Hacker News

Article URL: https://stencil.so/blog/prewalk

Comments URL: https://news.ycombinator.com/item?id=48922646

Points: 1

# Comments: 0

Hi HN! I've posted SirixDB here before, back in 2019 and again in 2023.

The core idea behind SirixDB is, that history is a first-class citizen. Every commit stores a lightweight, queryable revision. You can query any point in time, even individual nodes (for instance JSON values), diff arbitrary revisions, and efficiently track how data evolved without replaying events.

Unlike traditional event stores, historical states do not need to be reconstructed by replaying events nor do we have to think about projections. Revisions are directly queryable.

A simple example:

Jan 1: Record "Price = $100, valid from Jan 1". Stored on Jan 1 (transaction time). Jan 20: Discover price was actually $95 on Jan 1. Commit correction.

After correction, you can ask across both axes: - "What did we THINK the price was on Jan 16?" -> $100 (Transaction time) - "What WAS the price on Jan 1?" -> $95 (Valid time)

I've worked on this in my spare time since 2013, following its academic precursor (Idefix/Treetank) at the University of Konstanz. The architecture relies on an append-only physical log and a copy-on-write page trie.

A high level view of the architecture:

Physical Log (append-only, sequential writes) ┌────────────────────────────────────────────────────────────────────────┐ │ [R1:Root] [R1:P1] [R1:P2] [R2:Root] [R2:P1'] [R3:Root] [R3:P2'] ... │ └────────────────────────────────────────────────────────────────────────┘ t=0 t=1 t=2 t=3 t=4 t=5 t=6 → time

Each revision is indexed, whereas unchanged pages are shared among revisions (even unchanged nodes in pages with changes mostly):

Revision Roots Page Trie (persistent, copy-on-write) │ ▼ [Rev 3] ─────────────────┬────────────────┐ │ │ │ [Rev 2] ──────────┬──────┤ │ │ │ │ │ [Rev 1] ───┐ │ │ │ │ │ │ │ ▼ ▼ ▼ ▼ [Root₁][Root₂][Root₃] [Pages...] │ │ │ ▼ ▼ ▼

Unchanged pages are shared between revisions. Beneath the root pages sit node indexes and secondary indexes, which also use a novel sliding-snapshot algorithm to balance read and write performance. Everything is queryable using JSONiq via the Brackit compiler.

Back in 2019, and even in 2023, SirixDB was very slow due to GC pressure. Unlike most other document stores, SirixDB stores fine-grained nodes, and I came to realize that an on-heap (JVM) representation made up of lots of small objects simply didn't make sense. I measured it with async-profiler — with some help from Andrei Pangin himself — and the result was that the poor throughput was due to the sheer amount of allocations which scaled almost linearly with the number of open transactions.

Working a full-time software engineering job, I lacked the energy for a massive spare-time rewrite. About a year ago, I started experimenting with AI. It turned out to be ideal for automating the tedious, repetitive parts of migrating the storage layer to Java's Foreign Function & Memory API, storing pages completely off-heap.

Looking further ahead, the append-only, immutable-page design maps naturally onto object storage like S3 and distributed logs like Kafka for a cloud version, and initial prototypes already exist. Maybe that becomes a commercial service one day, but for now, I'm just thrilled to see these core design principles finally proven out.There's an interactive demo [3], documentation, and the code is on GitHub [1]. I'd love feedback and am happy to answer questions!

kind regards

Johannes

[1] https://sirix.io | https://github.com/sirixdb/sirix

[2] https://sirix.io/docs/architecture.html

[3] https://demo.sirix.io

[4] https://sirix.io/docs/

[5] http://brackit.io

Comments URL: https://news.ycombinator.com/item?id=48922631

Points: 1

# Comments: 0

Article URL: https://sa.my/slipstream/

Comments URL: https://news.ycombinator.com/item?id=48922625

Points: 1

# Comments: 1

today ran my own tool over the whole django repo it indexed it in 2.5 minutes and generated a whole graph for each function i searched, its always awesome to look at something you build on your own works. T-T

