An outsider's perspective on Scion and Mononoke
Web page · https://typesanitizer.com/blog/scion.html
First seen by Oriel on .
AI summary
A Databricks source-control engineer assesses Anthropic’s Scion against Mononoke and Git, finding familiar architecture, an unusually rapid migration, and several questionable scaling and performance claims.
- The author considers changesets, client-server asymmetry, APIs, horizontal scalability, sparse working copies, per-directory ACLs, and lock-free commits established or borrowed ideas.
- Anthropic reportedly migrated its monorepo from Git in five months with seven people while building Antique, code search, IAM, and code-landing workflows.
- Claude “superfans” monitor individual components through telemetry and logs, potentially detecting subtle issues beyond traditional alerts, though no concrete examples are provided.
- The author questions whether discouraging refs improves scalability and argues temporary refs could be cleaned up while long-lived refs remain manageable.
- Comparisons involving hundreds of Git mirrors, storage limits, and Scion performance percentiles appear insufficiently explained and inconsistent with the author’s Databricks experience.
#version control #scion #mononoke #monorepos
Generated by gpt-5.6-sol. Check the original for details.
Metric trends
Last 30 days · Up to 120 recorded samples per source and metric, with reposts kept separate · First and latest retained
Points
Score reported by the source.
- Lobsters: 151, latest observation ; +150 over observed period since
Comments
Comment count reported by the source.
- Lobsters: 20, latest observation ; +20 over observed period since
Aliases
| Source | External ID | URL |
|---|---|---|
| Lobsters | s2qeww | https://typesanitizer.com/blog/scion.html Read discussion |