A silly little TIL trick I learned recently. Comes in handy when debugging time-dependent races and such.
Posts by Chris
Exploring Apache Iceberg and SlateDB formats - with a repo link for additional exploration.
datapapers.substack.com/p/exploring-...
If you formed your opinion of agentic coding more than four months ago, I strongly encourage you to reevaluate.
Stu's talk is now available for anyone who couldn't attend South Bay Systems to watch: youtu.be/TeFsBVIYBis?...
I highly recommend watching it. It's ~30 minutes and goes into details on how ParadeDB achieves faster-than-Postgres JOINs.
🎂 Commonhaus Foundation Marks Two Years as Open Source’s ‘Missing Middle’ — and Maintainers Are Taking Notice.
Commonhaus’ 2nd anniversary
www.commonhaus.org/activity/305...
I think it's a both/and. :)
New post! I used to think the future was formal specs + model checkers. Now I think it’s English + tests + quality gates.
Welp, off to the factory to shovel tokens into the furnace!
1/4. Today, a bit of a different blog post. James from our team sat down with @fulmicoton.bsky.social, the main author of Tantivy, to discuss his journey and how he created such a world-changing project. 🧵
I am an open source maintainer. Nearly all code I write is AI generated. Nearly all PRs I'm receiving are AI generated. This is the world we live in.
New post! I now see open source projects as materialized agent output–cached code. Its purpose is to save me time and tokens.
Change is hard.
I started picking up on this vibe a few months ago. I started feeling it myself in the past month. Something is happening. I don’t know what it is, but it’s exciting.
rng.md/posts/someth...
Ah it appears you just fixerd it 👍
@buttondown.com your dashboard, archives, subscription links are fully offline.
🙏 Thank you! I really appreciate these kind words. It can be a bit nerve racking working through this stuff in isolation for so long. :)
I know a lot of you probably read or at least own the first edition of DDIA, but don't sleep on this, because it's a major update. Sixty new pages, chapters dropped, reordered, rewritten. @martin.kleppmann.com and @chris.blue have done a master job ensuring the 2nd edition up-to-date and relevant.
Spent some time with UniFFI this weekend. Experimenting with it for SlateDB's official bindings. It seems really nice! Anyone out there using it? If so, what has your experience been?
github.com/mozilla/unif...
Pleasure working on it with you! I'm looking forward to my copies. :)
A print copy of Designing Data-Intensive Applications, 2nd edition
Just received my author copies of the second edition of DDIA! After years of seeing it only on-screen, holding the final print book in hand is very special. @chris.blue
Feedback need: SlateDB's Python API uses pyo3, which supports Python/Rust async integration. The other bindings use slatedb-c (C FFI ABI), which is sync. Ideally, I want all bindings use one layer (FFI) to simplify maintenance. Should I ditch pyo3 and async Python support, tho?
"A background job queueing system built on top of object storage via slatedb" 🤩
Was just talking about the need for this with @almog.xyz today!
github.com/gadget-inc/s...
To serve, stick around.
Excited to enable pipelining by default in s2-lite with this upgrade! Expect lower latency and higher throughput.
github.com/s2-streamsto...
Finally, we published an official Java binding on maven central:
central.sonatype.com/artifact/io....
We did a full rewrite of slatedb-c to make this possible. All official bindings will use slatedb-c's FFI bindings moving forward. This gives us a single stable foundation.
A new slatedb-txn-obj crate is now available. Use it to implement transactional objects with epoch-based fencing, retries, conflict merging. It's a critical component of SlateDB, and it's now available to everyone.
crates.io/crates/slate...
The `DbReader` also got some attention:
- Added `DbReaderBuilder`
- Readers now support on-disk object store file caching
- Readers can preload (warm) caches
- WAL object store support
Writes now return a `WriteHandle` that contains the sequence number of the write. This is a first step in adopting time-travel reads on SlateDB. Users can also tie sequence numbers to other identifiers such as PostgreSQL LSNs or Kafka offsets.
slatedb.io/rfcs/0018-ex...