Data With Direction

Notes

Abstract

NOTER_PAGE: (4 . 0.100549)

Résumé

NOTER_PAGE: (5 . 0.100549)

Synthèse de la thèse en français

NOTER_PAGE: (6 . 0.100549)

Chapter 1: Introduction

NOTER_PAGE: (48 . 0.100549)

1.1 Objective

NOTER_PAGE: (48 . 0.147467)

This research fills a gap in project management theory and practise, which concerns how a project stakeholder is presumed to discover and obtain factual knowledge of the significant rules

NOTER_PAGE: (48 . 0.17752808988764046)

general- purpose solution to the problem of communicating which way is 'forward' when orienting decisions

NOTER_PAGE: (48 . 0.7887640449438202)

common efficient way to communicate obligation, permission or encouragement from rule-maker agents to rule-taker agents.

NOTER_PAGE: (49 . 0.5573033707865169)

any given agent may or may not be aware of certain rules, but would prefer to be notified about them. And the issuers of the rules may or may not know about any particular agent, but prefer to have a practical way of communicating with them.

NOTER_PAGE: (50 . 0.20749063670411985)

1.2 Problem Statement

NOTER_PAGE: (50 . 0.520497)

1.3 Structure of This Dissertation

NOTER_PAGE: (52 . 0.801472)

re-discovery and re-formulation of tabular declarative computing methods

NOTER_PAGE: (53 . 0.4449438202247191)

quaternary system with separate semantics on the input and output sides, instead of the widely known binary ‘truth table’

NOTER_PAGE: (53 . 0.50187265917603)

Chapter 2: Methodology

NOTER_PAGE: (54 . 0.100549)

2.1 Purpose and Methodology of a DBA versus a PhD

NOTER_PAGE: (54 . 0.147467)

This ‘pragmaticist’ tradition was captured by David Clark, the first chair of the Internet Architecture Board in the adage: "rough consensus and running code"

NOTER_PAGE: (54 . 0.5707865168539326)

2.2 Methodology of Middle Range Theory

NOTER_PAGE: (55 . 0.441864)

2.3 Design Research

NOTER_PAGE: (57 . 0.660182)

2.4 Design Success Criteria

NOTER_PAGE: (62 . 0.100549)

2.5 Design Virtues and Norms

NOTER_PAGE: (64 . 0.398925)

2.5.1 Design Virtues

NOTER_PAGE: (65 . 0.100549)

2.5.1.1 Human-Centred Automation

NOTER_PAGE: (65 . 0.131372)

NOTER_PAGE: (65 . 0.35805243445692886)

2.5.1.2 Free/Libre/Open Relationships

NOTER_PAGE: (66 . 0.133035)

2.5.1.3 Tolerance

NOTER_PAGE: (66 . 0.416267)

2.5.1.4 Interoperability

NOTER_PAGE: (67 . 0.18708)

2.5.2 Design Norms

NOTER_PAGE: (68 . 0.100549)

2.5.2.1 Simplicity

NOTER_PAGE: (68 . 0.131372)

2.5.2.2 Modularity

NOTER_PAGE: (69 . 0.493177)

2.5.2.3 Intuitiveness

NOTER_PAGE: (69 . 0.730085)

2.5.2.4 Decentralization

NOTER_PAGE: (70 . 0.186783)

NOTER_PAGE: (70 . 0.5797752808988764)

2.5.2.5 Least Power

NOTER_PAGE: (70 . 0.815725)

NOTER_PAGE: (70 . 0.8247191011235956)

2.5.2.6 Tabular Declarative Style

NOTER_PAGE: (72 . 0.100549)

NOTER_PAGE: (72 . 0.5632958801498128)

Chapter 3: A Reflective Review of Literature on the Nature of a ‘Rule’

NOTER_PAGE: (74 . 0.100549)

3.1 What is a Rule?

NOTER_PAGE: (74 . 0.147467)

A rule is any directional relation communicated among two or more people to associate what ‘is’ and what ‘ought’ to be.

