Levels of a Report

This is my third attempt to explain the levels of a digital financial report. Here is the first attempt.  Here is the second attempt.  This is similar to describing the levels of self driving in order to understand the meaning when someone says "self driving".  Or, see this Wikipedia article, Evaluation Assurance Levels, related to security.  Note the notion of "common criteria certification".

The purpose of this is to understand what is meant by the term "logical twin" (a.k.a. logical digital twin).  Below is my third attempt at defining the levels of a general purpose financial report.

1. Level 0 (Information provided physically): Not machine readable. An example of Level 0 is a clay tablet, papyrus, or paper as the report medium. (Example, imagine that this example is on paper.)
2. Level 1 (Information provided digitally): Machine readable, structured for presentation of the information, information metadata is nonstandard. (Examples: PDF, HTML, JPEG, and other forms of e-paper.
3. Level 2 (Information provided digitally, structured for meaning, nonstandard): Machine readable, structured for meaning, nonstandard, no taxonomy, no rules, no report model. (Example, an XML document; no model explaining report information.
4. Level 3 (Global standard report model and report syntax structured for meaning): Machine readable, global standard syntax, structured for meaning, with report model which explains some of the report (e.g. incomplete), and the terms used are not from a published common financial reporting scheme. (Example, XBRL-based report model and report but the report defines all report information locally in the report.)
5. Level 4 (Common dictionary of terms, structures, associations, rules): Machine readable, global standard syntax, structured for meaning, with report model, complete set of rules provided, incomplete high-level report model. (Example, an XBRL-based report with a XBRL taxonomy schema, with XBRL relations and resources, and with XBRL Formulas that completely describes the report.)
6. LOGICAL TWIN - Level 5 (Complete set of logical statements): Machine readable, global standard syntax, structured for meaning, with taxonomy, complete set of rules provided, complete global standard high-level report model; yields verifiably proven properly functioning system and understandable report information. (Example, an XBRL-based report with all the characteristics of Level 4, plus consistency cross checks, type-subtype relations, consistent and logical modeling of XBRL presentation relations, information that describes the correct representation of every disclosure within the report, and a reporting checklist that describes all required disclosures; verified to be a properly functioning logical system.)
7. Level 6 (Trust report logic not manipulated): All of Level 5 PLUS blockchain-anchored XBRL to increase trust. An XBRL-based report with all the characteristics of Level 5, plus information in the form of a Merkle hash within a digital distributed ledger that assures no one has tampered with the report. (No example to show; here is somewhat of a prototype)
8. Level 7 (Trust report transactions provenance): All of Level 6 PLUS blockchain-anchored accounting transactions and events. An XBRL-based report with all the characteristics of Level 6, plus information in the form of a Merkle hash that indicates that assures no one has tampered with transactions. (No example to show)

A logical twin would be line 6 above in fuchsia "Level 5 (Complete set of logical statements)". Alternatively, a logical twin schema boundaries might be agreed to by the stakeholders of the system making use of the logical twins.

Level 0 is not digital, it is physical, so it does not qualify as a logical twin.  Levels 1 to 4 are "digital twins" but do not meet the definition of a logical twin because the report is incomplete in terms of the logic boundaries specified for a report (in this case by the Seattle Method).

Level 6 and 7 add additional trust features, enhancing the capabilities of a logical twin.

* * *

There are many different technical alternatives implementing such systems (problem solving systems) but the implementation alternatives tends to fall into three primary buckets: semantic web stack, graph database, logic programming. The logic of the system would be expected to be the same regardless of the implementation alternative used.

Such systems need to be complete in terms of expressed logic in order to be able to prove that the system is a properly functioning logical system.  There tends to be a spectrum of approaches to representing knowledge for an area of knowledge (taxonomy, ontology, theory, etc.) with a "theory" being the most expressive.   

There tends to be a spectrum of level of logic for expressing knowledge with DATALOG being the best balance between expressive power and processing safety (avoiding catastrophic failure).  

There tends to be three types of reasoning approaches or tools (deductive, inductive, and abductive) and the right tool needs to be used for the task.  

An area of knowledge can be "kind" or "wicked".  The art is to balance these dynamics with the goal(s) and objective(s) of the system.

Additional Information:

• Systems Thinking
• DOW 30 Dashboard