Sometimes I find myself posting the same reference to people over and over. Other times I keep going back to the same reference. I thought I should keep a list.

  • It’s time for operating systems to rediscover hardware
    I just point to this every time I get frustrated at the Linux monoculture. That’s not really what it’s about, but it articulates a few reasons I have for actually being frustrated.
  • Faster zlib/DEFLATE decompression on the Apple M1
    I did work optimising zlib for a RISC-V CPU (unfortunately I based it on an upstream which didn’t want to maintain RISC-V code so most of the patches languished, and I wasn’t able to reparent them while I still had access to the same class of CPU). Then I discovered this effort which drew most of the same conclusions and produced a very similar implementation.
  • Quasirandom sequences
    This describes something I’d spent a long time thinking should exist but not being able to find. Then I found it. Unfortunately, while it seems mathematically solid, when I did experiments with it I found a very pronounced artefact where one axis cycles between four phases which seem to march slowly (meaning it has a ratio close to but not quite 1/4). Clearly that mathematical purity is not a model of the thing I actually wanted.

Other times I’ve just had a tab in my browser open for so long that it feels like it’s become a part of my life: