[–] [S] 1 point 19 hours ago

What it amounts to is a very neat technical trick, in the sense of a dog walking on its hind legs while barking 'Puttin' on the Ritz'. It's neat. It's not something you see every day. But it's also no basis for an industry, let alone an economy.

  • source
  • [–] 1 point 19 hours ago

    @glyph What drove me to walk away from it was realizing and internalizing that all the neat software toys in the world aren't worth much if the things you make aren't things other human beings want to use.

  • source
  • [–] [S] 1 point 20 hours ago (1 child)

    What I picture is, the OS/kernel/whateveryoucallit lives in NOR flash along with the FPGA's bitstream. User programs are limited to ~500KB for code, but can store non-executable data in the PSRAM as needed -- sprites, tiles, music samples, whatever.

  • source
  • parent
  • context
  • [–] [S] 1 point 20 hours ago (2 children)

    The more I study my own design, the less inclined to emulate the Commander X16 I become. The CPU performance is insane; even hard-limiting a user program to 500KB, you get two 150300MHz ARM or RISC-V cores to play with. Why waste that performance emulating a 65C816 when your game will be in C anyway, and you can get many times better performance running natively?

  • source
  • parent
  • context
  • [–] [S] 1 point 20 hours ago (3 children)

    Most RP2350B designs use HSTX to implement DVI over an HDMI connector for video out, with an in-SRAM framebuffer and the firmware running from NOR flash XIP. No problem there, but it's not how my design works. No, the video output has its own 128KB of VRAM inside the VERA FPGA, so the whole 520KB of SRAM is available for use.

  • source
  • parent
  • context
  • [–] [S] 1 point 20 hours ago* (4 children)

    The firmware can execute-in-place from the external NOR QSPI flash via a 16KB SRAM cache. No real big speed penalty to be had there. But if your OS wants to load a program from SD and run it from PSRAM with the same caching rule as flash (which you totally can do!) you then have the OS and the user application fighting over the cache, at least as I read it.

  • source
  • [–] [S] 1 point 21 hours ago (1 child)

    We might not have bought into the whole Free Software Foundation line of bollocks, and gone en masse with FreeBSD. Or maybe the FSF itself would have adapted and kicked out the sex pests and nonces and creepers. Yeah, right.

  • source
  • parent
  • context
  • [–] [S] 1 point 21 hours ago (2 children)

    Back in the day, we didn't even stop for a moment to consider the values of the people who made the software, or how they did it. Didn't even occur to ask.Looking back, some things might be better if we had done.

  • source
  • parent
  • context
  • [–] [S] 1 point 21 hours ago* (4 children)

    What people want is software the origins of which align with their own values, whether they made it themselves or got it from others. They value that above efficiency, speed, functionality, even security.

  • source
  • parent
  • context
  • [–] [S] 1 point 21 hours ago (5 children)

    But I learn from my mistakes. That's the important bit, here. What I learned was that where I ranked functionality above all, efficiency almost as high, others valued other things more -- and the opinions of the one who makes the thing aren't the ones that matter in the end. Software with no users might as well not exist at all.

  • source
  • [–] [S] 1 point 23 hours ago (5 children)

    @MuseumJoe I try not to beg except as a joke. I complain a lot with no expectation that anyone do anything because I have nowhere else to vent. But man do I appreciate it when people decide to help out. I wish I could do it for others.

  • source
  • parent
  • context
  • view more: next ›