Get the latest tech news

Regarding Prollyferation: Followup to "People Keep Inventing Prolly Trees"

Last month I said "people keep inventing prolly trees". Is that really true? What does that mean, really?

The above diagram is from the Inria paper, and it does a great job of conveying their data model while leaving very little to the imagination: keys stored in intermediate nodes create a usable tree map with a data-driven shape. This does make me suspect that XetHub could reduce the number of lookups required by storing key metadata in their Merkle trees in addition to having the data available in the underlying file. XetHub manifests data as mounted files that can be efficiently read using standard tooling, and requires additional content-aware analysis to perform version diffing.

Get the Android app