8 ppl attended
Announcements:
- CAA Maintainathon confirmed for 18–22 January 2027.
- Zack will circulate an information email.
Talk by Dries: NASSA
- Dries introduced himself and gave an overview of agent-based modelling (ABM) in archaeology.
- Introduced NASSA, an initiative supporting cumulative research in ABM through a library of reusable modelling components.
- Key features:
- Shared conventions, metadata standards, and review procedures.
- YAML-based metadata schema with automated validation.
- Semantic tagging for module discovery.
- Web interface and research directory for browsing modules.
- Discussed differences from CoMSES Net and the NetLogo User Community Models repository, particularly the focus on reusable modules rather than complete models.
- Invited participants to join and contribute.
- Resources:
- NASSA Hub
- Metadata schema
- Module directory
- Mailing list: nassaabm@googlegroups.com
- Presentation slides will be attached to the minutes.
Q&A and Discussion
Module reuse and contribution
- Question about how modules are identified, extracted, and transferred into NASSA.
- Dries: Currently a manual process.
- Jim: How can contributors identify useful components and isolate them from existing models?
- Initial contributors examined their own models for reusable components.
- LLM-assisted workflows for identifying modules and extracting metadata are being explored.
Metadata and interoperability
- Zack: How was the YAML schema developed, and how does it relate to existing standards and wider communities?
- Based on existing examples but adapted to the modules available.
- Broader interoperability and dissemination remain topics for future discussion.
- Clemens: Suggested mapping the metadata schema to an existing ontology; no Linked Open Data specialist was involved in its initial development.
Integration with Python and Mesa
- Matteo: Asked about integration with Mesa and the mapping of Python modules to NASSA’s structure.
- Dries: Not currently supported, but worth exploring. An automated NetLogo-to-Python translator was considered initially.
- Sophie: Suggested exploring this during the CAA Maintainathon.
Model critique and development
- Zack: NASSA could enable a “marketplace of ideas.” Does it facilitate critical reassessment and improvement of existing models?
- Dries: Individual modules are often too basic for extensive critique, but versioning supports their development.
- Examining how modules are combined and evolve into larger models could be particularly interesting in the future.
Institutional sustainability
Matteo: Has institutional support or long-term hosting been considered? - Dries: No institutional arrangement so far; some reservations about institutional dependencies and changing personnel. - Contributors’ time is currently a greater constraint than funding.
Archaeological specificity
- Caro: What makes the modules specifically archaeological?
- Dries: Not all modules are inherently archaeological; some address general processes, while others relate directly to archaeological questions.
LLM strategy
- Martin: Asked about NASSA’s strategy regarding LLM-based access and reuse.
- Dries: No explicit strategy yet; current approach assumes fair use and appropriate attribution.
- Clemens: GitHub hosting limits control over crawling, although the static infrastructure avoids performance issues.
- A policy on LLM-based use of metadata and code may be needed.
Licensing
- Zack: Asked about licensing across modules.
- Dries: Openly licensed modules are preferred; no significant licensing issues encountered so far.
- Clemens: Licences can be specified per module, and almost all modules have one.
Impact and feedback
- Zack: Are there mechanisms for tracking reuse or receiving feedback from users?
- Dries: Currently little direct communication with users; no established feedback mechanism.
Persistent identifiers
- Matteo: Asked about DOI assignment. -Dries: Not currently implemented.
