marked
                      with a structural tag.
    Node              One data reference in a collection that comprises a
                      cluster.
    Node ID           Identifier for a node in a cluster, e.g., the combina-
                      tion of Kid, Sid, Lid.
    Primary           The preferred one among a collection of aliases of
                      data reference names or schema names.
    Properties        Characteristics of a piece of data, a group of data,
                      or a structure of a group of data, e.g., name, ID,
                      definition, units, etc.
    Reconciliation - Dat Process for evaluating new data values as (a) being
    Values            the same as; (b) being equivalent to; or (c) being
                      different from one of the discrete data values already
                      in the environment for a particular data reference.
    Reconciliation - Dat Process of determining whether or not one data
    References        reference is an alias of another.
    Reconciliation -  Process for evaluating dictionary entries external
    Dictionary        to an environment as (a) matching with, or (b) being
                      equivalent to, or (c) being uniquely different from
                      an existing entry in an environment system
                      dictionary.
    Schema            See "Data Reference - Schema"
    Set of Information Collection of data values comprising one or more
                      occurrences of an instance segment together with
                      sufficient properties of those data values to enable
                      their understanding, interpretation and assignment
                      to discrete instance data sets with an arbitrary but
                      definable pattern to the ordering of the data values
                      in each instance data set.
    Sid               Identifier for data reference schema name
    Structural Tag    Three generalized attributes of a data value, viz.,
                      type, name, and organization, whose definition and
                      properties completely and unambiguously define a
                      data value within a cluster instance.
    Synonym           A name that has the same meaning as that of a
                      dictionary entry name.
    Time Basis        Frame of reference of time for either an instance
                      segment or an individual data reference in an in-
                      stance segment where null time is treated as "event".
    Transfer Form     Combination of one or more instance data sets and
                      their supporting dictionary information, all written
                      in neutral form and combined.
    Transfer Form -   One or more Instance data sets written in neutral
    Data              form.
    Transfer Form -   Dictionary entries employed by one or more in-
    Dictionaries      stance data sets written in neutral form.
    Union - Different All of the data values in two or more instance seg-
    Universes         ments are directly related to each other where their
                      basis P/D data value sets are the same.
    Union - Same      Union of (a) same universe instance segments UI.sub.B1
    Universes         and UI.sub.B2 and (b) a different universe instance seg-
                      ment UI.sub.A1 into which instance segment of U.sub.B1 or
                      UI.sub.B2 is nested where an x.sub.i in the query
     specification
                      resides exclusively in the instance segment of UI.sub.A1.
    Universe of A     Universe, U.sub.A, of all events of A expressed on the
                      same time basis, e.g., where A - Social Security
                      Number or A = Reactor Name, expressed as
                      instances.
    Universe of an    Universe of all events of A expressed on the same
    Instance o A      time basis, e.g., where A = Social Security Number
                      or Reactor Name, expressed as instances.
    Version           Designator which allows for the discrimination of
                      an existing cluster which has had one or more
                      instance data nodes added or removed.