2026-02-08 00:01:28 <KipIngram   > As soon as I'm able I'll use a tight machine code  loop to toggle a pin back and forth and seeiif I can  look really closely at it for the sort of jitter tpbsd was talking about earlier.
2026-02-08 02:18:05 <tpbsd       > KipIngram, awesome
2026-02-08 02:20:46 <tpbsd       > KipIngram, it was a regular cpu stealing interruption, clearly visible in the squarewave stream that stretched the squarewaves (iirc)
2026-02-08 07:23:03 <iv4nshm4k0v > 
ACTION listens to http://opengameart.org/sites/default/files/apple_cider.mp3 

2026-02-08 07:29:05 <tpbsd       > KipIngram, my AI thinks micropython has a garbage collector, that may intefere with your test ?
2026-02-08 07:29:49 <tpbsd       > "MicroPython contains a **mark-and-sweep** garbage collector that is enabled by default on every port."
2026-02-08 09:18:52 <veltas      > Well you can tell the computer that garbage collection doesn't run unless it's needed, so if he's not constantly allocating memory and re-uses lists etc he won't see that
2026-02-08 09:25:13 <tpbsd       > veltas, thats handy
2026-02-08 09:29:57 <veltas      > In my opinion MicroPython should use reference counting as well, or provide a build option for it
2026-02-08 09:30:55 <veltas      > That would allow most allocate/free scenarios to happen immediately without a mark-and-sweep step
2026-02-08 09:31:17 <veltas      > Garbage collection is only really needed when you have e.g. graph-like-tables that self-reference
2026-02-08 09:31:52 <veltas      > Reference counting is done by CPython, Lua, et al
2026-02-08 09:32:16 <veltas      > MicroPython didn't want the overhead, it increases memory usage of every allocated object a bit
2026-02-08 09:32:22 <veltas      > By one word really
2026-02-08 09:33:40 <veltas      > If anything mark-and-sweep should be optional, Python 1 didn't even support it
2026-02-08 10:19:19 <MrMobius    > ive wondered if you could implement a garbage collector and make some guarantees about the worst case behavior
2026-02-08 10:20:01 <MrMobius    > like an RTOS where you're guaranteed to always have 80% of cycles or something and it's guaranteed no code you write would exceed the remaining 20%
2026-02-08 10:20:28 <MrMobius    > so execution is always deterministic
2026-02-08 10:20:55 <MrMobius    > exceed the remaining 20% for doing garbage collection I mean
2026-02-08 10:22:08 <veltas      > I'm guessing not because the 20% wouldn't keep up if you kept allocating memory that was lost?
2026-02-08 10:23:04 <veltas      > It feels easier to allocate what you need up-front, and that is 'safe' regardless of whether you're in C or MicroPython or Java
2026-02-08 10:23:29 <veltas      > Obviously that's more restrictive but RT generally is
2026-02-08 10:25:26 <veltas      > By "kept allocating" I mean continuously in the 80%
2026-02-08 10:25:40 <veltas      > To guarantee you'd have to reduce the computation part to like 0.1% or something maybe
2026-02-08 10:28:41 <tpbsd       > or just use Forth and do everything on the stacks, no GC needed ?
2026-02-08 10:29:36 <MrMobius    > veltas: the point would be if you allocated like crazy it would still keep up because it's always doing maintenance in the background
2026-02-08 10:30:31 <MrMobius    > but maybe that would need way more than 20%
2026-02-08 10:30:46 <MrMobius    > I just wonder sometimes if there's a compromise other than no GC allowed if real time
2026-02-08 12:29:16 <KipIngram   > The standard move would be to try to implement your GC in a way that allowed you to carve it up into short bounded bits of work, that could live WITHIN your real time constraints.  I haven't ever tried, so I don't know how easy it would be, and it would likely be less efficient in terms of overall performance than an approach that just stopped everything and "got it done."  But that's the sort of
2026-02-08 12:29:19 <KipIngram   > tradeoff we face, right?
2026-02-08 12:30:43 <KipIngram   > Back at work over time we added more and more "background maintenance processes" to our SSD operation.  Housekeeping daemons, so to speak.  And it was VERY COMMON for such a thing to be deployed, only for us to discover that it had a visible impact on performance smoothness.  Usually they'd have to go back and change its implementation to "smooth out" the load it placed on things.
2026-02-08 12:31:06 <KipIngram   > I never quite got why they didn't just set that as a goal in the first place, but it seemed like no one ever thought to do that.
2026-02-08 12:31:15 <tpbsd       > hahah
2026-02-08 12:32:09 <KipIngram   > In SSD garbage collection, the longer you wait before you do your GC the more "stale" the write stripes get - there's less data to relocate when you GC a stripe.  But waiting gets you up against it, and you have to get a lot done before you go back to actual work.
2026-02-08 12:32:31 <KipIngram   > On the other hand, you CAN GC gradually, a little at a time continuously, but then you find yourself moving more "still live" data.
2026-02-08 12:32:47 <KipIngram   > So, higher write amplification.  There's no way to have both goals at once.
