This is nonsense. Old software was not fast. Computers used to take minutes to start up. That’s not an exaggeration. MS Word could easily take 30 seconds to open.

Obviously if you run it on modern hardware it’s going to be lightning fast, but not because the authors cared more about performance and efficiency than we do now.

Happy twelfth birthday!

That’s just nonsense. There are plenty of competent nice people working there. You can say that about leadership if you like.

It’s going to have terrible UX, a complete pain to build, the contribution process is going to be some git send-email mailing list nonsense, it’s going to expect you to have read the manual (probably in info just to be difficult) cover to cover before you even consider using it.

But on the other hand it probably has at least decent documentation, it probably works reasonably well, and there’s zero chance of rug pulls, closed source add-ons, etc.

Overall I would say it has negative connotations. If you said “check out this package manager, Fooly”, I’d think “ok might be good, might not”. If you said “check out this package manager, GNU Fooly”, I would say hell no. It’ll be awful.

It’s the software equivalent of books that have “how to read this book” sections.

Vim’s solution to fast editing also isn’t very compelling since multiple cursor editing was invented. You can get 90% of the editing speed by learning 1% of the shortcuts. And the UX is slightly nicer since you get immediate feedback.

I feel like this is due to my C-suite pushing for AI integrations in basically everything

I would put a small amount of money on it actually being because this guy was involved in setting up that workflow and sees your suggestion to fix it as criticism that it is shit (and by implication so is he).

Very common defence mechanism.

I don’t have a strong opinion on the beta site, but I do know that they need to stop listening to the exact people that killed their site (or allowed it to be killed by AI at least).

Actually they should have stopped listening to them a decade ago. Now is way too late.

I haven’t used Java for decades and never used .net so I’ll take your word for those. Absolutely not for C++ though.

Go’s standard library has:

• JSON, XML, etc.
• A web client and server
• Bigints
• Logging
• JPEG, PNG and GIF
• SQL client
• SSH client
• All the crypto algorithms
• Gzip, zlib, deflate, etc.

C++ has none of that. Hell C++ only got a function to check if a string starts with a prefix a few years ago.

I think the only mainstream language with a standard library that is both good and comprehensive is Go. All of the others either have smaller standard libraries (e.g. Rust) or poorly designed ones (Python).

The only reason to do this is if you’re directly integrated Rust into an existing build system (e.g. Bazel). It’s not going to help with this problem at all.

With caveats.

If you actually read his message you can see he is interested in hosting projects that aren’t open source.

I think mmap is unlikely to be the best option seeing as you’d be doing large sequential reads.

Yeah it’s clearly more than just experience. It’s at least experience + actually learning + actually caring. I wouldn’t rule out natural talent either, though I have seen plenty of smart people with poor programming taste.

For example I worked on a C++ SDK where the guy that wrote it was clearly very smart… But he had also written an enormous god object using CRTP to inherit about 20 classes. The aim was to make it somehow modular, but it absolutely wasn’t. Clearly poor taste.

Maybe it’s like religion. There are plenty of very smart people that believe in an imaginary friend. It’s almost orthogonal to “smartness”. Maybe taste is the same.

I mean, sure… But for most users of GitHub actions “it’s easy and free” is a huge benefit, easily outweighing any technical advantages Buildkite might have.

Python’s performance is too poor to do anything serious. Go and look at the screenshots from Pygame. They look like ZX Spectrum era games.

Compare that to something like PixiJS.

I don’t see why you couldn’t have relations in a object model database. Just allow fields that are references to other objects or their keys.

I would teach Typescript. Being able to write the types down and hover things to see what types they are will definitely help them.

I think C would put them off. I also wouldn’t go with Python, in case they want to do things like write games or make websites, which are common tasks you can do with Typescript but not very well with Python.

ORMs are a pain and so is hand rolling SQL queries and doing the mapping manually.

I definitely think there’s scope for NoSQL databases where the database “shape” matches the normal struct style of programming languages. Kind of like how JSON does and XML doesn’t.

But it seems like all we got was MongoDB and Firebase which are both shit.

Are there any good NoSQL databases? MongoDB and Firebase don’t even have schemas.

You don’t need ORMs to prevent SQL injection. Prepared statements have existed for decades.