I've been building Humbug for over a year, and it's one year this week since I released v0.1. For all that time I've been planning for a future v1.0 of Humbug, but no specific version felt like it warranted such a big bang change.
I started Humbug with a view that it could help engineers, product managers, security analysts, etc. all collaborate with AI in a new sort of tool.
That vision has worked out far better than I imagined. I now use it daily for building and testing software, for reviewing design ideas, editing and critiquing specifications, assessing third party code, and even maintaining my blog site. With lots of agentic features, and 100k net-new lines of code (with many hundreds of thousands more that came and went), the v0 prefix has been feeling increasingly stale.
The obvious thing would be to move to v1.0, but I've been starting to realize that aside from being rather arbitrary, I couldn't see an obvious way I'd ever move to a v2.0 in the future. Semantic versioning just doesn't feel right.
The end of SemVer for Humbug
I've used semantic versioning (SemVer) for a very long time (over 30 years) and until now it has served me very well. Everything I built before strove for determinism. Predictable logic delivered predictable results against predictable APIs. Any non-determinism only appeared at the level of the human user.
Humbug poses a new problem, however. We put a non-deterministic LLM on top of the deterministic tools, and now suddenly the behaviour seen by a human user is no longer so predictable. Even worse, AI models change and update so frequently that tool usage becomes more an art than a science. Minor changes in some aspect of a prompt, or innocuous backwards-compatible changes in a deterministic API, can suddenly cause one or more models to behave in new and unexpected ways.
The tools and core mechanisms inside Humbug all fit with the SemVer model, but Humbug itself really doesn't.
At this point there are really only 2 options: calendar versioning (CalVer) or a simple sequence number.
[Aside: yes, I know there are some other more obscure approaches. MAME has stuck at v0.x forever, the "perpetual beta" feel seems wrong.]
Decisions, decisions
If you read my Notes you'll see I like CalVer approaches to many things, but in the end I decided this would get messy. As the pace of releases picks up, CalVer invariably starts to get messy with either a sequence number for the year or month, or else ends up getting progressively more granular. If I'm going to end up with a sequence number anywhere, I might as well just use a sequence number. The bigger the number, the better the version - done!
Starting with the next release (today), I'm dropping the leading zero. The last release was v0.36, and this one is v37. The next one will be v38, etc.
So perhaps Humbug really came of age about a year ago with what I'd now call v1, I just never realized!