2026-02-08 12:35:01 <KipIngram   > For Forth RAM management, I've found it just make everything so much easier if all of your blocks are the same size.  Fixed size allocators never require you to move live data.  There are standard approaches that basically have a collection of heaps, each one with different block size.  You just take your block from the one that offers the right size, and then there just isn't any real "GC,"
2026-02-08 12:35:03 <KipIngram   > because nothing ever needs to be moved.
2026-02-08 12:36:26 <KipIngram   > I like making problems go away, as opposed to trying to "solve them."
2026-02-08 12:40:02 <tpbsd       > thats a classic 'thinking Forth strategy
2026-02-08 12:41:19 <KipIngram   > Heh.  It's always appealed.
2026-02-08 12:42:00 <KipIngram   > It's part of why I spend so much time "just thinking" at the outset of a project - trying to find the right approach that as much as possible just keeps difficulties from ever coming up.
2026-02-08 12:43:14 <KipIngram   > Certainly for a system that's going to run on a notebook, where we typically have more RAM than we'll ever know what to do with, it's a nice approach.  Plenty of room for all those heaps.  On a small embedded system more thought might be required.
2026-02-08 13:28:15 <veltas      > Programs these days are like gas, they expand to fill all the RAM they're running with
2026-02-08 13:29:39 <veltas      > It's a bit of a fallacy in my opinion to say that you have more RAM than you know what to do with, because it's a multi-tasking system.  So you might not know right now but later that RAM might be needed for your other apps running at the same time
2026-02-08 13:29:45 <veltas      > But it's your app, I mean do what you want
2026-02-08 13:30:49 <veltas      > I see this argument used to defend e.g. electron use, it's just nonsense, soon I need 16GB of RAM to do what 16MB was sufficient for not long ago
2026-02-08 13:52:52 <KipIngram   > True, but I just really doubt any of us are going to write FORTH that uses it all.  Maybe - who knows?
2026-02-08 13:53:15 <KipIngram   > I typically confine my Forths to a 4GB max zone anyway.
2026-02-08 13:53:52 <KipIngram   > I mean, the stack is 64 bits, so it can hold any address, but a lot of the internal structures use 32-bit fields, that are offsets from a "system base address" that's stored in some register.
2026-02-08 13:54:17 <KipIngram   > @ and ! and so on use 64-bit addresses.
2026-02-08 13:54:32 <KipIngram   > So I CAN get at any of my RAM, but I can't "put Forth code" everywhere.
2026-02-08 13:55:03 <KipIngram   > I flashed MicroPython onto the ESP32-C6; that works.  I made one of the two onboard LEDs turn on and off.
2026-02-08 13:56:01 <KipIngram   > It comes up on /dev/ttyACM0, not /dev/ttyUSB0.
2026-02-08 13:58:01 <veltas      > I'm thinking of when Forth resides next to other apps on an OS
2026-02-08 15:30:30 <KipIngram   > Well, sure - I get that.  But practically speaking I rarely ever push my computer that far.  I've got all my windows and all my browser tabs and everything I normally do - FreeCAD is open over there in a workspace, etc.   According to htop I'm using 15.26 GB of 31.16 GB.  So not even half.
2026-02-08 15:31:00 <KipIngram   > I'm sure I COULD find a way to bring the thing to its knees, but that would just be a terribly rare use case for me.
2026-02-08 15:32:05 <KipIngram   > So, given my "4GB zone" tendency in my Forths, I could run that and still be just 19.2 GB used, and then the question would only be "would I ever really fill up that 4GB?"
2026-02-08 15:36:30 <veltas      > Bear in mind though the reason you have 15GB used could be real, or could be death by 1000 "RAM is cheap" apps
2026-02-08 15:36:34 <KipIngram   > Looks like right now I've got 1) a console window with six tiles in it, 2) Chrome (Brave, actually) with 24 open tabs, 3) a workspace with KeePassX and BluMan open in it, 4) a FreeCAD workspace, and 5) a video in progress that's up on my network attached storage.
2026-02-08 15:36:39 <veltas      > Funny actually because right now RAM isn't cheap at all
2026-02-08 15:37:43 <veltas      > You've got a lot more running than that because of all the shared libraries and servers that are keeping your desktop etc running
2026-02-08 15:37:59 <KipIngram   > Oh, yes, I understand that.  And RAM shouldn't be "squandered" - I think it often is.  But here we're talking about a real REASON for structuring things slightly differently.  Yes, numerous heaps supporting different block sizes isn't going to use RAM AS effectively as one big GC'ed heap, but on the other hand you're getting a HUGE payoff for that: effectively no GC overhead at all.
2026-02-08 15:38:15 <KipIngram   > It's not really "waste" - it's a rational transaction.
2026-02-08 15:38:41 <veltas      > All ready to be rewritten in Rust, watch the memory usage explode
2026-02-08 15:38:53 <KipIngram   > Or you can choose to bear the cost of the GC - it is wholly up to you of course.
2026-02-08 15:39:42 <KipIngram   > Heh - not me; not on my system.  Forth, C, and Python are all t he tools I need.  Well, and an assembler, and occasionally Octave if I'm doing something numerical, but those are both pretty uncommon for me.