NOTER_PAGE: (74 . 0.17752808988764046)

3.2 What is an Algorithm?

NOTER_PAGE: (80 . 0.144972)

3.3 What is Agency?

NOTER_PAGE: (84 . 0.692669)

3.3.1 Direction from ‘Is’ to ‘Ought’

NOTER_PAGE: (84 . 0.729253)

3.3.2 Agency and Prerogative

NOTER_PAGE: (85 . 0.424106)

3.3.3 Source, Subjectivity and Strength

NOTER_PAGE: (88 . 0.421671)

3.4 Rule Transmission Systems

NOTER_PAGE: (90 . 0.23281)

3.4.1 Signal and Noise in Rule Transmission Systems: Insights from Information Theory

NOTER_PAGE: (90 . 0.269395)

uncertainty about the communication channel itself.

NOTER_PAGE: (91 . 0.8217228464419476)

Jaynes also critiques Shannon’s premise that there can be a set of possibilities with known probabilities of occurrence. That clearly requires considerable prior knowledge.

NOTER_PAGE: (92 . 0.16329588014981275)

3.4.2 Three Postulates for Optimal Rule Transmission Systems

NOTER_PAGE: (96 . 0.275037)

3.4.3 Considering Rules Transmission Systems Technology in Light of Systems Ecology Theory

NOTER_PAGE: (97 . 0.293032)

Chapter 4: Available Methods Review Relating to Rule Systems Design

NOTER_PAGE: (102 . 0.100549)

4.1 Available Methods for Expressing Rule Logic

NOTER_PAGE: (102 . 0.147467)

simplest form of rules-as-prose appears to be RuleSpeak, which reduces rules to declarative statements that are concise, consistent and unambiguous

NOTER_PAGE: (102 . 0.597752808988764)

guideline incorporating a set of ‘best practices’ conformant to the OMG's SBVR standard

NOTER_PAGE: (102 . 0.6449438202247191)

There are many ‘successful’ rules-as-code proof-of-concept projects, very few that are able to scale to tens or hundreds of thousands of rules

NOTER_PAGE: (103 . 0.33333333333333337)

descend from a successful pilot into infamy as the “incomprehensible failure”.

NOTER_PAGE: (103 . 0.42921348314606744)

can be read as ‘information’ by a human directly from a user interface without any specialized expressions or markup.

NOTER_PAGE: (103 . 0.5534495830174374)

4.1.1 Unstructured Natural Language Expression

NOTER_PAGE: (103 . 0.617243)

4.1.2 RuleSpeak (OMG, 2016c)

NOTER_PAGE: (103 . 0.780329)

4.1.3 Flowchart (Decision Tree)

NOTER_PAGE: (104 . 0.100549)

4.1.4 Procedural Imperative Programming Code e.g. PASCAL (Wirth, 1976)

NOTER_PAGE: (104 . 0.447209)

4.1.5 RuleML, OASIS (Boley, 2006)

NOTER_PAGE: (104 . 0.630012)

4.1.6 Notation3 (Berners-Lee & Connolly, 2008)

NOTER_PAGE: (105 . 0.100549)

4.1.7 Decision Model and Notation, DMN (OMG, 2019a)

NOTER_PAGE: (105 . 0.364122)

4.1.8 Truth Table (Input/Output Binary Decision Table)

NOTER_PAGE: (105 . 0.617243)

4.2 Available Methods for Logic Data Models

NOTER_PAGE: (106 . 0.47928)

The choice between numbers or letters may seem inconsequential, but logic tables with {0,1} or {-1, 0, 1} are precise, whereas there is an intrinsic ambiguity to systems that employ {T,F}. ‘True’ and ‘false’ may be interpreted in a variety of ways, and so-called ‘truth tables’ have multiple origins, styles and meanings for particular contexts

NOTER_PAGE: (107 . 0.42608036391205456)

This is not an abstract digression. Logical contradictions do occasionally arise in real-world law, business and policy.

NOTER_PAGE: (109 . 0.3525398028809704)

fish defined in section 45, as a term of art, is not limited solely to aquatic species. Accordingly, a terrestrial invertebrate, like each of the four bumble bee species, may be listed as an endangered or threatened species under the Act.