Comments URL: https://news.ycombinator.com/item?id=48922615

Points: 1

# Comments: 0

Article URL: https://arxiv.org/abs/2607.01316

Comments URL: https://news.ycombinator.com/item?id=48922609

Points: 1

# Comments: 0

Article URL: https://www.nbcnews.com/tech/apple/apple-openai-lawsuit-suit-trade-product-hardware-email-sam-altman-rcna587376

Comments URL: https://news.ycombinator.com/item?id=48922604

Points: 1

# Comments: 0

Article URL: https://www.macrumors.com/2026/07/14/lego-imac-g3-set/

Comments URL: https://news.ycombinator.com/item?id=48922587

Points: 1

# Comments: 1

Article URL: https://arstechnica.com/security/2026/07/now-defenders-are-embracing-the-prompt-injection-too/

Comments URL: https://news.ycombinator.com/item?id=48922551

Points: 1

# Comments: 0

Article URL: https://www.bloomberg.com/news/articles/2026-07-15/over-400-clos-tabbed-for-upgrade-in-pivot-that-fans-08-fears

Comments URL: https://news.ycombinator.com/item?id=48922546

Points: 1

# Comments: 0

Article URL: https://arxiv.org/abs/2607.05391

Comments URL: https://news.ycombinator.com/item?id=48922538

Points: 2

# Comments: 0

A while back, I decided I wanted to build a grocery store list app for my wife and I to share. Building a native app felt too difficult. I would have to deal with App Store review, TestFlight, and I’d still need to run a server to store data. Seemed like too much hassle for a grocery list app. Plus, I wanted to build a bunch more apps later, like a credit card perk tracker and a tracker for the family’s Pictionary games.

All of this turned into a full OSS platform. All apps in Homestead start out with the resource schema. Homestead creates fully authenticated REST APIs on top. You can use the built-in, agent-friendly CLI if you’re tech savvy or you can use the built-in web server to expose React apps. All the apps are easily compartmentalized, live on your machine, and each can have their own Home Screen icon.

If you’re interested in building personal software, please take a look!

Comments URL: https://news.ycombinator.com/item?id=48922536

Points: 1

# Comments: 0

I think we need to have something like npm, but for AI skills - with trackers & such; with proof and validation before installing / using a skill. Generalized's idea is this. :) I'd love for you to publish your skills.

Comments URL: https://news.ycombinator.com/item?id=48922498

Points: 1

# Comments: 0

Article URL: https://www.newyorker.com/magazine/2026/07/20/when-ai-is-a-member-of-the-family

Comments URL: https://news.ycombinator.com/item?id=48922487

Points: 1

# Comments: 0

Article URL: https://arxiv.org/abs/2607.00764

Comments URL: https://news.ycombinator.com/item?id=48922469

Points: 2

# Comments: 0

Article URL: https://www.wsj.com/pro/bankruptcy/brightspeeds-fiber-asset-bet-meets-11-billion-debt-test-33fe5260

Comments URL: https://news.ycombinator.com/item?id=48922463

Points: 2

# Comments: 0

Article URL: https://www.neomindlabs.com/2026/06/08/running-gemma-4-26b-at-5-tokens-sec-on-a-13-year-old-xeon-with-no-gpu/

Comments URL: https://news.ycombinator.com/item?id=48922434

Points: 3

# Comments: 1

Article URL: https://www.bloomberg.com/news/articles/2026-07-15/oneplus-once-popular-with-tech-fans-to-pull-out-of-us-and-europe

Comments URL: https://news.ycombinator.com/item?id=48922412

Points: 3

# Comments: 0

Article URL: https://www.soofi.info/

Comments URL: https://news.ycombinator.com/item?id=48921662

Points: 1

# Comments: 0

Article URL: https://www.dutchnews.nl/2026/07/police-break-up-e100m-a-month-investment-fraud-ring/

Comments URL: https://news.ycombinator.com/item?id=48921656

Points: 1

# Comments: 0

Article URL: https://dailyrung.com/

Comments URL: https://news.ycombinator.com/item?id=48921648

Points: 1

# Comments: 0