2022-12-10 10:48:08 <KipIngram   > Well, the general notion of race conditions is absolutely pervasive in digital logic design; this  has always led me to think that someone trained in that area would find thread programming "less daunting" that someone that came up through the standard software training path.  That shift from single-thread programming to multi-thread programming is indeed a serious new hurdle; suddenly they have to think
2022-12-10 10:48:10 <KipIngram   > about stuff they've never had to think about.  But digital guys have been thinking about that kind of thing from the start.
2022-12-10 10:48:39 <KipIngram   > Anytime you have multiple signal paths between two points in a digital circuit, you have the prospect for a race condition.
2022-12-10 10:48:50 <KipIngram   > The general term used is "reconvergent fanout."
2022-12-10 10:49:36 <KipIngram   > There are other similar hurdles in digital design, like signals that cross from one clock domain into another.
2022-12-10 10:49:52 <KipIngram   > There the prospect of metastability has to be dealt with.
2022-12-10 10:50:31 <KipIngram   > There are standard techniques for doing the hardware design that help minimize the problems associated with these things, but they carry overhead.
2022-12-10 10:50:43 <KipIngram   > I'm quite sure that's also the case in thread programming on the software side.
2022-12-10 10:51:07 <KipIngram   > If you've got a model that makes it a lot easier to avoid race conditions, etc., I'd bet dollars to donuts that it carries an innate performance penalty.
2022-12-10 10:51:29 <KipIngram   > Which doesn't mean you don't use it - for a lot of applications it's likely the way to go.
2022-12-10 10:51:56 <KipIngram   > But I imagine there are *some* applications, the ones where you need absolutely the best performance you can get, where you would want to face the challenges of thread programming.
2022-12-10 10:52:06 <KipIngram   > If you don't even have the ability to do that in your system, then you are limited.
2022-12-10 10:52:45 <KipIngram   > So I'd likely lean toward supporting both models.
2022-12-10 10:53:25 <KipIngram   > I don't like ruling out possibilities from the jump - especially possibilities that represent best performance options.
2022-12-10 10:54:27 <KipIngram   > But, of course, "processors will be faster next year - it won't matter" yadda yadda yadda.  Same story that always gets used to justify building new overhead into systems.
2022-12-10 11:05:18 <KipIngram   > Also, re: Forth, I strongly believe that Forth should never "wall off" capabilities.  It should provide "raw access" to anything that is possible.
2022-12-10 11:05:49 <KipIngram   > Forth shouldn't be in the business of trying to protect you.  It should, though, let you code up your own protection systems in an effective and efficient way, if you choose to.
2022-12-10 11:06:23 <thrig       > enough rope to make a fancy braid or to hang yourself with
2022-12-10 11:06:29 <KipIngram   > Yes indeed.
2022-12-10 11:06:49 <KipIngram   > So sure, provide a good foundation for an actor based system.  But don't *require* it - don't make it the only way.
2022-12-10 11:08:03 <KipIngram   > I complained the other day about people that accuse Forth of being "just an assembler."  But that doesn't mean that I think Forth shouldn't give you "assembly level access."
2022-12-10 11:08:19 <KipIngram   > It *should* be an assembler - when an assembler is what you want.
2022-12-10 11:11:48 <KipIngram   > So yeah, I think Forth certainly should LET you set up multiple threads with access to shared memory.  Maybe it gives you other, less risky tools as well.  Same as with blocks vs. file systems - if your Forth gives you a file system, that's fine.  But it surely should LET you operate at the block level if you want to.
2022-12-10 11:11:58 <KipIngram   > BLOCK shouldn't just "not be there."
2022-12-10 11:13:44 <KipIngram   > Well, BLOCK-RD AND BLOCK-WR shouldn't "not be there."  I don't think I'd condemn a Forth that had those but didn't offer BLOCK; you could make that yourself if you had the RD and WR words.
2022-12-10 11:13:58 <KipIngram   > BLOCK is actually a frill.
2022-12-10 11:22:36 <lagash      > https://syndicate-lang.org <---- pretty cool programming language that uses Actors as well!
2022-12-10 16:42:08 <KipIngram   > I've been watching the Highlander TV series the last few weeks.  I'm at the very end of season 4 and I think it's beginning to run out of steam for me.  Same thing happened with The X-Files a few months ago.  So many of these shows manage to pull off some slick stuff in the early seasons, but then "decline."
2022-12-10 16:42:54 <KipIngram   > They made six seasons of Highlander, but my understanding is that season 6 is more about trying out various female immortal characters to see who the audience liked, because they were planning a spin-off.
2022-12-10 16:43:51 <KipIngram   > And somewhere in season five they introduce a "demon" characters, which is completely  out of kilter with the mythology of the show up to then.
2022-12-10 16:49:52 <KipIngram   > Through four seasons there really wasn't a single hint of anything supernatural beyond these certain folks living forever unless they get decapitated.  Then suddenly, here comes a demon?
2022-12-10 16:51:44 <KipIngram   > Maybe I'll go back and finish re-watching Babylon 5.
2022-12-10 16:51:46 <thrig       > there's only so much amazing stuff to be had from any one idea
2022-12-10 16:52:19 <KipIngram   > Yes, I actually think they did a pretty good job of turning the crank on their beheading theme.  It's certainly a "novel" idea, but that doesn't mean it's a "rich" one.
