====== Model Critique ======

===== Definition =====

==== Is: What it is ====
==== IsNot: What it is not ====


===== Rationale =====
  * Models are not correct by virtue of inspection or by virtue that everybody worked on it.
  * Many different ways to implement same function and inevitably there is a trade-off when a choice is made. A critique of a model provides an outsider's perspective
  * Verify and validate the model for a given project, potentially as part of a non-advocate review
  * Evaluate the current state and achievements on an MBSE effort to determine whether it requires or would benefit from additional funding
  * Evaluate whether a sister project should adopt the MBSE approach based on the current MBSE efforts 
  * Evaluate a tool vendor


===== Performance Qualities  =====
: How to determine the performance of {Thing}
===== Problem  =====
: What issues are relevant to {Thing}
===== Precondition =====
: What conditions are necessary for {Thing} to operate
===== Postcondition  =====
: What conditions are necessary for {Thing} to produce
===== Ontology  =====
: What terms/taxonomy are relevant to {Thing}


===== Methodology =====

==== Method ====

=== Heuristics ===
  * Is there a potential for failure due to timing?
  * Has the model used patterns?
  * Does different components react differently to different stimuli
  * Is there consistency
  * Look at interfaces between SOI and context
  * What are the system dynamics and failure
  * What are the rainy day scenarios
  * Do observations on numbers

=== Proofs ===

==== Process ====
{{ :bok:eng:mbse:critique:week_03_still_04_b.jpg.png?400 |}}

  -  Functionality and Use Cases
    * What are the preconditions for the function
    * Is there any validation vs requirements
    * What happens in case of error/illegal use csae
  -  Structure of the Model
  -  Interfaces
  -  Dependability and Failure Modes
    * Check for any broken/circular relationship/redundant content
    * Touches on security/availability
    * What failure modes are anticipates
    * Has the modeler tested for resilience to failure mode
  -  Concurrency Analysis
    * There are two types concurrency failure (livelock and deadlock)
    * Livelock - When an element constantly changing resulting in a "runaway" state preventing model convergence
    * Deadlock - Two components attempt to share a common resource, blocking access to both
    * It is possible to be under-specified and over-specified
  -  Management and Deployment
    * Has allocated logical to function to structure
    * Who has authority to make changes?
    * Has model lifecycle management process been documented?
  -  Review Against Qualities of Great Models

===== Parts =====
==== Outputs ====

  * Remarks on whether the model is meeting the intended scope and purpose.
  * Set of observations and data on the model’s strengths and weaknesses. Your findings should relate back to specific instances of the model with data sampling to backup your feedback.
  * An evaluation against each of the qualities of great models, as well as supplementary behaviors or qualities you believe are relevant for an MBSE model.
  * Conclusions and recommendations for the specific rationale for the critique.


===== Constraints =====
===== Thing is refined by  =====
===== Use Cases =====
===== This is specialised by  =====
===== Generalization Sets: What sets subdivide {Thing} =====