NOTER_PAGE: (109 . 0.5852918877937832)

acknowledge that contradiction is sometimes a genuine persistent state, and proceed with it as such.

NOTER_PAGE: (109 . 0.7043214556482184)

For our particular purpose in the DWDS, we borrow certain quaternary logic concepts from catuṣkoṭi metaphysics, Turing’s universal machine, and molecular computing

NOTER_PAGE: (110 . 0.6557998483699773)

4.3 Available Methods for Rule Logic Processing: Distinguishing the Present Design Objective

NOTER_PAGE: (111 . 0.26791)

4.3.1 Distinguishing DWDS from Programmable Logic Controller (PLC) Systems

NOTER_PAGE: (111 . 0.524298)

4.3.2 Distinguishing DWDS from a Rules Engine

NOTER_PAGE: (113 . 0.380811)

Our design represents a different sort of pursuit: a general-purpose specification for communicating rules as data, in a manner that that is equivalently usable by any application

NOTER_PAGE: (113 . 0.5974222896133434)

Control remains at the edge.

NOTER_PAGE: (113 . 0.7035633055344959)

4.3.3 Distinguishing DWDS from a Decision-Support System

NOTER_PAGE: (114 . 0.100549)

4.3.4 Distinguishing DWDS from Artificial Intelligence

NOTER_PAGE: (114 . 0.507787)

4.3.5 Distinguishing DWDS from Business Process Workflow

NOTER_PAGE: (115 . 0.670813)

4.3.6 Distinguishing DWDS’s Rule Schema from a Domain Specific Language

NOTER_PAGE: (116 . 0.361509)

We forget about trying to avoid or minimize the deductive search, and simply do it, employing a rather extreme form of parallelism to get the job done quickly.

NOTER_PAGE: (117 . 0.6800606520090978)

4.4 Influences and Inspirations from 70 Years of Methods in Programmable Logic

NOTER_PAGE: (118 . 0.39649)

4.4.1 Data Structuring and Transmission

NOTER_PAGE: (119 . 0.100549)

Project Cybersyn

NOTER_PAGE: (119 . 0.47839272175890823)

A further step was taken by Michael Genesereth and Richard Fikes who designed the Knowledge Interchange Format (KIF) to embed first-order logic directly into Web documents

NOTER_PAGE: (120 . 0.37149355572403336)

through this decade XML schema proliferation resulted in a complicated labyrinth of competing standards.

NOTER_PAGE: (120 . 0.579226686884003)

large number and diversity of XML schemas which had come to be designed and implemented ‘bottom-up’ by diverse communities led to redundancy and inconsistency for the Semantic Web as a whole

NOTER_PAGE: (120 . 0.6383623957543594)

The need for interoperability among XML schemas led to the Resource Description Framework (RDF)

NOTER_PAGE: (120 . 0.7005307050796058)

Several competing standards emerged for rules expression

NOTER_PAGE: (121 . 0.1068991660348749)

OASIS published RuleML

NOTER_PAGE: (121 . 0.1463229719484458)

Object Management Group (OMG) published Semantics of Business Vocabulary and Business Rules (SBVR)

NOTER_PAGE: (121 . 0.18726307808946172)

produced Decision Model and Notation (DMN) through the OMG to express logic rules for Business Process Model and Notation (BPMN)

NOTER_PAGE: (121 . 0.25094768764215314)

The ISO has published and maintained several relevant standards.

NOTER_PAGE: (121 . 0.28658074298711145)

designed MetaLex as a proposed standard for jurisdiction-neutral, language-neutral XML encoding of legislation, as well as the Legal Knowledge Interchange Format (LKIF)

NOTER_PAGE: (121 . 0.42077331311599697)

majority of this has involved procedural imperative rules-as-code libraries for particular application environments

NOTER_PAGE: (121 . 0.5064442759666414)

With every BRMS vendor using its own proprietary representation for rule execution, and their own syntax for expressing rules across various metaphors, this task became increasingly difficult.

