Daniel Bartley

@T1SydneyTrains school kids vaping on the train entering artarmon carriage N5604

@mim_djo Lies. Damn lies. And calendars.

i don't like this at all !!!

@tomaspueyo does Wales have massive geographic disadvantages?

@karsyn_bailee Tbf, he looks worse. Everyone looks bad in fisheye

@nrlcentrall “Fire up bitch”

Parramatta have appointed Dragons legend Mark Gasnier as a specialist coach for their outside backs. Gasnier will mentor high-profile recruit Zac Lomax, who joined the Eels seeking a permanent centre role. #nrl #parraeels

@jbvsmo @charliermarsh Asdf?

@charliermarsh uv cannot update itself via uv pip install/sync on windows and having it as a dependency like this is a major blocker for us. We resort to "pip install uv" which is ultra lame... github.com/astral-sh/uv/i… And no, a system wide uv install is not a solution unfortunately

POV: you just migrated to uv for package publishing

@jbvsmo @charliermarsh You can’t install the binary with no dependency on Python?

@Ubunta @daft_dataframe @DataPolars What’s your bluesky handle?

I’m exploring @daft_dataframe to evaluate its performance and complexity in comparison to Apache Spark! At first glance, Daft is optimized for single-node use, but on a single node, I’d prefer @DataPolars for most cases. However, Daft becomes distributed by leveraging @raydistributed, so if you’re already familiar with Ray, there’s no new setup or learning curve for distributed configurations. github.com/Eventual-Inc/D… Daft is ready for Open Data Format and supports @ApacheIceberg writes 🙃, making it a potential go-to for managing Iceberg data! Daft shows impressive speed, though I’m curious if that’s due to Ray’s distribution capabilities or something unique within Daft. It’s too early to say it outperforms Spark, but I’ll continue using it to get a clearer sense!

@catalinmpit What’s your bluesky handle?

@rauchg What’s your bluesky handle?

Some engineering principles I live by: ✓ Make it work, make it right, make it fast ✓ Progressive disclosure of complexity ✓ Minimize the number of concepts & modes ✓ Most 'flukes' aren't… your tech just sucks ✓ Feedback must be given to users instantly ✓ Maximize user exposure hours ✓ Demo your software frequently to fresh eyes ✓ Sweat every word of product copy you render ✓ You're never done working on performance ✓ You're never done. Software ages like milk, not wine ✓ Visualizing traces of time is the best way to optimize it ✓ Ship frequently and strive to build in public ✓ Errors must have globally unique codes & hyperlinks ✓ Red is not enough to signal "error" (8% of men have red-green color blindness)

@ApacheBeam What is your bluesky handle

You can now access the first recordings of the conference on the Apache Beam Youtube Channel. Rewatch your favorite talks and catch any that you missed! A few talks are still pending, so stay tuned for updates once they're published. 📽 🎉 tinyurl.com/bdcthcf5

@pycoders What is your bluesky handle?

PySheets: Spreadsheets in the Browser Using PyScript bit.ly/48vEqv2

@Ezenegwu Is that a dog bed?

@charliermarsh Curious if they learned this from Microsoft’s feedback loops.

Open source dark pattern: someone at a company opens an issue, then a bunch of their coworkers upvote it and comment saying they have the same problem 🤡

@pawjast I agree that pyproject.toml is the starting point for any python configand you can install directly from that file’s specs.

@dbrtly_ You're right in saying there are always cases where you might need to extend or look for a completely different solution. And I'm not saying 'pyproject. toml' is a one-shoe-fit-all kind of solution. But it's a good start.

If you need to package your Python code, there's a tonne of tutorials that tell you to use 'setup .py' and 'requirements .txt'. But unless I'm mistaken, nowadays it's advised you replace these two with 'pyproject .toml'.

@samokhvalov @Database_Lab Gitlab should talk to the team at sdf.com.

Just discovered this great doc showing how GitLab uses @Database_Lab to automatically test all schema changes on full-size (dozens of TiB) thin Postgres clones provisioned in seconds gitlab.com/gitlab-org/dat… This comes in addition to ad hoc use of DBLab for query optimization by engineers: docs.gitlab.com/ee/development…

@pawjast Now try `uv sync`

@rakyll I still like gcp. I just dont want googlers to get bored with it and go looking for shiny things users dont care about. I want gcloud cli as a single binary like aws cli v2. Column labels & expanded trailing comma support in bigquery. An extra 5 gb in cloud shell.

It’s possible that Google’s most major problem is self inflicted gate keeping of its own tech. Earlier Google was a inspiring infra builder because it was easier to build publicly. Network effect created major communities around several of our foundational tech initiatives.

@blast_hardchese @Replit @astral_sh @charliermarsh Is there a blog post about this?

After a bit of burn-in time, I'm happy to share that we switched the @Replit Agent from Poetry to @astral_sh's uv and are seeing significantly faster installs. Thank you @charliermarsh and team!