2022-12-10 16:52:33 <KipIngram   > The original move from the mid 1980's is one of my favorite flicks.
2022-12-10 16:52:56 <KipIngram   > It's not really a "good" movie - it's pretty thoroughly B-grade, but there's just something about it that appeals to me.
2022-12-10 16:52:57 <thrig       > I first saw that subtitled in thai and it was all dark and I had no idea what was going on
2022-12-10 16:53:15 <KipIngram   > Director's cut is better than the original.
2022-12-10 16:53:29 <KipIngram   > I don't think there are "many" differences, but they're good ones.
2022-12-10 16:53:31 <thrig       > but VHS rentals in Islamabad have that problem
2022-12-10 16:53:46 <KipIngram   > That's true for the original Stargate movie too - even more so.
2022-12-10 16:54:00 <KipIngram   > I thought the theatrical release of that one kind of sucks, but the director's cut is quite good.
2022-12-10 16:54:22 <KipIngram   > And the first couple of seasons of that TV show are good, too.
2022-12-10 16:56:23 <KipIngram   > My understanding is that Christopher Lambert (Connor Macleod in Highlander) was practically blind shooting that movie, since he wasn't wearing his STRONG corrective lenses.  Most of what he was acting around he couldn't even see.
2022-12-10 16:57:13 <thrig       > not healthy to be around someone wildly swinging a bit of metal
2022-12-10 16:57:26 <KipIngram   > No kidding.
2022-12-10 16:57:31 <KipIngram   > Bad combo.
2022-12-10 16:57:54 <thrig       > even if it's a dull aluminum replica
2022-12-10 16:58:28 <KipIngram   > Speaking of swinging bits of metal, obviously there's a lot of that in the Highlander series.  I was wondering a day or two ago if it's anywhere close, at all, to "real" sword use, or if it's more like choreographed dancing.
2022-12-10 16:58:40 <KipIngram   > I have a feeling that "looks good on camera" isn't the same t hing as "winning a fight."
2022-12-10 16:59:53 <KipIngram   > My wife worked with a lady for a while who's son was one of THE "unnamed sword fighters" on "Master and Commander" with RusselL Crowe, and apparently he had to do a lot of training to get ready for that.
2022-12-10 17:00:09 <KipIngram   > Sorry - did not mean to capitalize "the" up there.
2022-12-10 17:00:23 <KipIngram   > No clue how that even happened.
2022-12-10 17:07:36 <eris[m]     > computers are fickle
2022-12-10 20:23:27 <KipIngram   > Interesting old Reddit thread on implementing Lisp in Forth here:
2022-12-10 20:23:29 <KipIngram   > https://www.reddit.com/r/Forth/comments/5wsbgr/forth_is_just_fast_lisp/
2022-12-10 20:23:38 <KipIngram   > And some other aspects (speed comparisons, etc.)
2022-12-10 20:23:49 <KipIngram   > Some arguing, which is sort of to be expected I guess.
2022-12-10 20:24:01 <KipIngram   > Forth and Lisp both seem to... "motivate" people.
2022-12-10 20:32:13 <decay       > largely because they're minimalist.
2022-12-10 20:32:23 <decay       > I should say deceptively minimalist.
2022-12-10 20:50:52 <KipIngram   > I've always been into Forth because of ease of implementation, and to some extent I think Lisp might share that.  I really should pay it more attention.
2022-12-10 20:52:29 <thrig       > Practical Common Lisp might be worth a read
2022-12-10 20:57:53 <MrMobius    > did Lisp even make it onto embedded devices like Forth?
2022-12-10 20:59:46 <thrig       > http://www.ulisp.com/ and others come to mind
2022-12-10 21:03:34 <MrMobius    > interesting
2022-12-10 21:04:51 <thrig       > and https://justine.lol/sectorlisp2/ is pretty small
2022-12-10 21:06:58 <decay       > the dichotomy between forth and lisp is the foundational model.
2022-12-10 21:07:13 <decay       > forth assumes a machine model. lisp fits comfortably in the abstract expression rewriting space.
2022-12-10 21:08:27 <KipIngram   > Wow - there are lots of nice little nuggets in that link.
2022-12-10 21:09:37 <KipIngram   > Among other things, there's a link to this, which I haven't yet digested, but it looks like it might be interesting:
2022-12-10 21:09:39 <KipIngram   > https://web.archive.org/web/20120211191732/http://archive.vector.org.uk/art10500710
2022-12-10 21:10:39 <KipIngram   > That talks about combining attributes of Forth and APL.
2022-12-10 21:15:17 <KipIngram   > It has an example where the code takes to integers off of the stack and pushes back a 100 element array, so it's starting to get into that kind of stuff I've talked about implementing with tags in the high bits.
2022-12-10 21:41:58 <KipIngram   > vms14 might want to check out this 'lang5'.  It's implemented in Perl, so it's kind of in the neighborhood he's been poking around in.
2022-12-10 21:50:33 <KipIngram   > Unfortunately I can't seem to find anything about the language's architecture / implementation.
2022-12-10 21:50:53 <KipIngram   > I'd be interested in seeing how they keep up with the stack item types.
2022-12-10 21:51:26 <KipIngram   > Might have to study the source, but the idea of wading through a bunch of Perl isn't fun.