Show more

We're getting quite a few people who thought the project was dead for some reason... Remember that you should follow our blog (and this account) for updates.

We released the first open-source accelerated GPU drivers for Apple Silicon just two months ago:
asahilinux.org/2022/12/gpu-dri

That's the world's first open source driver for Apple GPUs, and also the world's first Linux GPU driver written in Rust, but for some reason it didn't get almost any media attention...

But then we sent out some device trees for Linux 6.2 and that did?

We don't get it...

About yearly, we seem to get another article saying "look! are the biggest contributors to ". These articles come across as PR pieces since they point to data saying who is the biggest contributing company without putting it into context. Like, does contribute 0.5%, 5% or 50% of all activity? Instead the talk about things like "the top leading corporate contributors", leaving out non-"leading" contributors, non-corporate contributors, etc. theregister.com/2023/02/24/who

An internet toll will penalise consumers & breach #netneutrality. Today, Dutch Economic Affairs Minister Micky Adriaansens warned against the demands of Big Telco. She calls for a differentiated analysis of the problem first, before jumping at solutions. 🧵
reuters.com/technology/dutch-w

If your city couldn’t avoid making a massive city-damaging mistake in the first place, do the next best thing — remove the mistake, and repair your city. Like #Utrecht, Netherlands did with its central area river, then freeway, then river again.

(Image via @BicycleDutch) #Dutch #Holland #Netherlands #freeways #cars #bikes #cities

"The idea that complex social problems are amenable to cheap technical solutions is the siren song of the software salesman"

Re the Online Safety Bill/regulatory proposals invoking harm to the vulnerable as a problem mass surveillance could solve.

bbc.com/news/technology-645840

client is configured with two repos: Maven Central and the Google one. Yet running `./gradlew buildEnvironment --scan` downloads `org.gradle:gradle-enterprise-gradle-plugin:3.10.2`, which is not available on those two repositories. It seems that is adding repositories automatically, that seems sketchy to me. I confirmed this by running `gradle --write-verification-metadata sha256 buildEnvironment --scan`

Wiki Unblocked is also built by using the process, independently confirming that the binary APK shipped on f-droid.org matched the source code.

Show thread

Seeing the utopia that is promised just around the corner with AGI @clarkesworld closing submissions because of inundation by "AI" generated submissions. Feels like a DDOS attack.
dair-community.social/@clarkes

World: Can the US please just use the metric system, instead of making up weird measurements no one else understands

US:

If you use Debian container images, please note that "debian:bookworm" images are already using deb822-style repository sources manpages.debian.org/bullseye/a

Something that stuck with me from a previous job is the quote: “don’t underestimate things that have survived many attempts to kill them.”

Think: DNS, bash, C, TCP.

These things have survived this long for a reason. Find out the reason.

@cryptax if you post a droidlysis v3.4.1 ASAP, I can probably get it into the upcoming Debian/bookworm release. Also, I found a bug when using newer libmagic: github.com/cryptax/droidlysis/

Show thread

Just uploaded to the key inspection tools 2.7.0 and the latest from git, ahead of 2.5.2. All sorts of tools like and more rely on these for inspecting Android APK files.

The EU digital identity wallet might handle some of citizens’ most sensitive data. Its success highly depends on the trust people place in it. Undemocratic behaviour & the deletion of privacy-preserving features of the new ID are certainly the wrong way to gain society’s trust.

Show thread

Software deployments with many active engineers can work effectively by constantly deploying iterative changes and watching for feedback. That is a pattern used by many large software companies and startups alike. But that's not the only effective model of software development. Many projects still use stable releases since they allow progress without requiring constant attention. Once a stable release is deployed, it can be effectively maintained with a drastically smaller effort.

As a committer on a couple of medium-popularity projects I couldn't agree more with this:

"Maintaining a successful open source project is Good Will Hunting in reverse. You start out as a respected genius, and end up being a janitor who gets into fights."

— Byrne Hobart (@ByrneHobart@twitter.com)

I spent the morning poring over the Android 14 DP1 materials, especially the API differences report. Which means, once again, I have some random musings: commonsware.com/blog/2023/02/1

As usual, the early developer previews have the most stuff to report on. Thankfully, these posts get shorter as we march towards beta releases.

And, as usual, I don't pull many punches.

All the hype about souped up developer productivity using LLMs for coding reminds me of the original title of this 2014 paper, before it was milquetoasted in 2015 acceptance.

LLMs can help you rapidly acquire semi-plagiarized fragments of well-traveled code instead of using a quality library with vision of the problem domain. Might be great for KPIs, but this debt will come back to bite you, unless you're already gone. Will be painful for orgs to adapt.
research.google/pubs/pub43146/

Show more
image/svg+xml Librem Chat image/svg+xml