2026-02-08 15:40:06 <KipIngram   > I haven't paid much attention to Rust, just like I didn't pay much attention to Java.
2026-02-08 15:40:36 <KipIngram   > But I guess you mean that the people supplying those tools might switch to Rust?  Yeah, that would be a concern.
2026-02-08 15:40:40 <veltas      > People are all enlightened with cheap RAM
2026-02-08 15:40:48 <veltas      > 'enlightened'
2026-02-08 15:40:51 <KipIngram   > It seems to be "the Java for our times."
2026-02-08 15:41:04 <KipIngram   > I also never got the obsession with Java.
2026-02-08 15:41:21 <veltas      > They all think they're so smart writing 'arena allocators' for their programs to work around deficiencies in their language's safety model
2026-02-08 15:41:35 <KipIngram   > Generally speaking I think a lot of people are way too portability-minded.  Or, rather, "portability without working for it."  Everyone wants a magic bullet.
2026-02-08 15:42:01 <veltas      > I say 'arena allocator' because I could swear arenas used to refer to the separate per-thread allocation bookkeeping, not what people call it today
2026-02-08 15:42:16 <veltas      > What people call 'arena allocator' today is basically a stack allocator but worse
2026-02-08 15:43:17 <KipIngram   > Yeah.  I agree.  My usual way of doing the multiple-fixed-size allocator thing is to do... a fixed size allocator.  Two layers - one that hands out large blocks, each one intended as the source for smaller blocks of some size.
2026-02-08 15:43:26 <veltas      > None of these people understand e.g. how Doug Lea malloc works
2026-02-08 15:44:05 <KipIngram   > Well, I've browsed some of that stuff, but I doubt I could just sit down and explain it to you.
2026-02-08 15:44:19 <KipIngram   > Didn't actually "learn" it - just "perused" it.
2026-02-08 15:44:28 <veltas      > I gave a talk on it at university and have implemented it multiple times for different people
2026-02-08 15:44:37 <KipIngram   > Very cool.
2026-02-08 15:44:41 <veltas      > I implemented it for a larger baremetal project at work
2026-02-08 15:45:02 <veltas      > I've got a simplified version of dlmalloc I tend to use when I need to implement a heap
2026-02-08 15:45:23 <veltas      > I do also use stack allocators, allocation groups, block allocators etc.... when they make sense
2026-02-08 15:45:41 <amby        > arguably the only silver bullet for memory is virtual memory, because now you don't need to worry about fragmentation
2026-02-08 15:45:43 <veltas      > But today it feels like everyone is a sheep and the shepard is Rust's pigeonholing design
2026-02-08 15:46:00 <KipIngram   > Ugh.
2026-02-08 15:46:10 <amby        > if there are two 4k physical blocks free anywhere, i can have 8k contiguous virtual
2026-02-08 15:46:11 <KipIngram   > People do seem very averse to actually learning.
2026-02-08 15:46:30 <KipIngram   > And our culture seems to have a "denigration instinct" for people who do put a high priority on it.
2026-02-08 15:46:43 <KipIngram   > I don't know if I'd call it out and out "anti-intellectualism," but it's not too far from it.
2026-02-08 15:46:53 <veltas      > amby: Managing virtual memory space is a trial, it's just a damn sight easier than managing physical memory without virtual addressing
2026-02-08 15:47:54 <veltas      > 64-bit makes it a bit easier I'll admit
2026-02-08 15:48:01 <KipIngram   > It's never made sense to me - people have NO PROBLEM with, say, elite athletes, or elite musicians or artists.  In fact, they idolize them to some degree.  But let it be elite INTELLECT, and suddenly that gets an entirely different response.  I don't see how it's any different - it's just another way of being excellent at something.
2026-02-08 15:49:37 <veltas      > I tend to gravitate towards problems and areas where the benefit of working in a group is greatly diminished, which tends to be stuff requiring intelligence
2026-02-08 15:50:23 <veltas      > I think there's some research saying e.g. 5 maths students do a better job all on their own than they do working 'together'
2026-02-08 15:50:29 <KipIngram   > Me too - I much prefer doing things I can accomplish on my own.  I don't mind leading the work of others.  I guess it's more about me liking to be in control of things than about me "not liking to work with others."
2026-02-08 15:50:44 <KipIngram   > I don't like putting up with decisions I regard as poor.
2026-02-08 15:51:44 <KipIngram   > In those team situations, I did do a fair bit of hands on work, but I would always carve off a piece of it that I could then execute on my own.
2026-02-08 15:52:52 <KipIngram   > Like on the PnP / automated programming stuff back around 2000 or so, I did all the motion control and machine vision parts.  Those involved coordinate transformations, and I was likely the best math person in the group, so it made sense to do that part.
2026-02-08 15:53:09 <veltas      > Fragmentation is definitely still a thing with virtual memory but I think it's been demonstrated (and there's papers on this) that typical usage of allocators doesn't lead to a lot of fragmentation
2026-02-08 16:04:35 <veltas      > If you look at the manpage for malloc for glibc you'll see discussion of arenas for multithread performance, except they're allocated automatically in response to too much contention.
