gedhrel @ gedhrel @lemmy.world

Posts

0

Comments

128

Joined

2 yr. ago

Training, decent modern equipment, understanding and managing risk, gear redundancy, and wear a frickin' helmet (esp. when belaying).

Yeah, there are some yahoos, but there're also people who drop dead playing table-tennis.

With massive OOO pipelines, what's the alternative?

Sounds like a really bad decision - forced error due to some other reason to back out rapidly. What an awful tragedy.

I was under the impression that Musk was going to "fix" Boeing's delivery of AF1. I think his suggestions included relaxing the security clearance for new hires and getting everyone working 60-hour weeks.

News of that all went very quiet after the announcement though.

I think Kermit's cover of Cash's cover was brilliant.

He had a number of unchaperoned calls with Putin. I think it's more likely that he's acting in fear of an unpleasant death; sometimes direct credible threats work wonders.

It means you can take out the bride's party, or the groom's party.

Thousands separator.

I'm not sure if that opening sentence is fatuous or not. What errors in any industrial enterprise are not human in origin?

It doesn't, but if it did that'd explain why there isn't much of it around.

That's actually the idea. It's not general precrime, it's a decision support tool for predicting recidivism when deciding parole cases.

That doesn't mean it's not on decidedly shonky ground statistically speaking.

570 recorded homicides between March 2023 and 2024.

Data on "hundreds of thousands" of people can't provide the distinguishing markers to even have a stab at this.

It can reliably predict when people are black, though.

And as for your specific question: typechecked code doesn't get to production with a type error; it won't compile. There's a common phrase, "left-shifting errors". It means catching bugs as early in the development cycle as possible. In terms of things like developer time (and patience), it's far more cost-effective to do so.

I worked on OpenStack back in the day: millions of lines of untyped Python.

Let's say you've got an X509 certificate. You know you can probably pull the subject out of it - how? Were I using Java (for instance), the types would guide my IDE and make the whole thing discoverable. The prevalent wisdom at the time was that the repl was your friend. "Simply" instantiate an object in the repl then poke at it a bit.

And it's not just that kind of usability barrier. "Where is this used?" is a fantastic IDE tool for rapid code comprehension. It's essentially impossible to answer for a large Python codebase.

Don't get me wrong: python is still a great go-to tool for glue and handy cli tools. For large software projects, the absence of type enforcement is a major impediment to navigation, comprehension and speed of iteration.

Not just domain logic. The implementation logic is often weird too. Cobol systems have crash/restart behaviour and other obscure semantics that often end up being used in anger; it's like using exceptions for control flow, but exceedingly obscure and unfortunately (from what I've seen of production cobol) a "common trick" in lots of real-world deployments.

That's pretty nuch why every one of these cases has settled before it reached court. The first thing the prosecution would do would be to get documetation of how many times this had happened, and the met policy that knowingly encouraged it.

Wait until you hear why Cambridge exists.

Whilst it's gotten a lot better in the -17 and -20 iterations, the fact that there was recently a doorstop book published solely on the subject of C++ initialisation semantics is pretty telling.

I really like what Herb Sutter's doing around cppfront; I still wouldn't use C++ unless I absolutely had to.

To add something to this: linux has avoided internal SPIs for a long time. It's often lauded as one of the reasons it hasn't ossified.

However, some subsystems have a huge amount of complexity and hidden constraint in how you correctly use them. Some of that may be inherent, but more of it will be accidental.

Wrapping type-erased shims around this that attempt to capture (some of) those semantics shines a light onto the problem. The effort raises good technical questions around whether the C layer can be improved. Where maintainers have approached that with an open mind, the results are positive for both C and Rust consumers. Difficult interfaces are a source of bugs; it's always worth asking whether that difficulty is inherent or accidental.

https://lore.kernel.org/rust-for-linux/2025021954-flaccid-pucker-f7d9@gregkh/