I do two things well: Play music, and write code. This is what I do when I'm all out of music.


Stuff here is almost completely boredom-motivated. If you want to see something move up on my list of priorities, you should click on my Paypal button and tell me in the notes what you like to see me working on.

  • DCoN.
  • Lye. The Lybrary is slowly getting fleshed out, and it should be complete enough for useful things soon.
  • Gallium.
    • Galahad. Just about every assert in driver code should be moved up to the Galahad layer and shared between all drivers. This might already be done.
    • Docs. Gallium's docs aren't API-complete or feature-complete.
      • Consider adapting the current C header comments into the documentation, to avoid having to rewrite the original comments in ReST.
      • In addition to a technical overview and API, document some examples or other forms of developer brain dump in order to provide more useful documentation to newbies.
    • Pylladium. Listed here because it relies on Gallium. Needs to be finished, although it might not matter that much these days.
  • PyFluidSynth. Already far ahead of the other Python-FluidSynth bindings, but at some point it should get API-complete.
    • For that matter, Salsa needs to be completed, since that was the original reason for developing PyFluidSynth. Salsa's unfortunately on hiatus, but all the Salsa-related art and music is still on schedule.
  • Linux. A variety of projects await.
    • The LeapFrog Didj and Explorer have open-source kernel patches floating out there. A good, Linux-style set of platform drivers and platform support should be written at some point. Currently stalled on getting my Didj back in working order, or obtaining an Explorer.
      • The Didj and Explorer machine files need to have .video_start/.video_end filled out. This fixes the current problematic video setup, where VRAM isn't known to Linux but is instead reserved by limiting the total system memory. This should allow us to remove the entire configurable memory system and switch to standard RAM auto-detection.
      • The GPIO pins have a strange interrupt-sharing scheme. Other platforms use virtual interrupts and an interrupt demuxer; this is completely viable for Pollux as well.
    • KMS.
      • ATI Rage 128. Can be done inside the current radeon kernel module. Would require userspace updates. I have experimental patches for this that need to be solidified and sent upstream.
      • Voodoo Banshee. Would require userspace updates, but be so gratifying. James Simmons has already started on this. I now have multiple Voodoos, so nothing's really standing in the way.
      • XGI Volari. Supposedly a Direct3D 9-class chipset, meaning that Gallium drivers could be written once the kernel backend is in place. This is definitely a non-trivial task since the documentation for the chipset is sparse and almost entirely in C.
      • Matrox G450 and other G-series chipsets. There's a lot of code consolidation that could happen here. Anybody tackling this would get to experience the joyful adventure of making the Matrox DRI work again as well.
      • SiS USB adapters.


I wrote this stuff, but I'm not touching it at the moment either because I'm not that bored or because it doesn't need updates.

  • Construct. Full-on maintenance mode.


Never touching this again without good reason.

  • Hackabot. I didn't write this, and I don't need to write anything else for it. Mike's maintaining it and doing just fine.
  • Gallium r300 driver. Dave and Marek are maintaining r300g de facto.
  • Kong. Nothing more to write, I think. Additionally, I don't care about it anymore.
  • Tiger. Ditto.
  • alsa-patch-bay. I have no contact from original upstream, no contact from distros that don't have it, etc. The Fedora bug's started moving again.
  • utripper. It's just sitting there, really. Talk with the John guys indicates that there might not be much optimization possible, and more modern trippers seem to be doing GPU-based stuff.
  • ttk. Meh. There's no longer any point to this.
  • Pydra. http://pydra-project.osuosl.org/ and http://code.osuosl.org/projects/pydra
  • Bravo. I can't deal with Minecraft communities any longer. I won't work on it for free.
Valid CSS!