Hi all,

We are currently discussing software componentization on several scope levels:

• General discussion about componentization without direct relation to specific existing description format or solution.
• Minimal Viable Product (MVP) related discussion, where MVP is defined as minimalistic CLI tool which should be able to assemble buildable project from SW components.
• CMSIS pack specific discussion with focus on gaps identification.

I would like to share few resources and observations related to dependency resolution as part of general discussion.

There is segment of tools called EDA. Such tools use various design rules checkers based on logic rules evaluation.

Another segment is design of logic circuits especially area of circuit optimization. Optimization could be done by application of Boolean algebra or by more user friendly methods listed in switching circuit theory.

Specific methods of optimization are:

• Karnaugh map
• Quine-McCluskey algorithm
• Espresso minimizer

Result of those methods is minimized standard expression.

Dependency resolver should be able to transform expression into minimized standard form.

Another typical issue is Boolean satisfiability problem (SAT).

Dependency resolver should be able to validate whether expression could be resolved.

Example libraries and references:

• PyEDA (example of Boolean manipulation)
• https://doc.opensuse.org/projects/satsolver/11.4/
• http://www.lifl.fr/LION9/presentation_orator/lion9.pdf
• Solving Linux Upgradeability Problems Using Boolean Optimization
• https://www.cril.univ-artois.fr/documents/nanjing-leberre.pdf
• https://www.sat4j.org/

This is result of brief research, I have no personal experience with SAT libraries.
Do you have any specific experience with SAT ?

Regards,

David

David Jurajda

Software Engineer
MCU Software
NXP Semiconductors
1.maje 1009

756 61  Roznov pod Radhostem

Czech Republic

Phone: +420 571 665 339
david.jurajda@nxp.com