Understanding Method

1y ago
12 Views
2 Downloads
2.33 MB
45 Pages
Last View : 1d ago
Last Download : 2m ago
Upload by : Brady Himes
Transcription

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Understanding Method(Abridged)Best practices based method for creating a process control mechanismthat consistently yields high-quality XBRL-based financial reportswhere the model can be “reshaped” or “altered” by report creatorsBy Charles Hoffman, CPA (Charles.Hoffman@me.com)March 24, 2021 (DRAFT)“I skate to where the puck is going to be, not where it has been.” Wayne Gretzky,legendary Canadian hockey starExecutive summary:1 This document explains, at a high level, a proven, reliable, best practice method forimplementing XBRL-based financial reporting following the forthcoming OMG StandardBusiness Report Model (SBRM). This method is specifically designed to address issues which come about when theextensibility features of XBRL are employed which allow report creators to “reshape” or“alter” or other such modifications. Report creator alterations must be controlled in order to maintain report quality, avoidingpotential contradictions and inconsistencies. This method has been rebranded as the Seattle Method1.Seattle Method, f1

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Copyright (full and complete release of copyright)All content of this document is placed in the public domain. I hereby waive all claim of copyright in thiswork. This work may be used, altered or unaltered, in any manner by anyone without attribution ornotice to me. To be clear, I am granting full permission to use any content in this work in any way youlike. I fully and completely release all my rights to any copyright on this content. If you feel likedistributing a copy of this work, you may do so without attribution or payment of any kind. All that said,attribution is appreciated should one feel so compelled. The copyrights of other works referenced bythis document are established by the referenced work.2

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication This document explains at a high level a proven standard method of implementing a standarddigital financial report using the XBRL technical syntax leveraging the extensibility features ofXBRL which follow the forthcoming OMG Standard Business Report Model (SBRM)2.The intent of this document is to summarize know-how. This know-how, when documented inthe form of a useful method, eliminates the need for others to re-invent the wheel. Rather thanre-inventing the wheel; others can simply leverage a well-thought-through, world-classapproach that has been designed, created, rigorously tested, and carefully engineeredleveraging approaches that have been proven to work effectively results.These best practice approaches and techniques that have been generally demonstrated assuperior to any known alternatives because the techniques produce results that are superior tothose achieved by other means or because it has become a standard way of doing things aredocumented in this resource. It is anticipated that others will likely improve upon this methodover time.This method provides a process control mechanism that, when followed, will consistently yieldhigh-quality XBRL-based digital financial reports. For full details of this method, please seeMethod of Implementing a Standard Digital Financial Report Using the XBRL Syntax3.To understand this method, it is critically important to understand certain specific backgroundinformation, so that is where this document starts. First, we explain what it takes to achieveeffective automation. To get started we will provide important understanding about howcomputers work and a basic grounding in artificial intelligence.Control Rules Effective Automation (High Quality)If a process cannot be controlled then the process simply cannot repeatedly and reliably outputhigh-quality. If process output is not high-quality, automation cannot possibly be effective.So, control of a process is necessary in order for the process to be effective. How do youcontrol a process? You control a process using rules. Manual processes are controlled by rulesthat are read by humans. Automated processes are controlled by rules that are readable byboth machines (i.e., to execute the process) and humans (i.e., to make sure the rules are right).Who creates these machine-readable rules that are used to control processes that yieldeffective automation? Accountants must create these rules because the rules tend to beaccounting oriented. Technical rules tend to relate to syntax and such technical rules can be2OMG, Standard Business Report Model (SBRM), https://omgwiki.org/SBRM/doku.phpMethod of Implementing a Standard Digital Financial Report Using the XBRL thod.pdf33