2026-02-08 16:05:14 <veltas      > I don't know how current that is, i.e. maybe they allocate them as soon as you create new pthreads now?
2026-02-08 17:06:11 <KipIngram   > So this ESP32 -  you use esptool to flash it, which (if you flash MicroPython) gets you a serial accessible REPL.  Then you can use mpremote to access its file system - to get it to run something automatically on boot you just stick that in a file main.py and copy it over into that file system.  So that's really about as easy as you could want something to be.
2026-02-08 17:07:41 <KipIngram   > tpbsd: I think I will look into this square wave jitter thing using Python first, since that's slam-dunk easy to do.  I can't help wondering if the person you encountered that talked about the jitter just didn't quite do it as well as he could have.  I'm willing to delve  into it a little and see if I can nail down the "best way" and then do some measurements to see how good that is.
2026-02-08 17:09:10 <KipIngram   > Obviously no square wave is 100% perfect - do you have any thoughts on how good you'd want it to be to consider it "perfectly adequate"?
2026-02-08 17:19:09 <tpbsd       > KipIngram, I think jitter from a mcu in a squarewave is fine and normal on anything but a XMOS mcu, but ine shouldnt see repeditive large extensions to the 'square' wave at longer intervals than typical jitter ?
2026-02-08 17:20:01 <tpbsd       > iirc thats what the DSO screenpic showed, but it was shared online and I didnt generate that one
2026-02-08 17:22:49 <tpbsd       > KipIngram, certainly no mcu squarewave is perfect, they all have a fair bit of jitter as the wave is never exactly 50 %. I have test gear that does generate exactly 50% duty cycle square waves that are very stable, ie signal generators
2026-02-08 17:23:23 <tpbsd       > or just sens it into a flipflop to square it up
2026-02-08 17:23:33 <tpbsd       > send
2026-02-08 17:25:49 <rendar      > veltas: what about `2 3 eq? if. A B else C D end`   and for negation:   `2 3 eq? if! A B else C D end`  ?
2026-02-08 17:26:09 <rendar      > the '.' clearly shows you that the if is not referring to A, but to previous eq?
2026-02-08 17:54:57 <veltas      > I don't think the . will achieve that
2026-02-08 17:55:16 <veltas      > I think if/else/end instead of if/else/then works well
2026-02-08 18:11:56 <vms14       > if 2 3 equal then A B else C D end
2026-02-08 18:12:06 <vms14       > that's mine :D
2026-02-08 18:12:49 <vms14       > if you want the flag from the stack then an empty if then block
2026-02-08 18:13:00 <vms14       > 2 3 equal if then a b else c d end
2026-02-08 18:22:59 <KipIngram   > rendar: If I were writing that, I'd go for 2 3 =? if A B else C D end, and 2 3 !=? if A B else C D end.
2026-02-08 18:23:20 <KipIngram   > Though I'd probably not include the ? chars.
2026-02-08 18:23:40 <KipIngram   > I'd just have < <= = != >= >
2026-02-08 18:24:02 <KipIngram   > And possibly u-versions of each one, so u< u<= etc.
2026-02-08 18:24:49 <KipIngram   > Then you have to decide if you want "implicit 0 argument versions," which would just be 0= 0< etc. and would function identically to x 0 = etc.
2026-02-08 18:25:45 <veltas      > Got an SSD ordered for my NetBSD experiment on my desktop
2026-02-08 18:25:57 <veltas      > Damn stuff's so expensive now, I finally understand what it's like to be Canadian
2026-02-08 18:28:36 <veltas      > The Times 33 with Addition hash AKA DJBX32A
2026-02-08 18:29:10 <veltas      > DJBX33A I mean
2026-02-08 18:30:47 <veltas      > I've chosen as a sort of basic "don't care" string hash, not secure or clever but 'good enough' for lots of hashing and fast-ish to calculate on an 8-bit machine as well as a 64-bit machine
2026-02-08 18:32:09 <tpbsd       > around 2004 I purchased about six embedded netBSD devices from Ebay. They used roms for the image and were self contained, and pretty decent, nicely made for the Australian 2000 Olympic games as ticket related entry box terminals. They consisted of a slim device and a large PSU (containing a transformer). Ive since thrown them out but kept the psu's to power projects
2026-02-08 18:33:18 <vms14       > :0
2026-02-08 18:33:19 <tpbsd       > I should ahve kept at least one for review every 10 years, but I have to be brutal with old gear
2026-02-08 18:33:40 <tpbsd       > as I have so much of it
2026-02-08 18:34:26 <vms14       > 
ACTION imagines going to tpbsd home to see all that old tech and embedded devices

2026-02-08 18:35:01 <vms14       > like a pool of hardware where i can jump
2026-02-08 18:35:31 <tpbsd       > vms14, I live in a laege industrial shed not a home. I'm the kind of guy who always preferred a milling machine over a lounge room with a tv
2026-02-08 18:35:59 <vms14       > more space for hardware :D
2026-02-08 18:36:12 <tpbsd       > yes, exactly
2026-02-08 18:36:32 <tpbsd       > plus motorbikes and vehicles
2026-02-08 18:36:55 <tpbsd       > all in the one large shed, 20 metres by 15 metres
2026-02-08 18:37:16 <tpbsd       > and about 10 metres high
2026-02-08 18:37:35 <vms14       > i have a weird fascination for embedded devices for some reason
2026-02-08 18:37:46 <veltas      > This is why I was wondering how you keep cool in aussie summer
2026-02-08 18:37:48 <tpbsd       > with a mezzanine and enclosed office and bathroom
2026-02-08 18:38:05 <tpbsd       > cold showers when it's over 35 C
2026-02-08 18:38:23 <tpbsd       > I dont have aircon, just a couple of pedestal fans
2026-02-08 18:38:28 <veltas      > Ground water stays under around 20?
