ByteCode.News
Submit RSS Atom Sign in

Back to the channels

#nevet

  1. jreicher Sorry I've been a bit incommunicado. The Purim madness is mostly over now.
  2. jreicher blue: I have to do some fact checking, but AFAIUI one of the things special about markov chains and associated techniques is that unlike "traditional" statistics you don't need to know/assume the underlying distribution of the population. So for example it might be meaningful to do a "traditional" statistical analysis IN GENERAL of how English words are used, but that doesn't provide a meaningful measure of how I, personally, use
  3. jreicher English. A markov chain can do that by analysing my speech patterns without any knowledge of the distribution.
  4. jreicher This is the kind of commentary I'm thinking of, although what I said is probably not completely correct. https://en.wikipedia.org/wiki/Markov_chain#Physics
  5. nevet Markov chain - Wikipedia
  6. jreicher Also, I think you were saying earlier that AI is really bad at programming. I believe my garlic press is also really bad at programming. It doesn't seem a significant point. I really like my garlic press. It's very useful.
  7. * jreicher joined #nevet
  8. * jreicher joined #nevet
  9. dreamreal https://github.com/jottinger/bytecode.news/issues/91
  10. nevet Star Trader: multi-protocol economic simulation game · Issue #91 · jottinger/bytecode.news
  11. dreamreal Added some additional ideas here, this might be a viable idea
  12. dreamreal aight so has anyone looked at that issue?
  13. Chronos dreamreal: Make nevet scriptable so that things like simple games can be added using that scripting language itself.
  14. dreamreal Chronos: problem is only registration, that's actually super-easy. I already looked at jactl, but the capability is too easy to do - right now, it's classpath-based but there's *nothing* preventing it from being discoverable at runtime besides interest and time
  15. Chronos Never heard of Jactyl before. Reading about it now...
  16. dreamreal https://jactl.io/blog/2026/03/01/jactl-2.4.0-release
  17. nevet Announcing Jactl 2.4.0 | Jactl
  18. Chronos dreamreal: I don't see anything in Jactl that prevents script writers from causing OOM
  19. dreamreal well, you can limit API access, but yeah
  20. dreamreal it's not an actual sandbox
  21. Chronos Could start a second JVM with a strictly controlled max heap, and run Jactl in there
  22. * Chronos works on something at $WORK that can be scripted by customers, thus such things are always on his mind
  23. Chronos Any mistakes customers can make, they do make.
  24. Chronos (Alas, at $WORK, spinning up a JVM for customer scripts would not scale nearly enough)
  25. dreamreal Chronos: I don't like the idea of running user-supplied code in nevet: arbitrary code is bad
  26. dreamreal I don't want to rebuild glot.io or whatever. It's doable, but ... nontrivial esp with later VMs
  27. Chronos dreamreal: Running user supplied code in process seems to be a pretty rare requirement.
  28. Chronos Someday, I'd love to tackle a 100% safe in-process interpreter.
  29. Chronos More of a "when I retire" project.
  30. dreamreal glot.io and ideone do it. I guess I'm not sure what you're wanting in nevet.
  31. Chronos dreamreal: It's largely a nostalgia thing with no place in this modern world. When I was very young, I used a CDC Cyber, and taught myself BASIC on it. It even had some rudimentary multi-user capabilities, so I wrote some multi-user programs. :)
  32. * dreamreal points at star traders
  33. Chronos My nostalgia wants to build something similar for the modern world, even though, realistically, it's just not realistic anymore.
  34. dreamreal guess what system I first played startraders on
  35. dreamreal sentiment java
  36. nevet Error: No recent messages found for irc://libera/java
  37. dreamreal sentiment #java
  38. Chronos dreamreal: Yeah, but just like my nostalgia, who is going to play nevet-based Star Traders in 20xx? Alas...
  39. dreamreal Pffft, have you seen what I was planning on doing with it? :D And Star Traders has a RICH set of descendants... people play Elite:Dangerous every day
  40. Chronos Heh.
  41. dreamreal I loved the cyber-one at FSU
  42. dreamreal unreasonably, really, but even so
  43. Chronos On the Cyber where I stole time (heh), the BASIC had, if memory serves, a SHR keyword, which stood for "share" or "shared", and all instances of the program running shared that variable. I think it was always an array too...
  44. Chronos SHR A$(100) — 100 shared strings, enabling rudimentary multi-user capabilities!
  45. Chronos If memory serves. This was a *long* time ago.
  46. Chronos But pre-teen me figured it out enough to get chat programs and such working.
  47. dreamreal I don't remember that being an aspect of the cyber, but I was just stealing time after school with one of my friends
  48. Chronos On the CDC Cyber, I think it required an optional module called MULTI
  49. dreamreal we played star traders, a dungeon crawler with rudimentary first person rendering, etc
  50. Chronos dreamreal: Ah, fellow timeshare thief. Excellent.
  51. Chronos I stole unbelievable amounts of time.
  52. dreamreal I had a friend whose dad worked in the support lab, so we'd hang after school and play games
  53. Chronos Nice!
  54. dreamreal wouldn't say unbelievable amounts of time but it wasn't actually MINE, I was sort of borrowing my friend David's time which was given to him gratis :D
  55. Chronos Lucky dog!
  56. dreamreal Yeah, David's an imagineer now as far as I know
  57. dreamreal Keep wanting to talk to him, he's one of the best programmers I knew when I was young
  58. Chronos nice
  59. Chronos I'm a mediocre programmer, but *knowing* that helps a lot ;)
  60. dreamreal me too
  61. dreamreal i think I'm all right as a programmer but that's about it: I've been very blessed to know some absolute UNITS in the field though
  62. bot [jottinger/bytecode.news] New PR #150: Bump @hono/node-server from 1.19.9 to 1.19.10 in /frontend - https://github.com/jottinger/bytecode.news/pull/150
  63. bot [jottinger/bytecode.news] New PR #149: Bump hono from 4.11.9 to 4.12.5 in /frontend - https://github.com/jottinger/bytecode.news/pull/149