Now

A now page. Updated August 2026. Previously: July 2026.

Sketch of a desk with dual monitors, headphones, and a mug

Compilers and low-level work have turned into a genuinely nice place to be. Most of what I do now sits somewhere between an instruction set manual and a code generator, and that suits me.

The day job is still the system operator role at Cuscal Paymark in Auckland, TACL and COBOL and a fair bit of TAL, and the embedded fintech contract is still on. Both good, the shift work is brutal though.

Being honest about the harder part of this year. The attention around Booth was more than I expected and more than I wanted. A lot of startups got in touch promising a great deal, and very few of them followed up on any of it. The rest was a steady stream of people asking when their particular thing would be supported. I ended up programming from waking to sleeping, sixteen hour days, and having to deliberately make myself go for a walk. The quantum work was good fun and then the same thing happened all over again.

The true cost of that is two finished things I have not released. Kokako, a quantum chemistry suite, roughly Gaussian in a CLI and without the licence that stops you saying so. And a browser engine, which browses a fair chunk of the web already. Both sitting there because I know what releasing them looks like now. I will put them out, but on the understanding that they are things I made rather than products I am staffing.

Quantum itself was interesting and I am glad I went there. I got a Fortran runtime PR into qiskit-fortran and learned a lot. I do not think it is mine, though, so Ernest is going quiet rather than getting a roadmap.

One big lesson from this year, learned a few times now. If you want to build something inside somebody else's project, turn up with the patch rather than the question. I asked first, specifically and politely, partly so I would not duplicate anyone's work. Six weeks of silence later the work existed and I had not done it. An issue is a gift to whoever has more time than you, and in a project with paid full-time people that is always them. It was a bit disappointing because I was rather looking forward to it and it's put me off contributing for some time.

Where I am actually pulling is backwards. I have been enjoying BAL on the Apollo guidance computer project, and HLASM, and designing an emulated VSAM for z390.

Reading a lot of nuclear physics at the moment so I can take Moa past neutrons.

I've also been reading into Grace Hopper's work and have found some preserved copies of Math-matic and Arith-matic. I've already made Flow-matic usable for Python and I feel more confident in my compiler work so would love to get some of it running again.

The projects ticking along

Away from the terminal