2026-02-08 18:38:43 <veltas      > Fans are good if you're hydrated and can perspirate
2026-02-08 18:39:22 <tpbsd       > no, the water is warm in the middle of the day. I have to run the shower for about 30 seconds before I get in or it would scald me
2026-02-08 18:39:40 <veltas      > Wow
2026-02-08 18:39:55 <tpbsd       > only cold water around here in summer is in my fridge
2026-02-08 18:40:11 <veltas      > What do you do with the office?  Is that your bedroom?
2026-02-08 18:40:16 <tpbsd       > yeah
2026-02-08 18:40:44 <veltas      > I think I and a lot of people would be interested to see your 'house' in pictures, although I fully understand if you don't want to share that much!
2026-02-08 18:41:02 <tpbsd       > it used to have aircon but one 46C day it died overnight and I woke at about 3am to it blowing hot air. That was around 4 years ago
2026-02-08 18:41:08 <vms14       > i would like to live there directly xd
2026-02-08 18:41:17 <veltas      > I think if I was single at your age I'd like to live in a similar way probably
2026-02-08 18:41:45 <veltas      > I'd probably live in the steeples of a cathedral and swing from rafters etc
2026-02-08 18:41:51 <tpbsd       > imagine a large blue shed, corrugated steel on the outside, steel frame on the inside
2026-02-08 18:42:11 <veltas      > I can imagine it already but a picture is worth 1000 of the tokens you just emitted
2026-02-08 18:42:18 <vms14       > i want to see the hardware
2026-02-08 18:42:25 <veltas      > But it's okay if you don't want to upload pictures, it's personal obviously
2026-02-08 18:42:35 <veltas      > I don't upload pictures of my house either
2026-02-08 18:44:42 <veltas      > On 4chan they do 'battlestation' threads on the technology board, which are good sometimes.  The most interesting uploads are things like what tpbsd is describing, it's rare though
2026-02-08 18:44:59 <veltas      > Mostly people uploading pictures of cozy pastel chicanery and mechanical keyboards
2026-02-08 18:45:22 <vms14       > xD
2026-02-08 18:45:56 <vms14       > mine is just an old monitor working as a table with cardboard folded so a cheap android tablet can stand + usb keyboard
2026-02-08 18:46:00 <veltas      > There is one guy uploading actually who seems to live in an old school hall, I think it's a commune though
2026-02-08 18:46:01 <vms14       > i guess i win :D
2026-02-08 18:46:27 <veltas      > vms14 I've said before and I'll say it again, please tell me where the gofundme is :P
2026-02-08 18:46:43 <tpbsd       > I'm uploading a pic at the back of my shed now
2026-02-08 18:47:10 <vms14       > actually if i had money i might not buy a computer anyways
2026-02-08 18:47:31 <vms14       > the first thing would be a phone with integrated keyboard
2026-02-08 18:47:59 <veltas      > Do they still do those?
2026-02-08 18:47:59 <vms14       > with termux + proot distro i have a fake debian and i miss almost nothing
2026-02-08 18:48:09 <veltas      > Have you tried the android terminal?
2026-02-08 18:48:15 <veltas      > I've heard android's got its own terminal now
2026-02-08 18:48:21 <vms14       > very few companies like unihertz
2026-02-08 18:48:24 <vms14       > no :0
2026-02-08 18:48:37 <veltas      > I don't know how it compares to termux, might be a waste of time
2026-02-08 18:48:46 <veltas      > Probably is a waste of time knowing Google
2026-02-08 18:48:57 <veltas      > without being rude, please hire me google
2026-02-08 18:49:07 <vms14       > i would buy this one
2026-02-08 18:49:16 <vms14       > https://www.unihertz.com/es-es/products/titan-2
2026-02-08 18:49:42 <vms14       > i had the titan pocket and i never have been that happy
2026-02-08 18:49:56 <vms14       > this one is kind of meh, but doubles the specs of the pocket
2026-02-08 18:51:10 <tpbsd       > here is a pic at the rear of my shed, it shows my Starlink install and looks east: https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/starlink-tp-rearshed-sml.jpg/download
2026-02-08 18:53:15 <tpbsd       > and here is one at the front which shows the whole shed and the carpark taken before I fitted the starlink terminal or youd see it on the lafthand top corner: https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/building.jpg/download
2026-02-08 18:53:44 <veltas      > I don't know why but I just pictured an older building than that in my head
2026-02-08 18:53:44 <KipIngram   > vms14: I'd never be happy with that little keyboard.  I'd never do anything with it other than the bare minimum.  In a phone format I'm a devoted Swype fan.  Still doesn't match a real full-size keyboard I can TYPE on, but it's the best I think you can do in that form factor.