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication hidden from business professionals. What is left is the business logic and accounting rules thatare used to control information and control process workflow. As such, the creation ofmachine-readable rules must be “self-service”. Business professionals must be empowered tocreate, adjust, maintain, and otherwise manage the rules that are used to control and thereforeffectively automate processes. Once you have the machine-readable rules, you need softwarethat can process the rules; this is sometimes called a rules engine or reasoning engine or asemantic reasoner. We will get to that in a bit, but first let’s be sure you have some criticallyimportant background understanding.Computer Empathy and AI in a NutshellThe following is a brief summary of the document Computer Empathy4 which points out thatboth computers and specific aspects of accounting work per the rules of mathematics.If accountants can (1) improve their understanding of how computers work and (2) leveragethat understanding and represent some of their accounting knowledge in a more formalmachine-readable way, this will lead to accountants and technology providers having muchmore productive conversations and pave the way to computers being able to do some of theaccountant’s repetitive, mechanical, monotonous manual work.To understand how to get a computer to do work, it is important to understand the strengths ofcomputers and the obstacles that get in the way which we will highlight now along with a fewother important details.Strengths of ComputersComputers seem to perform magic. How computers do what they do tends to be a mystery tomany people. But computers are simple machines that follow very specific instructions; nomagic is involved. The strengths of computers can be summarised as follows. Computers can: store informationretrieve informationprocess stored informationmake information accessible to individuals or other machines or softwareObstacles – Communication & UnderstandingThe accounting profession is yet to fully leverage the strengths of computers mainly due to thefollowing general obstacles that tend to get in the way:4Computer Empathy, puterEmpathy.pdf4

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication https://creativecommons.org/publicdomain/zero/1.0/ accountants use different terminologies to refer to exactly the same thingaccountants differ in their understanding and interpretation of accounting standardsaccountants don’t understand technologies’ limitationsIT professionals use different technology stacks and languages to achieve the sameresultIT and business professionals have an oversimplified view of accountingComplexity and OrderDifference systems have different levels of complexity. Systems can also be ordered ordisordered. The Cynefin Framework5 is a conceptual framework that helps you understand thedynamics that are at work within different types of systems.The following graphic helps one understand the different levels of complexity: simple,complicated, complex, and chaotic. The graphic also helps one understand the differencebetween disorder and order.5Cynefin Framework, fin-framework.html5

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication The video Using Cynefin to Prioritize and Analyze Features, User Stories, and FunctionalRequirements6 provides an excellent walk through of these ideas. Another video, Complexity,Cynefin, and Agile7; provides additional useful insights related to understanding how to dealwith complexity.This method leverages “safe to fail” experimentation to understand complexity and to createthe necessary control mechanisms necessary to create XBRL-based digital financial reports thatare also provably properly functioning logical systems.Different skill sets are necessary to be able to create simple, complicated, and complex systemsthat work effectively.Data vs Information vs KnowledgeWe are working with information, not data. The difference between data and information isthat data is the raw facts and numbers where information is data in context. This is importantto understand as most problems faced by accountants are an information problem, rather thana data problem. Getting data is easy. Knowing what that data represents and how the data fitstogether is more challenging. Representing information in the form that a machine such as acomputer can understand and use that information is difficult and takes a skilled professional.Knowledge is a set of data and information and a combination of skill, know-how, experiencewhich can be used to improve the capacity to take action or support a decision making processby categorizing, collating, associating the data and information8.6YouTube.com, Using Cynefin to Prioritize and Analyze Features, User Stories, and Functional Requirements,https://www.youtube.com/watch?v L5fnxahydXM78YouTube.com, Complexity, Cynefin, and Agile, https://youtu.be/-F4enP8oBFMYouTube.com; Data, Information, Knowledge; https://youtu.be/3NxN0OgVN2k6

