What is the technical rationale behind warning diagnostics for vector/matrix block input conversion?
1 view (last 30 days)
The Simulink documentation recommends a setting of "error" for the vector/matrix block input conversion diagnostic in "safety precaution" applications. Several modeling standards for high-integrity systems similarly recommend a warning or error output for that diagnostic. Is there a published white paper or other rationale that describes the types of errors or non-deterministic behavior that can result from automatic vector/matrix conversions?
My application includes numerous matrix multiplication applications that are performed on arrays or 1-D vectors that lead to numerous nuisance warnings of this type, and I would like to understand the implications of disabling the diagnostic or adding signal definition or reshape blocks to resolve it, as no adverse behavior has been observed.
TAB on 17 Oct 2018
Its part of High Integrity System Modeling guideline.
Please read the description at Check safety-related diagnostic settings for type conversions.
The diagnostic that detects vector-to-matrix or matrix-to-vector conversions at block inputs is set to none or warning. When the Simulink software automatically converts between vector and matrix dimensions, unintended operations or unpredictable behavior can occur.
In my opinion, purpose of the guideline to ensure that Vector/matrix conversion happens as developer wants (means explicitly and not automatically/implicitly) to avoid any unexpected problem.
Just like explicit Data type convrsion (which is also applicable for C)