2026-02-08 18:53:52 <veltas      > Actually looks like a nice patch of land
2026-02-08 18:54:29 <KipIngram   > Yes it does.
2026-02-08 18:55:33 <veltas      > Very nice
2026-02-08 18:55:37 <tpbsd       > looks can be deveiving, it's industrial and I have a various smelly food factories on one side and a sewerage processing plant on the other, but in general apart from the heat, it's ok
2026-02-08 18:55:53 <veltas      > Nice loading bays
2026-02-08 18:56:22 <vms14       > KipIngram: yeah i was really missing a standard keyboard with that phone
2026-02-08 18:56:35 <vms14       > but i miss more the ability to program anywhere anytime
2026-02-08 18:56:47 <vms14       > and to have the whole system always running in my hand
2026-02-08 18:57:25 <vms14       > and you just get used to the keyboard, but obviously nothing is better than a normal one
2026-02-08 18:57:33 <tpbsd       > it's a gated premesis, another shed identical to mine and 14 lock up storage units all surrounded by lockable sliding gates and a high fence all topped with barbed wire
2026-02-08 18:58:01 <vms14       > the smell is a real problem
2026-02-08 18:58:07 <vms14       > anything else, meh
2026-02-08 18:58:11 <tpbsd       > veltas, yeah the bays are pretty big, they take a truck with a 20' shipping container no problem
2026-02-08 18:58:49 <tpbsd       > vms14, I only get any smell very occasionaly bepending on the wind and heat
2026-02-08 18:58:56 <tpbsd       > depending
2026-02-08 18:59:38 <tpbsd       > as I say, it's a industrial area, not residential. Its dead quiet at night, no neighbours to yell etc
2026-02-08 18:59:54 <vms14       > that's the best for programming
2026-02-08 19:00:22 <vms14       > you and your own thoughts, what else do you need
2026-02-08 19:01:14 <veltas      > Sounds good when you say that
2026-02-08 19:01:32 <tpbsd       > nothing really, it's the epitomy of seclusion
2026-02-08 19:04:17 <tpbsd       > and finally, Ive reserved the best pic for last, if you guys thought the previous pics were nice ...
2026-02-08 19:04:32 <tpbsd       > https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/terrys-shed.jpg/download
2026-02-08 19:04:43 <tpbsd       > take a gander at that one!
2026-02-08 19:06:31 <veltas      > wow
2026-02-08 19:06:53 <tpbsd       > the background is the caldera of an ancient volcano
2026-02-08 19:06:55 <veltas      > Big farm equipment seller in foreground?
2026-02-08 19:07:33 <tpbsd       > it was a annual farm machinery show/sales venue
2026-02-08 19:08:05 <tpbsd       > my shed backs onto that large area on the other side of the security fences
2026-02-08 19:08:44 <veltas      > Right
2026-02-08 19:08:53 <tpbsd       > Ive flown a helicopter all around the caldera including up to the 'Gold Coast' about 100 km from here
2026-02-08 19:09:14 <veltas      > Did you take that pic from a heli?
2026-02-08 19:09:38 <tpbsd       > no, I didnt take the pic, that was the show vendor using a drone
2026-02-08 19:10:27 <tpbsd       > it was a tiny Robinson R22 I used to fly
2026-02-08 19:16:18 <veltas      > No interior shots?
2026-02-08 19:18:03 <veltas      > Flying a heli sounds cool
2026-02-08 19:26:14 <tpbsd       > I'm just uploading a interior of my workshop, of course it only shows a tiny part
2026-02-08 19:26:38 <tpbsd       > veltas, it's the most exciting thing Ive ever done, flying a R22
2026-02-08 19:26:53 <tpbsd       > I recommend you do the same before youre too old
2026-02-08 19:27:11 <tpbsd       > and Ive done a LOT of exciting things
2026-02-08 19:28:49 <tpbsd       > a shot of part of my 'electronics area'(its about twice that size, while Im building outdoor long range wifi P-P links for a remote island location: https://sourceforge.net/projects/mecrisp-stellaris-folkdoc/files/workshop.jpg/download
2026-02-08 19:31:59 <vms14       > that's cool
2026-02-08 19:33:05 <tpbsd       > it's a old pic, it's very cluttered in here atm
2026-02-08 19:33:34 <vms14       > more wires makes it more cool
2026-02-08 19:33:58 <vms14       > wonder the kind of experiments you make there
2026-02-08 19:35:09 <tpbsd       > the island only had solar power and was the birthing area for giant green sea turtles of the northern coast. I made a network of about 5 wifi units that talked to the main base in victoria via a sattelite uplink
2026-02-08 19:35:49 <tpbsd       > they were used to connect to networked cameras
2026-02-08 19:36:34 <tpbsd       > Ive made a lot of these units, all expensively machined here and one of my designs
2026-02-08 19:36:49 <tpbsd       > extensively
2026-02-08 19:37:34 <vms14       > i yet dream of the day i start making a led blink xD
2026-02-08 19:37:39 <tpbsd       > haha
2026-02-08 19:38:00 <tpbsd       > $10 and Mecrisp-Stellaris and youre there!