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Knowing what that data represents and how the data fits together is difficult. Representinginformation in the form that a machine such as a computer can understand and use thatinformation is difficult.StandardsStandards can help overcome the obstacles above but won’t eliminate them. Good examples ofstandards that have helped change the world are standard shipping containers, uniformproduct codes (barcodes), and standard electrical outlets. It is highly unlikely to get everyone toagree so providing options can be a good thing. XBRL is a global standard for business reportingand is an ontology-like thing (explained below) that can represent financial reports digitally.XBRL can be leveraged for automation of accounting, reporting, auditing and analysis processesand tasks. To do that, you use a knowledge based system.Knowledge Based SystemsThe better the capability of a system to represent knowledge, the better the ability for asoftware application to read and process that knowledge and perform useful work for the userof the system using that machine-readable knowledge. A dictionary would be a simple flat inventory of terms with no relations.A thesaurus would document some relations between broader and narrower terms.This is more useful than a simple dictionary.A taxonomy provides descriptions and a limited amount of structure generally in theform of one information hierarchy. This is more useful than a thesaurus.An ontology is a model that tends to provide formal descriptions and multiple structuresand therefore tends to have more than one hierarchy, e.g. a graph9.A logical theory is a set of models (ontology like things) that are consistent with thelogical theory. A logical theory provides a way of thinking about a domain by means ofdeductive reasoning to derive logical consequences of the theory.I have created a logical theory that describes the mechanical aspects and dynamics of afinancial report10. But to get a knowledge system to work, you have to put knowledge into thatsystem.9Wikipedia, Graph Theory, https://en.wikipedia.org/wiki/Graph theoryLogical Theory Describing Financial lTheoryDescribingFinancialReport.pdf107

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Logical theoryIn very simple terms, a logical theory is a set of models that are permissible per that logicaltheory. Those models are constructed by making logical statements which specify: Terms (things used by that model)Associations (relations between things) e.g. “type-subtype” of thing, structure “haspart”Structures (sets of associations between things)Rules (assertions that certain things and associations follow specific patterns)Facts (values that are described by terms, fit into structures, follow specific rules)World view (e.g. closed world assumption, unique name assumption and negation asfailure)Exchanging Information EffectivelyXBRL is a media11 for exchanging complicated/complex information in either human-readableor machine-readable form. For example, the general purpose financial report is a payload ofcomplex information12:That complex information, such as a general purpose financial report, is the payload in aninformation exchange:11Understanding the Role of XBRL, ncial Report Articulation, cheme/proof/referenceimplementation/PROOF Articulation.jpg8

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication The diagram above shows a general purpose financial report as a payload that is exchangedbetween an information bearer and an information receiver. Both the information bearer andreceiver share common background knowledge, common inference logic, and a common worldview.This system works because nothing is left to chance. A proven (fail-safe)documented theory, framework, and method document good practices. Clever software9

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication engineers leverage the theory, framework, and method models and metadata to makesoftware easy enough for business professionals to reliably perform the tasks and processesnecessary to do their work in new and more efficient ways.For an information exchange to be useful, the exchange must be reliable. To be reliable, it mustbe controllable. Rules are used to control the system.Creating the knowledge to store in the systemThere are two general approaches to creating knowledge to store in a knowledge base:1. Inductive reasoning: Let the computer work it out by using AI, machine learning orother approaches. This means, feed the computer a load of data and let it figure out thepatterns. (pattern-based, machine learning)2. Deductive reasoning: Tell the computer what the knowledge is. Accountants andauditors are highly trained and have the knowledge in their head. All we need is a wayof capturing that knowledge and storing it in an ‘ontology’ and a knowledge base ofrules. (rule-based, expert system)It is not an either-or question. But option 2 needs to be prioritised because it will provide thefoundation for AI and machine learning to build on. Machine learning excels where there is ahigh tolerance for error. There is an extremely low tolerance for error in financial accounting,reporting, auditing, and analysis.A knowledge based system draws upon the knowledge of human experts, i.e. accountants andauditors. The more knowledge in the knowledge base, the more the knowledge based systemcan do. The right information can literally supercharge what can be achieved.To understand the capabilities of a knowledge based system, it is important to understand thecomponents of such knowledge based systems.Components of a Knowledge Based SystemThis information is stored in a fact database and a knowledge base. The system applies problemsolving logic using a problem-solving method. The knowledge based system supplies anexplanation and justification mechanism to help users understand the line of reasoning used toreach conclusions. The system then presents that information back to the user.Nothing is a “black box”. The origin of information used to reach conclusions is alwaysapparent.10

