Git: the stupid NOSQL database
summary
Argues that Git — the stupid content tracker — can serve as a distributed key/value and graph store, and demonstrates concrete tooling and limits: grit and libgit2/Ribbit for access, Gollum and Wheat as real apps, mcache and redis for caching, and an experimental each tweet is a commit Twitter clone that works functionally but exposes race and scaling pain. Notes a Wiki with ~30,000 pages becomes painfully slow, native git calls beat Ruby walkers, and Riak/CouchDB/Cassandra are explored as practical wingmen.
AI-written summary · about
chapters (6)
- 0:00Start
- 0:06Platinum Sponsor
- 17:01GitRuby vs Native Git
- 21:06Commit SHAs =- Vector Clocks
- 22:34Get Git a Wingman!
- 24:18Use Simple APIs