NOTER_PAGE: (122 . 0.12964366944655042)

ow we see yet another emerging standard for the new world of decision management – the Decision Management Notation (DMN).

NOTER_PAGE: (122 . 0.1865049279757392)

4.4.2 Tabular Logic Programming

NOTER_PAGE: (122 . 0.695282)

One of the most prominent industry implementations of tabular declarative programming was on mainframes for global banking, financial services and industrial organizations

NOTER_PAGE: (123 . 0.601213040181956)

They combined the methods of Datalog with input/output tables

NOTER_PAGE: (123 . 0.6338134950720242)

data modelling (goodness of fit) in academic venues, versus algorithmic modelling (decision trees/tables) in business and industry.

NOTER_PAGE: (124 . 0.26004548900682334)

4.4.3 Procedural Logic Programming

NOTER_PAGE: (124 . 0.687739)

collaborated to create Datalog as a simpler subset of Prolog, consisting solely of declarative facts and rules, without operational functions.

NOTER_PAGE: (126 . 0.601213040181956)

Procedural imperative programming found its niche in general market computing where programmers, as ‘software engineers’, are expected to rapidly show ‘good-enough’ prototypes for highly competitive milestone-driven clients.

NOTER_PAGE: (127 . 0.2403335860500379)

Tabular declarative programming found its niche in large-scale, complex industry use cases in which operational integrity is mission-critical (banking, nuclear control systems, insurance). In such scenarios, analysts and programmers, as applied ‘logicians’, are provided the time to understand and solve for whole system problems.

NOTER_PAGE: (127 . 0.3260045489006823)

4.4.4 Data Models for Quaternary Logic

NOTER_PAGE: (128 . 0.525188)

Chapter 5: DWDS Technical Rationale and Design Summary

NOTER_PAGE: (130 . 0.100549)

5.1 ‘Data With Direction’ from Concepts to a System Specification

NOTER_PAGE: (130 . 0.163858)

differentiates the imperative, declarative and empirical aspects of rules

NOTER_PAGE: (130 . 0.47990902198635327)

5.2 Reference Implementations Under Current Development in Software

NOTER_PAGE: (137 . 0.067351)

5.3 Methods for High Performance Decentralized Distributed Computing

NOTER_PAGE: (138 . 0.675743)

5.3.1 Computing Fast and Slow: Externalize Computational Work from Run-Time

NOTER_PAGE: (138 . 0.699202)

Richard Feynman characterized the difference between these two data processing styles: “This often goes under the name of artificial intelligence, but I don't like that name. Perhaps the unintelligent machines can do even better than the intelligent ones."

NOTER_PAGE: (139 . 0.4495830174374526)

5.3.2 Transforming Complex Natural Language to Simple Structured Natural Language

NOTER_PAGE: (141 . 0.212083)

To convert unstructured free-form natural language into uniformly-structured natural language requires the capability to compose clear sentences. This can seem obvious to the point of condescending

NOTER_PAGE: (141 . 0.33661865049279754)

DWDS RuleSchema specifies a limited phrase-structure grammar with just one declarative sentence type. Despite this very rigid syntactic constraint, there are no boundaries on semantic scope. This is inverse to the more common Semantic Web technique of supporting complex expression using rigid semantic schemas (e.g. RuleML) and tolerant syntactic structures (e.g. SGML). The two approaches are not mutually exclusive; they are complementary and can be deployed concurrently.

NOTER_PAGE: (141 . 0.5261561789234268)

The DWDS leaves the management of semantics in the hands of people who have the prerogative, motivation, domain knowledge and socio-cultural familiarity to tailor the expression of each sentence of each rule, and who are motivated to make a genuine effort to provide a faithful reproduction of the full normative intent of the original rule, with minimal distortion.

NOTER_PAGE: (141 . 0.6421531463229719)

DWDS Rule Schema (RS) is a non-executable syntactic specification for tuple-oriented data containing normative propositions between rule-maker agents and rule-taker agents across a network. It packages this with with metadata and optional descriptive data, to provide rule-as-data packages.. This is less than a ‘domain-specific language’ (DSL). The syntactic structure of sentences expressible in a DWDS logic table is constrained, but there are no constraints on the semantic scope.