2026-02-08 19:38:02 <vms14       > i would love to get enough knowledge to steal components from old devices and combine them to see how they explode and laugh
2026-02-08 19:38:26 <tpbsd       > reclaim you mean :)
2026-02-08 19:38:34 <tpbsd       > recycle
2026-02-08 19:39:08 <vms14       > i guess with today's tech is harder to steal components than before
2026-02-08 19:40:13 <tpbsd       > vms14, I once spent 3 weeks in Hong Kong doing a VMS controller course with AAP-reuters at the 'Cable and Wireless' building
2026-02-08 19:40:37 <vms14       > seems you have done a lot of stuff
2026-02-08 19:40:58 <tpbsd       > no, in many ways it's easier to remove SMT components, only hot air rework tools are needed
2026-02-08 19:41:22 <tpbsd       > well Ive had plenty of time to do everything as I'm 71 years old
2026-02-08 19:41:46 <vms14       > was going to ask for age but was thinking it does not really matter
2026-02-08 19:41:55 <vms14       > what matters is what you do with your time
2026-02-08 19:42:03 <tpbsd       > vms14, by the time youre 71, you will have done a lot also!
2026-02-08 19:42:17 <vms14       > wonder if i'll ever arrive, and how i would be if so
2026-02-08 19:42:34 <tpbsd       > it only matters regarding time to do things, experience takes time
2026-02-08 19:42:53 <tpbsd       > heheh, you arrive at old age, a tiny bit at a time
2026-02-08 19:43:10 <tpbsd       > and you don't even notice the journey
2026-02-08 19:43:11 <vms14       > i do not plan to do many things, just keep gathering knowledge about programming and eventually go to hardware
2026-02-08 19:43:34 <vms14       > i do realize programming is my passion though, but i assume hardware will be as exciting
2026-02-08 19:44:07 <tpbsd       > I did many things, I made sure to do them while I was young, because I reasoned that I wouldnt be able to do a lot once I was in my old age
2026-02-08 19:44:40 <tpbsd       > so I travelled a lot, rode across australia on a trail bike when I was 18
2026-02-08 19:45:07 <tpbsd       > all kinds of things, my life has been packed with adventures
2026-02-08 19:45:55 <vms14       > i never had interest in travel, but japan and china get my attention
2026-02-08 19:47:05 <tpbsd       > yeah, china is already the world tech and manufacturing  leader imho. They already do many 'impossible'things
2026-02-08 19:47:19 <vms14       > like what impossible things?
2026-02-08 19:47:37 <tpbsd       > too many to list
2026-02-08 19:47:42 <vms14       > now they seem focused on ai, they've set a goal to be the best country on ai in 2030 i guess
2026-02-08 19:47:52 <tpbsd       > they have their own space station
2026-02-08 19:48:06 <tpbsd       > they will be first with a moonbase imho
2026-02-08 19:49:22 <tpbsd       > they have more warships than the USA, including aircraft carriers with working electromagnetic aircraft launchers
2026-02-08 19:49:51 <vms14       > that makes me wonder about ww3 and the power to destroy the whole planet we already have
2026-02-08 19:49:58 <tpbsd       > I think the rate of technological inovation in China is going exponential
2026-02-08 19:51:56 <tpbsd       > we sure have crazy ability to do a lot of damage, but hopefully sane heads will prevail
2026-02-08 19:52:39 <tpbsd       > people have worried about the end if times since the invention of the crossbow, most of us are still here tho
2026-02-08 19:53:52 <tpbsd       > remember, while we are worrying about a nuclear winter, an asteroid the size of the moon could slam into us at mach 100 and split the earth in two!
2026-02-08 19:55:17 <tpbsd       > there have never been any guarantees, and ww2 lasted 7 years, everyone was exhausted, sadly a lot of the survivors have since died of old age, and the world has forgotten to never do that again!
2026-02-08 19:58:48 <vms14       > they say it's not probable
2026-02-08 19:58:58 <vms14       > but thought in 2032 it might
2026-02-08 20:00:44 <tpbsd       > I think there are too many billionaires who couldnt bear to be seperated from their gold to allow it to happen
2026-02-08 20:01:07 <tpbsd       > imagine having vaults full of gold and no where to spend it ?
2026-02-08 20:01:30 <tpbsd       > just smoking ruins everywhere?
