sam_bristow

Stealth mountain.

I've used an even less-kind framing in the past: incompetence is multiplicative. Two half-wits together gets you a quarter-wit.

Are they planning to pull a Jimmy Hofa with Jesus...

A mechanically-minded man reasons that the rules of grammar are fixed by certain, almost mathematical principles. By exploiting this idea, he is able to create a mammoth machine that can write a prize-winning novel in roughly fifteen minutes. The story ends on a fearful note, as more and more of the world's writers are forced into licensing their names—and all hope of human creativity—to the machine.

Wait, are these the same GPUs that were diverted from Tesla into xAI a couple of years ago?

The linked page seems to think it does.

"As of June 29th 2020, CTA, the CERN Tape Archive, started to be operated as the successor of CASTOR and gradually replaced it."

I think that a large part of Atlassian's problem is their network of extension vendors. Every time you look at a feature request for simple things like "as a user I'd like to see a total of the worklog hours I've added to tickets this week" you get a chorus of vendors popping up to sell you their $10/month plugin that approximately covers your request.

If Atlassian actually fixed these papercuts their product would be better but they wouldn't be getting a cut of all subscriptions from the dozens of plugins people install to make it a usable tool.

Someone should chuck a clean-up ticket on the backlog.

It's a completely different failure mode, but this reminded me of an infuriating situation I had a while ago with single-use URLs and safe-links functionality in MS Outlook. I can't remember the exact details but the gist was that Outlook would replace the actual URL with an indirection which would scan the page to check it was safe when you clicked. The result was that every time I would request an activation link it would be expired by the time I clicked it. I ended up having to reverse engineer the safe-links encoding algorithm to get the original URL. I wasted hours.

No, but it seems to be par for the course for enterprise software.

Any idea if their internal version has improved dramatically since they stopped maintaining the public version?

What does Facebook use internally these days. I'm amazed that the state of review tools is still at or behind what we had a decade ago for the most part.

I believe that once the OpenBSD team started cleaning up some of the other gross coding style stuff as part of their fork into LibreSSL that even fairly simplistic static analysis tools could spot the underlying bugs that caused heartbleed.

I'll give you most of those criticisms, but I'm a little surprised you think our tax system is overly complicated. For the vast majority of people it's pretty much just a progressive PAYE income tax handled by your employer and a flat 15% GST/VAT on purchases without all the carve outs that seem common elsewhere.

Genuinely curious what I'm missing.

Assuming your discomfort is around the defense side of things here's an example of a diving reberather system using Spark/Ada from a while back.

https://youtu.be/zL9vVs5vHuQ?si=-thG-FkelkW6oFfb

One of my the papers I share around a lot is "Nobody ever gets credit for fixing problems that never happened (2002)"[1]. I like it because it's not purely about software so the examples resonate better with some exec level people in other teams I work with.

[1]https://ieeexplore.ieee.org/document/1167285

I currently pay for YouTube premium but I'm strongly considering stopping again. For me it's a combination of prices creeping up (small part) and the worsening UX and engagement-bait (big part). It's the same reason I dropped Spotify a few years ago.

2025 had a bunch of bits of fun mathematical trivia:

45² = 45 x 45 = 2025

Also,

9² x 5² = 2025

40² + 20² + 5² = 2025

My favourite?

1³+2³+3³+4³+5³+6³+7³+8³+9³ = 2025

I imagine pink is considered too close to red tail lights.

Honestly, if I was writing some code that depended on dicts being ordered I think I'd still use OrderedDict in modern Python. I gives the reader more information that I'm doing something slightly unusual.

One key point that people overlook with that paper is that they were applying the coding standards retroactively. Taking an existing codebase, running compliance tools, and trying to fix the issues which were flagged. I think they correctly identified the issue with this approach in that you have all the risks of introducing defects as part of reworking the existing code. I don't think they have much empirical evidence for the case where coding standards were applied from the beginning of a project.

In my opinion, the MISRA C++ 2023 revision is a massive improvement over the 2008 edition. It was a major rethink and has a lot more generally useful guidance. Either way, you need to tailor the standards to your project. Even the MISRA standards authors agree:

"""

  Blind adherence to the letter without understanding is pointless.

  Anyone who stipulates 100% MISRA-C coverage with no deviations does not understand what the are asking for.
  
  In my opionion they should be taken out and... well... Just taken out.
    - Chris Hill, Member of MISRA C Working Group (MISRA Matters Column, MTE, June 2012

"""

Back in my undergrad days I built a similar clip out of a broken clothes peg, hot glue, and some 2.54mm headers. It worked well.

I think the web itself would be stronger if it was served in pure HTML and not 95% created by JS SPAs.

Stoke Space[1] uses a similar system, letting people write arbitrary code to generate a static configuration for their launch vehicle. It means you get all the power of something like Python during development but also a deterministic, bounded config for the critical flight systems. I think their config files are just TOML that is consumed by Rust.

I'll try dig out a link to the talk one of their Flight Software Engineers did on the concept.

[1] https://www.stokespace.com/

I've only had a cursory look at Varlink, but it almost felt too simple. In particular the lack of unsigned or sized integers.

This might enf up being be fine, but it gave me pause when I looked at it previously.

I found Hillel Wayne's series of articles about the relationship between software and other engineering disciplines from a few years fairly insightful [1]. It's not _exactly_ the same topic but a lot of overlap in defining wht is "real engineering".

[1] https://hillelwayne.com/post/are-we-really-engineers/

Wasn't there a phenomenon with the GPUs being retired from crypto mining operations being basically cooked after a couple of years. Likely because they weren't keeping temperatures in check and just pushing the cards to their limits.

Multi-version approaches to developing software aren't as good at reducing common-mode failures as many people expect[1].

[1] J. C. Knight and N. G. Leveson, “An experimental evaluation of the assumption of independence in multiversion programming,” IIEEE Trans. Software Eng., vol. SE-12, no. 1, pp. 96–109, Jan. 1986, doi: 10.1109/TSE.1986.6312924.

Yes. Synology introduced the requirement to use first-party drives earlier this year and it was such an unmitigated disaster for them that they rolled it back just a couple of days ago.

https://arstechnica.com/gadgets/2025/10/synology-caves-walks...

One feature I wish the Arch wiki had last time I used it was conditionally hiding sections. It presented various options throughout their guides and depending on which options you chose later sections weren't relevant. I often found I'd get partway through a step only to discover it wasn't relevant.

It would be great if, when presented with different options, you could indicate which one you'd selected and have it hide the irrelevant stuff further down the page