Dev.to author Karm Patel writes about different standardized ways to go about git branching for different software development workflows in 2025. From Git Flow to Forking Workflows, each is evaluated exposing their pros and cons weighting each approach's complexity vs its contribution to best modeling your workflow.
"Git Branching Strategies: A Comprehensive Guide"
In the responses to my #rp25 talk this year I got a lot of "Now digital rights organizations are libertarian? When did that change?" and I am sorry to say that digital rights discourse has been imbued with if not based on libertarian belief systems since the beginning. That is the heritage we need to move beyond.
I think I did only two things that really mattered to that org in the long term:
1. Stopped them from buying a really dubious CMS / web app engine that would have cost the org (a nonprofit) millions it didn’t have to be vendor-locked into a dead-end platform that would have set their site back years.
2. Selected a new developer hire who was far more dependable and mature than I was, and became their lynchpin. She wrote tons of software that did actually matter, and stayed with them for years.
Whatever dev skills I had turned out to be good for (1) saying “stop, don’t” and (2) recognizing an excellent person in a job interview. Note that neither involved the code I wrote.
2/2
Prediction of Geoeffective CMEs Using SOHO Images and Deep Learning / First Observations of a Geomagnetic Superstorm With a Sub-L1 Monitor: #SolarStorms: https://science.nasa.gov/science-research/heliophysics/nasa-missions-help-explain-predict-severity-of-solar-storms/
Rust's Stable MIR project is being slept on. It's pitched as a tool for formal verification, but it's more general than that. Once stabilized, it will almost be a stable compiler plugin API!
It doesn't just include what you get from `--emit=mir` (IR for each function), but also has full type and trait information, and will likely even support interactive queries like "does type implement trait”.
In short, it'll be perfect for metaprogramming!
[1/2]…
LLMs are 3D printers for code. Sure, the output is nowhere near as robust or precise as what you'd get off a CNC mill, and it might take 5-6 tries to get something that works, but it enables local production where none would previously have occurred.
“TNC is where it all comes to life.”
In this video message, our CEO Lise Fuhr shares excitement ahead of her very first TNC, and why this global gathering matters so much to our Research & Education community.
Watch her personal welcome message below, as we all get ready for next week’s #TNC25 in Brighton 🇬🇧
And if you’re not coming in person, don’t forget to g…
this *sounds* like SFMTA agreed to painting full 20-foot red curbs at every intersection, rather than the mere 10 feet at stop-controlled intersections that has been traffic engineer Ricardo Olea's policy. if so, good! parent organizing worked!
https://www.k…
After the whole "I asked ChatGPT" as talk opener I've recently seen a lot of "Look, my kids are using AI to build their own games and that's beautiful" stuff in presentations.
Makes me sad that instead of wanting kids to learn how to build something they get taught to accept what the kinda-passable code generator craps out. What they learn is not how to conceptualize or build something, what they learn is that shit comes from nowhere if you just match your e…