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication The following is a summary of the components of a knowledge based system. Each of thecomponents shown in the graphic will be described and examples provided in the followingsections.Business Professional User InterfaceThe business professional user interface are the components that are exposed to the businessprofessional using the system. Business professionals need transparency as to the terms,associations, structures, rules, facts, line of reasoning, problem solving logic, problem solvingmethod, and the plausibility of all conclusions reached by the system.The following is one of a number of screen shots13 of the working proof of concept softwareapplication Pesseract which provides an example of a user interface with which a businessprofessional could likely interact:13Additional Pesseract User Interface Screenshots, https://photos.app.goo.gl/cWeZYaMBEbmSSm7v811

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication The user interface is non-technical requiring only business and accounting knowledge toeffectively understand the software application and how to use it.Justification and Explanation MechanismThe justification and explanation mechanisms of the software application explains and justifiesand provides transparency into how conclusions are reached by the software application. Therules used, facts used, line of reasoning, and origin of all facts are knowable to the businessuser of the software. There is transparency into all conclusions that are reached by thesoftware application. Nothing is a black box.Below you see the fundamental accounting concept relations continuity cross check verificationchecks provided by XBRL Cloud’s Evidence Package14 which is a review tool that can be used toverify XBRL-based financial reports:14XBRL Cloud Evidence Package, y.html12

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication If you look at the fundamental accounting concept relations continuity cross check verificationresults you see that the business user can trace each fact two it’s origin, understand all rulesused by the software to reach conclusions, etc.Pesseract provides similar functionality:XBRL Cloud’s Disclosure Mechanics and Reporting Checklist15 provides the rules used, line ofreasoning used, and conclusions reached for determining if a disclosure is structured consistentwith its expected specification:Disclosure mechanics rules:15XBRL Cloud Disclosure Mechanics and Reporting nd%20Reporting%20Checklist.html13

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Line of reasoning:14

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Conclusions reached:Similar functionality is offered by Pesseract:Disclosure mechanics rules:Line of reasoning:15

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication Conclusions reached16:Similar mechanisms exist for all other categories of rules verified using the method that hasbeen created which leverages OMG’s Standard Business Report Model (SBRM) 17:Reasoning, Inference, Rules EngineThe reasoning, inference, and rules engine use the machine-based rules, a line of reasoning forsolving problems using some problem solving logic and problem solving method (i.e. forwardchaining, backward chaining) to reach conclusions about facts and all other statements madewithin the logical system. This includes capabilities to logically derive or infer new facts orother information based on existing facts and rules. It also includes the capability to determineconsistency of facts with the systems knowledge base of rules.16Pesseract disclosure mechanics verification of 94.8% of all 124 disclosures ototype/Microsoft/Microsoft2017 Discovery.jpg17SBRM Progress Report, -progress-report.html16

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication As described by RuleML.org, there tends to be three primary problem solving paradigms whichare used to build a rules engine18:1. Knowledge Graphs (i.e. the W3C semantic web stack; RDF, N3, OWL, SHACL, SPARQL,RDF triple stores)2. Graph Databases (i.e. Neo4j and other labeled property graphs, Graph Query Languageor GQL, graph databases)3. Logic Programming (i.e. Prolog, SQL, relational databases)It is unlikely that every enterprise will use the same approach. This graphic shows how thesedifferent problem-solving paradigms relate to one another and the intersection or “sweet spot”between these paradigms19:XBRL-based financial reports are consciously architected such that they fit into the PSOA “sweetspot” which means that an XBRL-based financial report can be bidirectionally convertedbetween all three of these primary problem solving paradigms.18Primary Problem Solving Logic Paradigms, ary-problemsolving-logic-paradigms.html19Primary Problem Solving Logic Paradigms, ary-problem-solving-logicparadigms.html17

