2022-12-15 06:17:42 <veltas      > I note that in C++ etc people tend to just rename or type out in full namespaces rather than using 'using' etc
2022-12-15 06:17:46 <KipIngram   > In a lot of Forth systems you can SEE <word>; SEE "decompiles" the word.
2022-12-15 06:18:02 <veltas      > Hivemind KipIngram and I just access IRC at the same time
2022-12-15 06:18:26 <veltas      > SEE decompiles and LOCATE gets the source if available
2022-12-15 06:18:27 <KipIngram   > :-)
2022-12-15 06:18:56 <veltas      > And then there's ' <word> 64 DUMP
2022-12-15 06:19:13 <veltas      > Which I often find necessary especially with data etc
2022-12-15 06:22:47 <KipIngram   > I've never actually included header fields to support LOCATE in my systems, but I really should.  There are a couple of things I should probably add to my headers, but I may do it by putting in a mechanism to have arbitrary optional content in headers.
2022-12-15 06:22:59 <KipIngram   > Maybe a JSON string or the like.
2022-12-15 06:23:35 <KipIngram   > For that eventual support for typed stuff I could then include the word's stack effect description using that.
2022-12-15 06:24:55 <KipIngram   > That would be kind of fatty, though.
2022-12-15 06:28:34 <KipIngram   > As it's set up now the PFA pointer is optional, and isn't included in primitive headers.  That's not done in any generic way, though.
2022-12-15 06:29:06 <KipIngram   > It's currently the only optional thing, and if it's there it preceeds the CFA field.
2022-12-15 06:30:40 <KipIngram   > I used to have a more or less FIG traditional header, with the name first.  These days I put the name last, since it's the only variable size part of the header.  Avoids me ever needed to traverse the name field.
2022-12-15 06:30:45 <KipIngram   > needing
2022-12-15 06:32:59 <KipIngram   > Right now it's [PFA] CFA LINK NAME, where name is a counted string.  PFA and CFA are 32 bits, LINK is 16 bits.
2022-12-15 06:36:24 <KipIngram   > I used to have the immediate flag bit in the count byte, but these days I've moved it into the link word, so that the name is a fully standard counted string instead of a funky one that has to be masked.
2022-12-15 06:38:20 <KipIngram   > One of you guys actually suggested that layout change a year or two ago.
2022-12-15 06:42:34 <KipIngram   > It occurred to me in the last day or two that the way we define new words is actually not as "Forth-ish" as it might be.  Something like this would be more Forth-like:
2022-12-15 06:43:04 <KipIngram   > " definition" " name" :
2022-12-15 06:43:38 <KipIngram   > Not that I find that preferable; the usual way is a strong habit now.
2022-12-15 07:37:42 <KipIngram   > vms14: have you done any performance measurements on your system?
2022-12-15 08:18:20 <veltas      > KipIngram: LOCATE in swiftforth and gforth is really nice
2022-12-15 08:19:15 <veltas      > They have motivation to have a good LOCATE because SEE just gives you a code disassembly on many words now
2022-12-15 08:19:30 <veltas      > They're all optimised STC forths so SEE is not that nice to use anymore
2022-12-15 08:20:09 <KipIngram   > Yeah, SEE has limitations.
2022-12-15 08:20:32 <veltas      > It works best with ITC, DTC, and tokenised forths
2022-12-15 08:20:33 <KipIngram   > LOCATE makes a lot of sense, and supports nice debugging, since you can then open to an error point automatically if you want to.
2022-12-15 08:21:32 <veltas      > I think the job of any forth is amongst supporting a good programming library, also providing comfort features to the interactive interpreter
2022-12-15 08:21:52 <veltas      > Because Forth is often claimed to be a full 'operating system', so the interpreter is our shell
2022-12-15 08:22:08 <veltas      > And you can use swift-forth and gforth like a shell
2022-12-15 08:22:23 <veltas      > They should have built-in editors too, and unfortunately they tend to just call vi et al today
2022-12-15 08:22:35 <veltas      > But your forth has an editor right?
2022-12-15 08:23:37 <veltas      > In the Computer Chronicles episode that Elizabeth Rather was on she used the starting forth editor in the demo, and it looked quite nice
2022-12-15 08:24:01 <veltas      > She also seemed really proficient with it, almost as if she had been working in it for years :P
2022-12-15 08:25:26 <veltas      > https://youtu.be/D5osk9lrGNg?t=768 if anyone is interested
2022-12-15 09:23:31 <KipIngram   > I haven't put the editor into this particular implementation yet, but yeaH, I write internal editors.  The last imp had one 4kB block of source that implemented the editor, and it was a fairly nice screen editor that would show me any two blocks side by side, let me navigate around, do basic editing, move lines from one of the blocks to the other, and so on.
2022-12-15 09:23:53 <KipIngram   > I was pretty happy with it, but I've been torn between just re-using that vs. making something a little more powerful.
2022-12-15 09:24:04 <KipIngram   > The kind of "torn" that keeps you from getting anything done.  :-(
2022-12-15 09:24:49 <KipIngram   > It used vim command names, basically, but was a little closer to emacs in terms of how the control key was used.
2022-12-15 09:24:53 <KipIngram   > Sort of a blend of the two.
2022-12-15 09:25:28 <KipIngram   > It only worked with fixed-length 64-char lines, and was thoroughly a "block" editor.
2022-12-15 09:25:53 <KipIngram   > Part of the "torn" thing is a temptation to support variable length lines and make it ready to extend easily to doing files rather than blocks.
2022-12-15 09:26:14 <KipIngram   > Which would mean being able to edit a contiguous series of blocks as one entity.
