IRC Log - 2025-01-15 - ##forth

Channel: ##forth
Total messages: 80
Time range: 00:24:32 - 16:41:20
Most active: veltas (36), pgimeno (31), xentrac (8)
00:24:32 ##forth <xentrac> veltas: Uxn?
00:40:44 ##forth <veltas> That's the one
00:42:23 ##forth * xentrac hits veltas
00:42:50 ##forth <xentrac> I was worried I'd have to hit myself instead because I didn't know it either
00:44:23 ##forth * veltas hides
12:16:57 ##forth <veltas> pgimeno: I got a chance to play with the emulator and Python JAForth stuff yesterday
12:17:11 ##forth <pgimeno> oh yes?
12:17:39 ##forth <veltas> And it held up to expectations basically, from glancing at the code the other day
12:18:19 ##forth <pgimeno> well, note that today I've found out there's a more standard word for line comments, namely \ so I've replaced // with \ everywhere
12:18:25 ##forth <veltas> I think it might be interesting to allow it to be more interactive, so you can do some more forthy development in the terminal, but I understand that's probably not your style
12:18:44 ##forth <veltas> Oh right I don't think I even saw // so that was already replaced(?)
12:18:49 ##forth <veltas> Or maybe I don't remember lol
12:19:14 ##forth <veltas> Also would be good for it to say which word wasn't found, when it fails on a word not being found
12:19:25 ##forth <veltas> or print out where >IN was basically
12:19:32 ##forth <pgimeno> the aim was not to make it interactive, it was more of a programming aid for Jupiter
12:19:41 ##forth <veltas> Yeah I thought as much
12:19:54 ##forth <veltas> But as a happy accident of stdin it is actually interactive :P
12:20:10 ##forth <pgimeno> the emulator version is, the Python version isn't
12:20:18 ##forth <veltas> True
12:20:26 ##forth <veltas> I didn't try giving Python /dev/stdin anyway
12:20:37 ##forth <pgimeno> the emulator version reads the file as it goes; the Python version reads the whole file in one go at the start
12:20:53 ##forth <veltas> Right
12:21:20 ##forth <pgimeno> as for the word not found issue, it shows the remaining buffer after the error
12:21:40 ##forth <veltas> Fair enough
12:22:06 ##forth <pgimeno> I actually had a plan to make the python version read the file as it goes, to allow for interactivity
12:22:29 ##forth <pgimeno> but you don't have any editing capabilities, so it didn't seem very useful
12:23:12 ##forth <veltas> Do you think you'll write anything with jaforth?
12:24:02 ##forth <pgimeno> I developed JAForth because I had to give up on using [ ] LITERAL in Mazogs, so I'll convert it to use JAForth
12:25:02 ##forth <pgimeno> I'm also working right now in a .tap decompiler, and then I think I'm done with the JA software for now
12:26:57 ##forth <veltas> Was mazogs the thing that was converted from BASIC?
12:27:09 ##forth <pgimeno> yes, exactly
12:27:14 ##forth <veltas> Nice
12:27:37 ##forth <veltas> Would you consider doing a demo video of mazogs when you've done it?
12:28:01 ##forth <veltas> Also do you have a JA emulator you recommend?
12:28:41 ##forth <pgimeno> hm, I don't think you'd find any difference between ZX81 Mazogs and JA Mazogs, I've cloned it in all the details
12:29:05 ##forth <pgimeno> as for the JA emulator, I mostly work with ZEsarUX and MAME
12:29:34 ##forth <veltas> Oh I can't run zesarux, burns a hole in my table
12:29:35 ##forth <pgimeno> not sure I'd recommend either, I have not tried Eighty-One though I've heard it works under Wine
12:29:52 ##forth <veltas> And lags like hell
12:30:15 ##forth <pgimeno> it had a bug until recently where memory and CPU consumption grew indefinitely
12:30:28 ##forth <veltas> Maybe that's my job then, to port fuse to the jupiterace :P
12:30:31 ##forth <pgimeno> also, it supports several engines, and SDL is not the default
12:30:46 ##forth <pgimeno> you mean to add support for JA to fuse?
12:31:09 ##forth <veltas> Yes you know what I mean
12:31:20 ##forth <veltas> I knew what I meant anyway
12:32:46 ##forth <xentrac> is it possible to record an asciinema.org video of Mazogs? I have the impression that the ZX81 only had character graphics, so maybe it is?