NOTER_PAGE: (142 . 0.1607278241091736)

5.3.3 Externalize Linguistic Complexity from Rule Structure, to Simplify Function

NOTER_PAGE: (142 . 0.495256)

explicit decomposition of syntactic structure and logic structure from semantic expression.

5.3.3.1 Making the Logic Relations Explicit

NOTER_PAGE: (142 . 0.694154)

5.3.3.2 Making the Syntactic Elements Explicit

NOTER_PAGE: (143 . 0.237443)

NOTER_PAGE: (143 . 0.40561031084154664)

NOTER_PAGE: (143 . 0.7816527672479151)

NOTER_PAGE: (144 . 0.11296436694465505)

NOTER_PAGE: (144 . 0.5231235784685367)

5.3.3.3 Making the Syntactic Structure Explicit

NOTER_PAGE: (145 . 0.622113)

NOTER_PAGE: (149 . 0.2714177407126611)

NOTER_PAGE: (152 . 0.5170583775587566)

5.3.3.4 Making Rules Easily Readable and Efficiently Computable

NOTER_PAGE: (142 . 0.5231235784685367)

5.3.4 Externalize Engagement of Semantic Web Standards to Rule Makers and Rule Takers

NOTER_PAGE: (155 . 0.58351)

5.3.5 Externalize Computability by Requiring Rule Expression to be NOT Turing-Complete

NOTER_PAGE: (157 . 0.100549)

A rule may take time to compute, but inspection can validate that it will halt.

NOTER_PAGE: (157 . 0.27748294162244125)

it is acceptable within the DWDS RuleSchema specification for a rule author to include a pointer to one or more external references where the required procedural code can be obtained.

NOTER_PAGE: (158 . 0.18574677786201668)

5.3.6 Externalize Control Data and Logical Relations Data by Separating Data from Procedure

NOTER_PAGE: (158 . 0.372971)

[M]uch of the technical difficulty of this subject has to do with negotiating the transition between imperative statements (from which programs are constructed) and declarative statements (which can be used to deduce things).

NOTER_PAGE: (158 . 0.5064442759666414)

5.3.7 Externalize the Data Processing Burden with Purposeful Structuring of Data Into Tables

NOTER_PAGE: (160 . 0.712921)

5.3.7.1 Data Topology Overview

NOTER_PAGE: (160 . 0.750277)

the potential of a decentralized global network of computationally consistent tabular logic gates has yet to be realized. The DWDS is our contribution to this pursuit.

NOTER_PAGE: (161 . 0.7065959059893859)

5.3.7.2 Cartesian Product Topology (DWDS Lookup Tables)

NOTER_PAGE: (162 . 0.524416)

5.3.7.3 Vertical Stack Topology (DWDS Logic Gates)

NOTER_PAGE: (164 . 0.388472)

5.3.7.4 Horizontal Tape Topology (DWDS ‘RuleReserve’ Tables)

NOTER_PAGE: (164 . 0.683523)

5.3.8 Externalize Reusable Algorithms (In-Memory Retrieval of Cartesian Product Tables)

NOTER_PAGE: (166 . 0.324271)

5.3.9 Externalize Declarative Conditions and Assertions from Logical Relations

NOTER_PAGE: (168 . 0.34001)

5.3.9.1 Roles and Purposes

NOTER_PAGE: (168 . 0.370833)

5.3.9.2 The DWDS Normative Logic Gate

NOTER_PAGE: (172 . 0.315244)

5.3.9.3 Discussion: The ‘DWDS Logic Gate’ Differs from a ‘Decision Table’

NOTER_PAGE: (176 . 0.533562)

5.3.9.4 Discussion: ‘DWDS Logic Elements’ Differs from the ‘Wright / Ostrom School’

NOTER_PAGE: (177 . 0.661608)

