User avatar
James Scholes @jscholes@dragonscave.space
2d
If the way LLMs bump version numbers in vibe coded projects is anything to go by, people have largely failed to use semantic versioning as intended. A patch release wasn't meant to carry a note like, "and it's a big one," not that I think most software projects should use SemVer in the first place.
2
2
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
2d
@jscholes I dunno. Even before vibe coding I pretty much assigned version numbers randomly. I never understood how they're supposed to work. In my head things starting with 0 are alpha/pre-production, and then anything else is really just about how important or big or exciting I feel a change is. It's all vibes, man.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
2d
@jscholes Like polyglot-secure is on 0.0.4 because, you know, two zeros! That means it's really early days. I'll change the second 0 to a 1 when it feels, you know, like a beta. And then the first zero becomes a 1 when it "feels" ready.
1
0
0
0
User avatar
James Scholes @jscholes@dragonscave.space
2d
@fastfinge Exactly. My assumption is that the training data was full of projects taking a similarly hand wavy approach, inexperienced vibe coders don't know to question it, and it turns out not a lot of supposedly real devs cared about it anyway.

So SemVer might be dead. Long live the next thing which will revolutionise software versioning if only we can agree on any aspect of it for long enough.
1
0
0
0
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
2d
@jscholes Unfortunately, we can't do versioning by date, because vibe coding means we're all making 485 releases per day.
1
0
0
0
User avatar
Andre Louis @FreakyFwoof@universeodon.com
2d
@fastfinge @jscholes 487, actually.
0
0
1
0