2026-02-08 20:01:46 <tpbsd       > the super rich couldnt beat the thought
2026-02-08 20:01:54 <vms14       > i believe in God and he tells me anything in the earth has any value
2026-02-08 20:02:37 <vms14       > and that the love of money is the root of all evil
2026-02-08 20:02:48 <vms14       > but more importantly that money cannot pay for your soul
2026-02-08 20:04:38 <tpbsd       > it's important to believe in something much bigger than yourself I believe, but lets not get too far of the track here, this is a Forth channel :)
2026-02-08 20:04:53 <vms14       > yeah sorry about that
2026-02-08 20:05:31 <tpbsd       > well Im not a OP, but we havent been kicked off yet :)
2026-02-08 20:07:24 <tpbsd       > so youre a software person ? planning to get into electronics hey ?
2026-02-08 20:07:40 <tpbsd       > Forth is the perfect vehicle for both
2026-02-08 20:08:17 <tpbsd       > Im only a electronics guy, not a programmer, tho I do program because embedded requires it
2026-02-08 20:08:50 <tpbsd       > but Im a pedestrian quality programmer, very uninspired
2026-02-08 20:10:17 <KipIngram   > I'll just make one comment on that - overdoing the money thing is indeed a negative, but on the other hand we've built a world in which it's indispensable - the trick is in the balance.
2026-02-08 20:11:47 <tpbsd       > KipIngram, absolutely
2026-02-08 20:12:00 <veltas      > Money is resources, manage your resources, don't worship the resources
2026-02-08 20:12:52 <tpbsd       > well, I'd expect logic and critical thought like that from Forthers, it's why I frequent here
2026-02-08 20:14:03 <tpbsd       > I think the stats are that only something like 1% of the people in the world have food, a roof over their head, and money in the bank
2026-02-08 20:14:17 <tpbsd       > the rest dont
2026-02-08 20:14:56 <tpbsd       > tho maybe that stat has changed in the last 40 years, especially in China ?
2026-02-08 20:19:12 <veltas      > 1% is probably true depending on how you define "money in the bank"
2026-02-08 20:19:28 <vms14       > tpbsd: about forth the real reason i actually started learning about it was because a guy told me forth embraces the bottom up approach better than any language
2026-02-08 20:19:59 <vms14       > never cared about hardware, but i wanted to become better at bottom up because i thought it would make me a better programmer
2026-02-08 20:20:23 <vms14       > then forth made me realize i could make my own rpn abomination and i got stuck there :D
2026-02-08 20:20:58 <vms14       > now i'm actually mixing both lisp and a stack rpn lang in the same runtime
2026-02-08 20:21:07 <vms14       > 1 2 (put (+ 1 2))
2026-02-08 20:22:11 <vms14       > https://gitlab.com/vms14/oh/-/blob/master/oh.js  that's the code although it's not finished
2026-02-08 20:22:50 <vms14       > never cared about hardware i mean the aspect of forth being used there
2026-02-08 20:23:15 <vms14       > i have a weird fixation for embedded devices and i see electronic components as a woman might see joys
2026-02-08 20:23:41 <vms14       > but yeah, if i ever go there finally forth will make more sense than my rpn abominations
2026-02-08 20:29:41 <tpbsd       > haha
2026-02-08 20:29:51 <tpbsd       > KipIngram, "### ✅ **1. Mecrisp-Stellaris (RISC-V fork)**
2026-02-08 20:29:52 <tpbsd       > - **What**: A native code Forth for ARM and RISC-V MCUs.
2026-02-08 20:29:52 <tpbsd       > - **Status**: There is a **RISC-V port** of Mecrisp-Stellaris, and it **can run on the ESP32-C6**, but it requires some **manual work**:
2026-02-08 20:31:23 <tpbsd       > vms14, I think Forth does both 'bottom up' and 'top down' very well, I know I use it for both
2026-02-08 20:32:06 <vms14       > i wanted to learn about this because i couldn't find documentations or anything just that being mentioned by lispers actually
2026-02-08 20:32:28 <vms14       > but it felt bottom up is the way a program should be made
2026-02-08 20:33:09 <vms14       > i guess i have learned it since i have been making my own random langs since then, and what's more bottom up than start writing your language first
2026-02-08 20:33:39 <vms14       > although my languages are not serious and i cheat as much as i can
2026-02-08 20:34:04 <tpbsd       > ah, well thats all above my paygrade as Im a hardware guy
2026-02-08 20:34:39 <tpbsd       > prdestrian-programmer-is-us
2026-02-08 20:35:09 <vms14       > i want to learn hardware before i die, but i also want to get a good background in programming
2026-02-08 20:35:26 <tpbsd       > well Im off for a few hours to get some more sleep, back later
2026-02-08 20:35:32 <vms14       > see you
2026-02-08 20:38:05 <veltas      > Night
2026-02-08 20:38:20 <veltas      > Thanks for the pics, was interesting, wife was interested too
2026-02-08 20:50:19 <KipIngram   > vms14: Design top down, build bottom up.
2026-02-08 20:50:31 <KipIngram   > That way you can test as you work your way up.
2026-02-08 21:12:17 <vms14       > i usually reach for perl when i have no idea on how to do something xd
2026-02-08 21:12:32 <vms14       > but i tend to start by defining data and operators on that data