5.3.9.5 Discussion: ‘DWDS Logic’ is Combined with Metadata to Find and Fetch

NOTER_PAGE: (179 . 0.549538)

5.4 Rules as Data

NOTER_PAGE: (180 . 0.746714)

5.4.1 Data Structure of [rule.dwd] Records

NOTER_PAGE: (180 . 0.783298)

5.4.2 Transmission Protocols for Data with Direction

NOTER_PAGE: (185 . 0.100549)

5.4.3 Identifiers for [rule.dwd] and [lookup.dwd] Resources

NOTER_PAGE: (186 . 0.194147)

5.4.4 Diagnostic ‘Rule 256’

NOTER_PAGE: (189 . 0.695757)

5.5 Data Sifting

NOTER_PAGE: (191 . 0.100549)

5.5.1 Feature Criteria vs Conjunction Criteria

NOTER_PAGE: (191 . 0.129472)

5.5.2 RuleReserve Uses [is.dwd] as a [sieve1.dwd] to find [rule.dwd]s ‘In Effect’ and ‘Applicable’

NOTER_PAGE: (193 . 0.578877)

5.5.3 RuleTaker Uses [is.dwd] and [rule.dwd], Creating a [sieve2.dwd] to Obtain Assertions ‘Invoked’

NOTER_PAGE: (197 . 0.100549)

Chapter 6: Proof-of-Concept Reference Implementations

NOTER_PAGE: (208 . 0.100549)

6.1 Development of Operational Software Based on the DWDS Design

NOTER_PAGE: (208 . 0.147467)

6.2 RuleMaker at Version 3

NOTER_PAGE: (209 . 0.292498)

6.2.1 RuleMaker Overview

NOTER_PAGE: (209 . 0.329082)

6.2.2 Prior Implementations of RuleMaker

NOTER_PAGE: (210 . 0.100549)

6.2.3 RuleMaker Version 3.x Implementation Details

NOTER_PAGE: (210 . 0.629002)

6.2.4 RuleMaker Experiments

NOTER_PAGE: (214 . 0.100549)

6.3 RuleReserve and RuleTaker at Version 3

NOTER_PAGE: (219 . 0.741131)

6.3.1 RuleReserve and RuleTaker Overview

NOTER_PAGE: (219 . 0.777716)

6.3.2 Prior Implementation of RuleReserve and RuleTaker

NOTER_PAGE: (220 . 0.40427)

6.3.3 RuleReserve and RuleTaker Version 3.x Implementation Details

NOTER_PAGE: (221 . 0.16956)

6.3.4 RuleReserve and RuleTaker Experiments

NOTER_PAGE: (226 . 0.100549)

Chapter 7: Conclusion

NOTER_PAGE: (230 . 0.100549)

7.1 Purpose and Outcome

NOTER_PAGE: (230 . 0.147467)

7.2 Original Contributions and Useful Restorations

NOTER_PAGE: (234 . 0.181022)

7.3 Limitations of this Research

NOTER_PAGE: (241 . 0.100549)

7.4 Future Research

NOTER_PAGE: (241 . 0.717553)

7.4.1 Various Suggestions Received

NOTER_PAGE: (241 . 0.754138)

7.4.2 Ongoing Technical Methods Research

NOTER_PAGE: (242 . 0.603524)

7.4.3 Improved Computational Linguistics for Digital Transformation

NOTER_PAGE: (243 . 0.38378)

7.4.4 Common Sense-Making Through a Period of “Incommensurable Paradigms”

NOTER_PAGE: (244 . 0.645988)

Appendix A: Thesis Project Timeline

NOTER_PAGE: (247 . 0.100549)

Appendix B: The Geometry and Forces of a Bubble Cluster as an Applied Metaphor in the Design of Multi-Entity Projects

NOTER_PAGE: (248 . 0.100549)

Appendix C: External References to This Research

NOTER_PAGE: (255 . 0.100549)

Appendix D: Informal Comments by Technical Contributors

NOTER_PAGE: (262 . 0.100549)

Reference List

NOTER_PAGE: (274 . 0.100549)