Update on Raku Persistent Data Structures
summary
Implements a Hash Array Mapped Trie style persistent vector and maps for Raku aiming to give immutable data without full copies by using path-copying and node sharing. Work is underway with about 20 hours implemented, plans to target a branching factor 32 for cache friendliness, and uses index or hash bits to traverse shallow trees so appends copy only O(log32 N) nodes. Prior art: Clojure’s implementation; naming options include “sip” or “hat” and the API design is still unresolved.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 0:47Presentation Overview
- 2:24Why
- 4:27The Big Idea
- 6:10Under the Hood
- 9:25Similar Data Structures
- 10:27Name
- 11:25Name Elimination
- 13:27Two Hard Problems
- 15:00Naming
related talks
40m"Gaining Constant time Lookup over Unorganized Data"
39mEventually Consistent Data Structures
21mTangram: Bridging Immutable and Mutable Abstractions for Distributed Data Analytics
17mScaling Enterprise-Grade RAG: Lessons from Legal Frontier
16mEmpowering Fearless Data Engineering
47mDesign Decisions Behind Cube Store - a Blazingly Fast Open-Source Rollup Storage Layer
29mRhyme: A Data-Centric Expressive Query Language for Nested Data Structures
41mMaximising Data Storage Efficiency with Ceph as a Caching Layer for Tape Archiving