Request for feedback on an experimental ASPECT learning and modeling workflow for geoscience students

ASPECT
1

Hello ASPECT community,

I have been working on an early experimental project called Aspect_Yuan.skill:

The project is intended primarily for geology and geophysics students who are interested in computational geodynamics but have limited experience with ASPECT, numerical methods, or scientific software development.

The goal is not to replace the official ASPECT documentation, hide numerical choices, or turn geological modeling into a one-click black box. Instead, I am exploring whether a structured Codex skill—consisting of instructions, scripts, templates, checks, and example workflows—could help beginners move more carefully through the following process:

  • translating a geological question into explicit model choices;

  • learning from existing ASPECT cookbooks and benchmarks;

  • generating, explaining, and checking parameter files;

  • identifying ASPECT versions, plugins, and external files used in published studies;

  • checking run logs and model outputs;

  • developing reproducible postprocessing and plotting workflows.

The current version is an early beta. It includes beginner workflows for mantle convection, subduction, and rifting, as well as preliminary tools for parameter-file validation, paper-reproduction planning, output inspection, and plotting. I am aware that many parts are incomplete, may overlap with existing ASPECT resources, and may require substantial revision.

Before developing it further, I would be very grateful for critical feedback from ASPECT developers and experienced users, particularly on the following questions:

  1. Which parts of this workflow could be scientifically or technically misleading for beginners?

  2. What safeguards should be required before a generated model is considered suitable for scientific interpretation?

  3. Which functions should simply direct users to the official ASPECT documentation rather than attempting to reproduce existing guidance?

  4. What minimum tests, benchmarks, and version checks would you expect from such a project?

  5. Which difficulties faced by new ASPECT users would actually be useful to address?

I would also appreciate examples of realistic beginner or paper-reproduction tasks that could be used to evaluate the project. I do not expect this tool to automate scientific judgment or guarantee reproduction of published results.

This is an independent experimental project . I would be happy to revise its scope, terminology, and technical direction based on community feedback.

Thank you for taking the time to look at it and for any suggestions or criticism.