Interjected Future

As part of biasing towards action, I thought it was time to reimplement the Prolly Tree in Zig. I wanted to focus on being cache-friendly this time, now that I have lower-level control. But that resulted in a deep dive into data-oriented programming. The question is how to represent a

My adherence to immutability has put me on the hard path for everything so far. Not just for the unicit Prolly tree, but also for the CRDT. Most practical CRDT libraries use an op-based CRDT or a delta-based CRDT, but many do not keep the entire history around. I think

I started working on making an LLM Zine. Not to worry, the work on the database is still in the works, but this has been a burgeoning interest and opportunity. LLMs are a fundamental technological breakthrough, and the effects of this technology will take at least a decade to work

I've now implemented an Op-based and a Pure op-based CRDT. I couldn't quite make the leap before to merkle-based CRDTs. While vectors clocks are mechanically simple, I wasn't quite sure how they mapped to a merkle-DAG node if replica IDs weren't available. After seeing them in action, I'm fairly certain

This continues my exploration into CRDTs, beyond the typical tutorials. The more I looked, the more threads there were to pull on. I currently have a couple of lines of questioning to answer: * How do we compose CRDTs so they both converge to a single state without coordination but with

After the implementation of the Prolly Tree, I set about writing an immutable map. It was pretty straightforward as most of the heavy lifting is done by the Prolly Tree implementation. The immutable map just keeps a simple list of roots after every change. There are a bunch of improvements

It took a long time, but the incremental insert for a Prolly Tree is finally done. I don't think I've ever spent this long as such a little part of code. The two aspects that had me struggling were the bottom-up breadth-first traversal, and how to replace the swath of

The punchline first: the incremental insert is finally done on the Prolly Tree. I don't think I've ever spent so much time on such a small, core piece of code. It was also a lesson for me to lean into the struggle and talk about it. Last time, I talked

A problem was that traversal in a layer would be backward. I'd need to traverse left-to-right to match the build process when building the Prolly Tree from scratch. No problem, I'll just loop through the children backward when processing top-down.

I've been stuck on the incremental insertion of a key in a Prolly Tree, and feeling like I'm in a bog. The version where a Prolly Tree is built from scratch was relatively straightforward. I was able to write it in a day. The incremental version has undergone four or

Last week, I built Prolly Trees from scratch and probed some of its properties. This week, I started building a Prolly Tree incrementally, by inserting one key at a time. It was more challenging than I suspected at first glance, seeing as I implemented it three times. In a typical

Alright, party people. This week I went back to reimplementing Prolly Trees. I wanted to finish out the Lattice Merkle Tree. But given that my goals aren't to just fiddle around in applied research and Prolly Trees already work, I reimplemented Prolly Trees from scratch. I was delighted that the

In Lab note #029 [https://interjectedfuture.com/lab-notes/lab-notes-029-balancing-nodes-in-a-unicit-tree/] , I was thinking about a different way to construct a unicit tree. Can we create both a description of the splits in a tree and a comparator that would make it a partially ordered set? And with that partially ordered set

In my previous lab note #028 [https://interjectedfuture.com/lab-notes/lab-note-028-consistent-splitting-for-node-sizes/] , I was worried that Prolly trees [https://github.com/attic-labs/noms/blob/master/doc/intro.md#prolly-trees-probabilistic-b-trees] were not unicit. When I inserted keys in different orders, there was a small chance that splits would propagate to the end

Last time [https://interjectedfuture.com/lab-notes/lab-note-028-consistent-splitting-for-node-sizes/] , I found balancing nodes seemed to be in conflict with the property of a unicit tree. A balanced tree is desirable for efficient queries. A unicit tree is desirable for Merkle Trees for fast equality comparisons. Was there a way to get both?

👉See Lab note #30 [https://interjectedfuture.com/lab-notes-030-prolly-trees-are-unicit] for corrections. Prolly Trees are unicit, and don't exhibit the bad behavior I've outlines below.Last week [https://interjectedfuture.com/lab-notes/lab-note-027-breadcrumb-to-prolly-trees/] , I was looking at Prolly Trees [https://github.com/attic-labs/noms/blob/master/doc/intro.md#prolly-trees-probabilistic-b-trees] . A major concern

When I found the splitting of nodes for Merkle Search Trees to be problematic [https://interjectedfuture.com/lab-notes/lab-notes-025-split-condition-for-merkle-search-trees/] , I stumbled upon Prolly Trees [https://github.com/attic-labs/noms/blob/master/doc/intro.md#prolly-trees-probabilistic-b-trees] . They were first described by the Noms, a git for databases [https://github.com/attic-labs/

Most desktop, mobile, and web apps use a form-based metaphor to manipulate the state of the digital objects sitting in the database. This is a little odd when you think about it. Most things in the real-world don't work like this, aside from our complex machines. What would it look

I started outlining how to compose different CRDTs together to make an Immutable Conflict-free Resolution Relations [1] construction, based on the Merkle Search Tree. When designing it, I first tried sketching out the entire construction by drawing, but I was left confused. I wasn't sure if all the pieces actually

A major implication not often included in discussions of local-first software is how to deal with migrations. The degree of the problem depends largely on how you fence the boundaries of your system. If your local-first system allows any number of replicas to come in and out of the network

This week the todo list is coming along. The app works like a regular todo list. There's no server, and yet it persists the data between reloads. I wanted to see how it would feel to have a database as the state management library. Instead of having an app state