2022-12-15 09:26:25 <KipIngram   > Then later the file system would extend that to arbitrary sequences of blocks.
2022-12-15 09:27:00 <KipIngram   > Also, since writing that before I've added support for UTF8 to my console interface, so that would be a complicating factor as well.
2022-12-15 09:27:52 <KipIngram   > I have done some work on EXPECT, though, that nicely facilitates the variable line length thing.
2022-12-15 09:28:06 <KipIngram   > I just need to do all the "pulling together" of the various bits.
2022-12-15 09:39:48 <veltas      > KipIngram definitely have a look at that video if you've not seen it
2022-12-15 09:39:54 <veltas      > The forth environment looks so professional lol
2022-12-15 09:39:58 <veltas      > Filling the screen and everything
2022-12-15 15:56:28 <KipIngram   > Was it intentional for that link to start in the middle of the video?
2022-12-15 15:59:09 <KipIngram   > This is the first time I've ever actually *seen* Liz Rather.
2022-12-15 15:59:26 <KipIngram   > She has a southern accent - I wonder where she's from?
2022-12-15 15:59:38 <KipIngram   > I'd guess somewhere in the Southeast.
2022-12-15 16:00:39 <KipIngram   > Her Wikipedia article doesn't say.
2022-12-15 17:34:20 <KipIngram   > Interesting article on page 24 here on Forth in a CPLD:
2022-12-15 17:34:22 <KipIngram   > http://www.forth.org/fd/FD-V21N1,2.pdf
2022-12-15 17:34:31 <KipIngram   > Couple of other interesting articles there too.
2022-12-15 17:34:46 <KipIngram   > One on Forth as a HDL.
2022-12-15 17:37:10 <KipIngram   > I would disagree with the first sentence of that CPLD article, though - "Because Forth's performance isn't compromised by a limited number of registers..."
2022-12-15 17:37:17 <KipIngram   > Well, YES, it is.
2022-12-15 17:37:39 <KipIngram   > It's absolutely clear to me that you can get better performance if you have a good number of registers to deploy for special purposes.
2022-12-15 17:37:54 <KipIngram   > I've discovered that by first-hand experience.
2022-12-15 17:38:46 <KipIngram   > Maybe they mean that in a special way - maybe they mean overall average appication performance after you've gone through and assembly optimized all the critical bits.
2022-12-15 17:38:51 <KipIngram   > That might be true.
2022-12-15 17:45:10 <KipIngram   > But anyway, just the general achievement of designing a Forth processor into a CPLD is nifty.
2022-12-15 17:55:36 <KipIngram   > You know, web searches just aren't much good when you search for something like "low power CPLD" and get hits on products from 2003 leading the results list.
2022-12-15 17:55:56 <KipIngram   > All of the new shiny AI algorithms ought to know that I'm not interested in 20 year old products.
2022-12-15 17:58:32 <thrig       > they ought to know what product to steer a consumer towards
2022-12-15 18:03:41 <KipIngram   > Hey - THIS is a nifty looking little gadget:
2022-12-15 18:03:43 <KipIngram   > file:///home/kipingram/Downloads/FPGA-DS-02008-1-9-iCE40-UltraPlus-Family-Data-Sheet.pdf
2022-12-15 18:03:47 <KipIngram   > Ooops.
2022-12-15 18:03:50 <KipIngram   > Stand by.
2022-12-15 18:04:13 <KipIngram   > Try this instead:
2022-12-15 18:04:16 <KipIngram   > https://www.latticesemi.com/-/media/LatticeSemi/Documents/DataSheets/iCE/FPGA-DS-02008-1-9-iCE40-UltraPlus-Family-Data-Sheet.ashx?document_id=51968
2022-12-15 18:04:43 <KipIngram   > And Digikey has it for about $7.
2022-12-15 18:24:49 <MrMobius    > oh hmm. does that link on forth.org has a lot of issues
2022-12-15 18:29:21 <MrMobius    > *that link on forth.org has a lot of issues
2022-12-15 18:30:38 <veltas      > KipIngram: Yeah was intentional for video to start in middle, that's where the Forth content starts
2022-12-15 18:44:01 <KipIngram   > I figured; thanks.
2022-12-15 23:09:39 <KipIngram   > Linux can be damn frustrating at times.
2022-12-15 23:09:40 <KipIngram   > I'm trying to get the Lattice software for that part I linked earlier installed.
2022-12-15 23:09:50 <KipIngram   > it is installed, and it will let me run through a project setup.
2022-12-15 23:10:13 <KipIngram   > But when i click "Finish" it bombs for QT + Wayland vs. Xorg related reasons.
2022-12-15 23:10:40 <KipIngram   > My best guess is that there's a rat's nest of library versions floating around on my system.
2022-12-15 23:39:02 <KipIngram   > Well, I've exhausted all available Highlander entertainment now.  Finished off The Raven; there were a few pretty poor episodes in there, but there were a couple of quite good ones too, along with a bunch of middlin ones.
2022-12-15 23:39:31 <KipIngram   > Not to hard to see why it wasn't renewed, but it's a bit of shame they didn't get to resolve the season 1 cliffhanger.
2022-12-15 23:40:09 <KipIngram   > I didn't actually watch all of the original series - I left off season 6 and the last episode of season 5.
2022-12-15 23:40:40 <KipIngram   > I know something that happens in that episode and I'd just prefer not to see it happen on-screen.
2022-12-15 23:55:53 <KipIngram   > Anyway, apparently it's possible to select Xorg vs. Wayland at boot time, so I really should reboot under Xorg and see if that works.  But I just don't quite feel like rebooting right now.