Alan Barrett

Introducing PoliceData ZA 🇿🇦 Free, open crime statistics dashboard for South Africa — built because SAPS publishes data as PDFs that nobody can actually use. CrimeStatsSA charges for this. Ours is free. policedata.online

Happy to publicly announce that MasqFit is now available in the iOS App Store! apps.apple.com/ca/app/masqfit…

"I calculated that civilization needs just 50 machines to build everything from scratch. And what people can't believe, is that I posted the full plans, designs, instructions and how anyone can build these machines for themselves."

Headteacher of Hayling College [UK] introduced HEPA filtration units into classrooms 2 years ago. Goal was to reduce illness, keep children & teachers in school & support learning. According to school’s attendance data, the impact has been striking portsmouth.co.uk/education/the-…

@poiThePoi @Duderichy Not built in, but there's @litestreamio which lets you continuously stream updates to multiple replicas without pausing the main database. Behind the scenes it's reading WAL files. litestream.io/how-it-works/

@Duderichy Does sqlite have replication?

How to pick a DB: Do you need users and multiple thread access? yes -> postgres no -> sqlite

@esrtweet Before CVS there was RCS, and before that was SCCS. I agree that they were not suited to the needs of distributed developers.

Another key point (I'm leaving notes for myself because I'm going to turn this into a longer narrative history) is that we didn't have any form of version control in the early days. CVS, which was the first version control tool matched to the needs of networked development, only became generally available after 1990.

Calling all ancient hackers. I am about to sketch a history of the beginnings of distributed open-source development. I was there for much of it, but I'm not sure I have all the facts - I wan't everywhere. Correct me in the comments. There's a model of distributed open source development that we take very much for granted nowadays. It consists of passing around patches among a geographically distributed, decentralized group of developers. This had to be invented, and it has a timeline. There was a nebulous early period of academic software collaborations with source passed around by email and distributions published on FTP s. One of the very early ones was Donald Knuth's TeX starting in 1978. These projects were hosted at particular institutions, with patches from elsewhere relatively rare simply because at the time wide-area networking was in its infancy. To join one of these projects you had to know someone on the development team in meatspace. There were no software forges yet, and even the patch utility did not exist - it wouldn't be invented until 1984. The smallest unit of code you could pass around was an entire source file. In 1979 USENET news was invented. This gave people who didn't have access to the ARPANET a distributed bulletin board using batch transfers over relay networks made from copper telephone wires. It rapidly became customary to publish source code in dedicated newsgroups; outside contributors could email back patches which might appear in later releases of the code. The convention of having a README file propagated through USENET, though it can be traced back further to DECUS source tapes. All-caps names for other metadata files like NEWS and TODO started there too. The heavy-duty development was mostly happening on the ARPANET, though. Several projects in the early 1980s began to systematize distributed development via email and FTP. BSD Unix, the X window system, and the GNU projects. This is when I got involved starting in 1983; I published some things on sources newsgroups, and was an early GNU contributor, particularly to Emacs. The whole culture was very small then then, no more than the low hundreds worldwide. The next step was a subtle one; nethack. This seems to have been the first project for which the location of the FTP server was an accident. It wasn't tied to any particular institution or development group, and the developers were truly geographically distributed rather than being a relatively tight-knit group at one institution with occasional contributors from elsewhere. Nethack was, as far as I know, the first distributed development project about which it is not certain that any two of the developers ever met face to face - you just joined a mailing list. I was a nethack dev. We didn't understand at the time that what we were doing was socially innovative. Starting in 1990 Linux development would pick up nethack's model and put rockets on it. Then beginning in 1999 with SourceForge, software forges happened. And that brings the early era of distributed development to a close.

Don't worry: time isn't broken. NIST's backup systems kept providing accurate time even during the power outage in Boulder. Our clocks drifted about 4 microseconds or millionths of a second, which we will correct when power is fully restored.

There's an entire parallel scientific corpus most western researches never see. Today i'm launching chinarxiv.org, a fully automated translation pipeline of all Chinese preprints, including the figures, to make that available.

It is done. Here are clickable links of over 400 books in the western canon, separated by age group. It's meant for all of you who want the best for your child's education and for your own as well. The good news? Emily Dickinson was right when she said of books: "How frugal is the chariot that bears the human soul." Link to the library below: ⬇️

@JustinWolfers Were the shooters in compliance with existing Australian gun laws? (Were the weapons licensed to the shooters?) If not, then how would changing the rules affect the behaviour of people who are known to break the rules?

Just want to share with my American friends how the Australians respond to a shooting tragedy. Action, rather than thoughts and prayers.

1000ppm CO2 is bad. The paper linked from the quoted tweet tested commercial pilots' performance at 700ppm CO2, 1500ppm, and 2500ppm. I wish that they had tested more than those three levels.

This is the kind of measurement that should be published for all air filtering fans that are intended to be used close to the face.

@RhysSullivan Aider (aider.chat) seems fairly successful at prompting the underlying models to output diffs. It sometimes has to iterate when the format is wrong.

anyone know why cursor & other coding agents prefer to regenerate files using a ton of tokens rather than just running a find and replace? have tried to adjust this behavior with rules, prompts, etc but they really love sticking to it for some reason - feels token inefficent

We have been gagged! We cannot say by who or why. Attempts to silence public-interest journalism threaten everyone’s matters of public interestt. We’ll challenge this order with our attorneys at PowerLaw Africa. More information to follow. opensecrets.org.za

This is a terrible idea. It's OK to give a warning, or have a hard-to-find setting that needs to be changed, but bad to prohibit users from installing apps at their own risk.

No one, no matter their age, should have to hand over their passport or driver’s license just to access legal information and speak freely. eff.org/deeplinks/2025…

Zero‑Knowledge Proofs let users prove they are over 18 without exact birth dates, but they don’t stop verifiers from collecting information like your IP address. These alone aren’t enough to protect privacy and shouldn’t be pushed forward without proper protections in place. eff.org/deeplinks/2025…

Addendum: It's important that privacy goes both ways. No site should know your ID, and nobody with your ID should know which sites you use, nor be able to figure it out. Zero knowledge, as I said in another post.

Governments will not give up on their nasty proof of age ideas. We need a technological solution that preserves privacy while proving credentials such as age. Let them know that your age is more than N years, but don't let them know your name or other identifying information.

Of course the people who want online censorship and ID verification have their own "global network" for coordination, and of course their three-year plan includes "Building regulatory coherence across jurisdictions".