Hi everybody,

I've now established that Ooloi’s foundations actually hold when pressed into the places where notation software traditionally breaks down. That meant staying with semantics, architecture, and validation longer than is fashionable, and accepting that this is simply how this kind of work proceeds. That work is now done. It took three months, longer than I expected, but it was absolutely necessary.

A sequence of blog posts came out of this period, each dealing with a different pressure point:

• Re-reading Read – https://www.ooloi.org/home/re-reading-read

• The Belly of an Architect – https://www.ooloi.org/home/the-belly-of-an-architect

• The Eternal Return – https://www.ooloi.org/home/the-eternal-return

• Breather – https://www.ooloi.org/home/breather

• Computer Science for Musicians – https://www.ooloi.org/home/computer-science-for-musicians

• Cross-Platform, Verified – https://www.ooloi.org/home/cross-platform-verified

• Knuth–Plass – https://www.ooloi.org/home/knuth-plass

• Twice – https://www.ooloi.org/home/twice

Two results stand out.

First, remembered accidentals — long treated as 'mostly solvable' — turned out to be fully deterministic once the surrounding semantics were right.

Second, measure distribution across systems and pages — another traditionally heuristic area — reduced cleanly to a Knuth–Plass–style dynamic programming problem once layout was properly decoupled from meaning. This is now formally specified as ADR-0037: Measure Distribution Optimisation.

Alongside this, collaborative sessions and hybrid transport are now fully specified as a foundational property of the system, not a bolt-on: ADR-0036: Collaborative Sessions and Hybrid Transport.

This phase also clarified something about how I need to work. I’ve stepped away from all social media, including LinkedIn. Not as a gesture, but because this project cannot be built under the dopamine-driven, 'move fast and break things' rhythm so fetishised by Silicon Valley.

Ooloi’s motto is the opposite, and always has been: move slowly and mend things.

What remains on the semantic side is small: one final step in hierarchical key-signature handling. After that, the focus moves fully to the frontend foundation. Compared to what is now in place, that work is concrete and bounded.

Happy New Year, everyone. I'm looking forward to 2026.

/ Peter Bengtson
ooloi.org