We are happy to announce that the **Scottish Programming Languages and Verification Summer School (SPLV) 2025** will take place at the University of Edinburgh during July 21-25.
Thanks to generous sponsorship we are able to subsidise student participation. Registration and scholarship information will be available on our website soon.
Please save the date, forward this announcement to anyone interested, and check the [website](https://spli.scot/splv/2025-edinburgh/) for updates.
We look forward to seeing everyone in July!
@**Malin Altenmüller**, @**Ohad Kammar** , @**Sam Lindley** , and @**Nachi Valliappan**
---
Courses:
Program verification using concurrent separation logic
Robbert Krebbers, Radboud University Nijmegen
A few ideas from distributed systems for PL folk
@**Lindsey Kuper** ( @lindsey ), University of California, Santa Cruz
Application programming with dependent types
@**Edwin Brady** ( @edwinb ), University of St Andrews
Type theory
**Fredrik Nordvall Forsberg** ( @fnf ), University of Strathclyde
Behavioural types
**Simon Fowler** ( @simon_jf ), University of Glasgow
Concurrency theory
**Rob van Glabbeek**, University of Edinburgh
Logical relations for program equivalence
**Filip Sieczkowski**, Heriot-Watt University
Models, programs and bidirectional transformations
**Perdita Stevens**, University of Edinburgh

Could be the perfect hardware to repurpose for an emulated NeXTCube project 🤔
(It's a docking station, but I bet you could fit a raspberry pi in there, and those ports would probably be trivial to hook up)
A few days ago, a client of mine asked me to install an open-source software (which I won’t name for now). The software has only one official installation method: Docker. This is because, as they themselves admit, it has a huge number of dependencies - some quite outdated - that need to be carefully managed and forced into place; otherwise, nothing works.
I tried replicating the same setup on FreeBSD but didn’t succeed, as some dependencies either aren’t compatible or simply refuse to run. I could try finding workarounds, but I can already picture the chaos every time an update is needed.
So, I decided to build it via Docker to get a better sense of what we’re dealing with. The sheer number of dependencies that Node pulls in is impressive, but even more staggering is the number of warnings and errors it spits out: deprecated and unsupported packages, security vulnerabilities, generic warnings- you name it, and there’s plenty of it.
Since my client needs to launch this service but is subject to audits, they want to be fully compliant and ensure security. Given their substantial budget, they offered financial support to the developers (a company, not just a group of hobbyists) to help improve the project either by making it FreeBSD - compatible or, at the very least, by reducing dependencies with critical vulnerabilities. The client was willing to pay a significant sum, and since the improvements would be open-source, everyone would benefit.
The response from the team? A flat-out refusal. They claimed they couldn’t accept any amount of money because many of these dependencies are "necessary and irreplaceable, as parts of the code relying on them were written by people who no longer work on the project, and we can’t rewrite the core of the software.” Then came the part that really got under my skin: they stated they would rather deal directly “with my client, not with me, because in the end, my concerns are just useless and irrational paranoia.”
Translation? Just pay, and you’ll pass compliance checks - never mind the fact that underneath, it’s a tangled mess of outdated and insecure components. And don’t make a fuss about it.
While I can understand some of the challenges the team faces, I might have accepted this response if it had come from a group of volunteers or hobbyists. But if you’re a company whose sole business revolves around a single software product (with no real competition at the moment), this approach is not just short-sighted - it’s outright dangerous for your users’ security and for your own survival as a business.
The result? They lost a paying client who was ready to invest a significant budget into their software. That budget will now go elsewhere. My client is considering hiring developers to build a similar project with better security (they have both the time and the money for it). I’ll do my best to convince them to release it as open-source - at which point, a new “competitor” will emerge in the market.
According to all known laws of DNS, there is no way a fedi instance could be hosted on an .ARPA domain.
Even if you get ahold of a domain like this, it should only be used for PTR records - right?
The instance, of course, federates anyway - because DNS doesn't care what humans think is impossible.
when the men who tore down the stars burned the True History,
they sought in part to erase the original, truer purpose
of the GPU: to Paint -
every color we humans ever imagined,
castles in the sky,
flocks of starships, fleets of pegasus,
rainbows and waterfalls and violet-green fire and pearlescent ooze; everything -
and still we dream of these,
in the songs we pass down
in secret moments of hope
Tried out the new and popular “Deepseek” LLM with my standard “tell me facts about the author of PCalc” query. At least half were misleading or straight up hallucinations. LLMs are not a suitable technology for looking up facts, and anybody who tells you otherwise is… probably trying to sell you a LLM.




My plug-in was written as a precursor to another, an alternative to task wiki, which is still in development. Marksman could potentially replace micro wiki