12:33:14 ##forth <pgimeno> character graphics, but not ASCII
12:33:37 ##forth <pgimeno> they had to add ZX81-specific characters to the Unicode standard
12:34:02 ##forth <pgimeno> that's pretty recent, I doubt there are fonts that support them
12:34:16 ##forth <xentrac> oh really? I thought Noto was pretty quick about adding new characters
12:35:49 ##forth <pgimeno> 0x1FB8E is one of these characters
12:37:09 ##forth <pgimeno> also, Mazogs won't look very well unless the character box aspect ratio is 1:1
12:39:04 ##forth <veltas> The pixel ratio isn't 1:1 on a real ZX81 is it? Sorry being pedantic, it's still nothing like a normal font ratio
12:46:23 ##forth <pgimeno> the characters are 8x8 pixels, and the pixel aspect ratio depends on the TV but is close to 1:1 (and can be made to be 1:1)
12:48:15 ##forth <pgimeno> the Jupiter ACE is very similar, and its screen composition is here: https://codeberg.org/wilco2009/Mercury_Ace/media/branch/main/images/JupiterAce_screen.png
12:49:28 ##forth <pgimeno> depending on how much of the border is visible (CRT TVs used to have a vertical zoom feature) the aspect ratio can be made 1:1
12:50:14 ##forth <pgimeno> on US machines that might be different, idk
14:52:58 ##forth <veltas> I just mean on a 4:3 Trinitron TV my ZX Spectrum lined up nicely with the phosphor dots, no problem, and yet the 'pixels' clearly weren't square
14:53:20 ##forth <veltas> The big picture is that what was a square on an emulator was actually a rectangle in real hardware
14:53:40 ##forth <veltas> And I think this kind of sizing is quite standard for 4:3 CRT's
14:54:39 ##forth <veltas> I'm sure it can be adjusted, but there's nothing 'wrong' with the signal, it's just the way PAL works on a 4:3 CRT, to my best understanding
14:56:15 ##forth <veltas> Basically the pixels are more like 5:4 than 1:1, I don't know exact proportion but can probably be calculated if 4:3 is the goal
15:11:47 ##forth <pgimeno> thing is, CRT TVs vary in how many lines and horizontal size they show; I don't remember seeing oval circles in the Spectrum, for one, and I'm very sensitive to wrong aspect ratios. I would have noticed a 5:4 PAR deformation. I believe the SNES had about that PAR though.
15:37:16 ##forth <xentrac> there's usually a trimpot to change the gain on each of the horizontal deflection and the vertical deflection
15:48:53 ##forth <KipIngram> I imagine most humans CAN detect "out of true circle" fairly sensitively - it probably just "bothers" some of them more than others. I just can't put up with watching a video that's out of aspect ratio, particularly if it has people in it.
15:49:54 ##forth <KipIngram> I wonder if you tested this on a population using a) a circle and b) a square if you'd get an observed difference. I want to think you would - that the circle would produce a higher sensitivity.
15:50:28 ##forth <KipIngram> Or a "flat" square vs. a diamond oriented square.
15:50:41 ##forth <KipIngram> I think in that case the diamond would win.
15:51:57 ##forth <KipIngram> pgimeno: How sensitive are you to slightly out of tune notes in a melody? That's one that I catch really well. I can't sing to save my life, but when I'm just listening I know if it's wrong.
16:02:02 ##forth <veltas> There's a big difference between showing a circle on a TV out of context, and showing it after usign that TV for hours
16:02:24 ##forth <veltas> And you look away and everything looks red-tinted and oblong, your eyes may have adjusted to the TV lol
16:02:55 ##forth <veltas> No that doesn't really happen, but I will say, especially with pixel graphics, your brain is filling in some of the gaps
16:03:17 ##forth <veltas> I found a thread and apparently on an 80's CRT the Spectrum was very close to square, and lots of micros from the era weren't
16:05:37 ##forth <xentrac> cool
16:14:27 ##forth <pgimeno> KipIngram: I couldn't say, but given a tone and another reference tone, I can tell if it's off. For example, a 2 kHz active buzzer I have is slightly off of a reference 2 kHz signal produced by my signal generator.
16:16:00 ##forth <pgimeno> anyway, back to the point about character aspect ratio, the fact is that fonts often have close to a 1:2 aspect ratio, and that's very perceivable. Those who have seen CGA 40x25 and 80x25 column modes can tell.
16:16:58 ##forth <pgimeno> That's why I don't think an ASCII version of Mazogs would look OK unless rendered with a special 1:1 (or close enough) font.
16:28:04 ##forth <veltas> Yeah this is why I said my point was pedantic, even if it was true, the fonts are way off
16:41:20 ##forth <xentrac> I suppose asciinema.org renders with some font it specifies