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication The following is a graphic which shows the structure of a disclosure within a model of afinancial report provided by Pacioli20:Fact DatabaseThe fact database is essentially equivalent to the facts that are reported within an XBRLinstance. The separation of the facts reported from the knowledge base of rules that supportthose reported facts is somewhat arbitrary.There are many approaches to storing facts within a database21. Each approach has a set ofPROS and CONS; no approach is 100% the best or 100% the worst. What appear to be the mostviable information storage alternatives include: SQL database: These are the most pervasive and the most popular today.RDF triple store: These are popular for working with the W3C Semantic Web Stack.These are sometimes implemented within a SQL database.Graph database: Graph databases such as Neo4j22 are increasing in popularity, standardquery languages are being developed like Cypher23.NOSQL databases: NOSQL databases such as MondoDB are increasing in popularitybecause they require no schema which can be a feature or a bug depending uponwhether you desire a database schema.DATOMIC: Datomic24 is a fact database or cell store25 that has a built in DATALOG rulesengine.20Mini Financial Reporting Scheme, Report Analysis, tml21Understanding Database/Query Options (Part o4j, .html23Cypher, https://www.opencypher.org/24Datomic Cloud, https://www.datomic.com/25Ghislain Fourny, PhD, Cell Stores, https://arxiv.org/pdf/1410.0600.pdf18

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication https://creativecommons.org/publicdomain/zero/1.0/ Cell store: Reportix26 is an example of a cell store that is specific to XBRL-basedinformation.What is the right database alternative to use? That is a decision that should be made byqualified technical professionals.Knowledge BaseThe knowledge base is essentially equivalent to the information that supports reported factsthat is represented within XBRL taxonomy schemas, XBRL linkbases, and other informationprovided in the form of XBRL Formulas. The knowledge base is essentially machine-readablestatements based on factual and heuristic knowledge created based on experience andpractices of the best domain experts.The following are example/prototype knowledge bases for several financial reporting schemes: US GAAP27IFRS28IPSAS29FRF for SMEs30US GAAP Not-for-Profit31Other testing, prototype, and other such XBRL-based financial reporting schemes wererepresented in order to collect information which could yield information useful to create onframework for representing all financial reporting schemes. That information is summarized inMastering XBRL-based Digital Financial Reporting32.What I call the PROOF BASELINE33 representation takes everything that is common between allother prototype financial reporting schemes and distills it down into the simplest yet completerepresentation possible. This Proof Baseline representation is used to explain and test.26Reportix, https://www.reportix.com/products cellstore.phpUS GAAP financial reporting scheme, cheme/usgaap/documentation/Index.html28IFRS financial reporting scheme, heme/ifrs/documentation/Index.html29IPSAS financial reporting scheme (prototype), heme/ipsas/documentation/Index.html30FRF for SMEs financial reporting scheme, html31US GAAP Not-for-Profit financial reporting scheme, heme/nfp/documentation/Index.html32Mastering XBRL-based Digital Financial Reporting, oof Baseline, cheme/proof/documentation/Index.html2719

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication In essence, it is possible to represent any financial reporting scheme34 using the notion ofprofiles35 to adjust for any minor differences between how each financial reporting schemechooses to implement XBRL-based digital financial reporting.But how do you get the knowledge that ends up in a knowledge base? You need some sort ofmechanism for acquiring knowledge.Knowledge Acquisition MechanismThe power of any knowledge based system is proportional to the key ingredient of theknowledge based system which is high-quality machine-readable domain knowledge availableto that system. Knowledge acquisition is the process of obtaining that domain knowledge.There are three approaches to acquiring knowledge:1. A rules-based approach which involves humans creating machine-readable knowledge.2. A patterns-based approach which involves machine learning to capture domainknowledge which is useful when there is a high tolerance for error. Further, extensivemachine-readable training data is necessary to use this machine-learning basedapproach.3. A combination of approaches #1 and #2 to create a hybrid approach to acquiringknowledge.For the domain of financial reporting, there is ZERO probability that approach #2 (i.e. machinelearning) can be used to acquire the initial financial reporting domain knowledge.However, after some unknown period of time when enough machine-readable information hasbeen created by human domain experts; then that human created machine-readableinformation can be leveraged to create additional new information.For example, information about disclosures36 can be used to learn how to create algorithms foridentifying other such disclosures simply by probing existing XBRL-based financial reportssubmitted to financial regulators such as the SEC and ESMA. That machine-readableinformation along with humans to guide and tweak the process can be used to identify rules forother unknown disclosures by looking for specific known patterns.34Comparison of Financial Reporting Schemes High Level brary/ReportingSchemes-2018-12-30.pdf35XBRL-based Digital Financial Reporting Profiles and General Business Reporting rary/Profiles-2018-10-22.pdf36Disclosure Best Practices, Name IncomeStatement20

CC0 1.0 Universal (CC0 1.0) Public Domain Dedication We don’t want every enterprise or regulator creating proprietary approaches to creatingknowledge based systems for storing and working with financial reports. A better approach foreveryone is to have high-quality global standard models which makes creating software moreefficient and therefore less costly.It takes skill and experience of a domain to create knowledge for a domain. Businessprofessionals have that skill and experience and will need software which they can realisticallyuse to put, collate, categorize, associate, and otherwise create useful machine-readableknowledge.Understanding the SBRM Meta-Meta ModelA meta-model is a model whose purpose is to describe and process models that subscribe tothat meta-model. Models and meta-models both prescribe and describe what is permissibleand what is not permissible per some model or meta-model.Utility of MethodOne reason for this is t

Mar 24, 2021 · A logical theory is a set of models (ontology like things) that are consistent with the logical theory. A logical theory provides a way of thinking about a domain by means of deductive reasoning to derive logical consequences of the theory. I have created a logical theory that describ

Related Documents:

EPA Test Method 1: EPA Test Method 2 EPA Test Method 3A. EPA Test Method 4 . Method 3A Oxygen & Carbon Dioxide . EPA Test Method 3A. Method 6C SO. 2. EPA Test Method 6C . Method 7E NOx . EPA Test Method 7E. Method 10 CO . EPA Test Method 10 . Method 25A Hydrocarbons (THC) EPA Test Method 25A. Method 30B Mercury (sorbent trap) EPA Test Method .

The Problem with Root Cause Analysis Method A Method B Method C Method G Method E Method H Method J Method F Method D Method I No‐one can agree on "what is a root cause." Everyone says they do "root cause analysis,"yet everyone is doing something different!

Object built-in type, 9 Object constructor, 32 Object.create() method, 70 Object.defineProperties() method, 43–44 Object.defineProperty() method, 39–41, 52 Object.freeze() method, 47, 61 Object.getOwnPropertyDescriptor() method, 44 Object.getPrototypeOf() method, 55 Object.isExtensible() method, 45, 46 Object.isFrozen() method, 47 Object.isSealed() method, 46

Humidity Test Method MIL-STD-1344, Method 1002 Immersion Test Method IEC 60529:2001, IP68 UV Resistance Test Method MIL-STD-810, Method 505 Vibration Test Method MIL-STD-202, Method 214 Weatherproofing Method Butyl and electric

method EPA 218.6 (ug/L) Chromium, total dissolved by method SW 6020 (ug/L) Iron by method SW 6010B (ug/L) Iron, dissolved by method SW 6010B (ug/L) Manganese, dissolved by method SW 6020 (ug/L) Molybdenum, dissolved by method SW 6020 (ug/L) Nitrate/Nitrite as Nitrogen by method EPA 353.2 (mg/L) pH by method SM 4500-H B (PHUNITS) Selenium .

is 7.678622465 times better than the Bisection method. 2 BISECTION METHOD. The Bisection Method [1] is the most primitive method for nding real roots of function f(x) 0 where f is a continuous function. This method is also known as Binary-Search Method and Bolzano Method. Two initial guess is required to start the procedure. This method is based

(Packer, 2007). It is this coincidence that has finally connected psychology and cinema to this day. The cinema education method or method of therapy using film media is a method developed from the bibliotherapy method (Mamahit, 2013). The cinema education method is an innovative method and

7. Dave Elman Method 18 8. Direct Gaze Method 20 9. Drop Object Method 22 10. Eye Fixation with Imagery Method 24 11. Fixation Object Method 26 12. Favorite Place Method 27 13. Forest And Stream Method 29 14. Gil