00:01 <neuro_sys> Amazing, I missed this important fact.

00:02 <neuro_sys> It seems like it eliminates the need for namespacing as I previously thought.

00:03 <veltas> I suppose it does generally make namespacing unnecessary

00:03 <neuro_sys> Also looked into `here`, `,` and `allot`, it seems like no need to use `allocate` either. That way the memory management is simpler yet more restrictive I guess.

00:03 <veltas> Absolutely no need to use allocate

00:04 <veltas> Did you see my forth solutions to day 1? :P

00:05 <veltas> A heap with all its poor determinism and fragmentation is not something you should dirty normal forth programs with

00:05 <veltas> But it's nice to have anyway sometimes

00:07 <veltas> Someone needs to write a modern forth tutorial neuro_sys, you are making all the same revelations I had to make a year ago

00:09 <neuro_sys> The first line in your day1 has [undefined], I think that's a way to execute that line which contains an if word without actually defining a word?

00:13 <veltas> neuro_sys: It's kind of like #ifndef

00:13 <neuro_sys> Ah no, just read what [undefined] again. Yeah.

00:14 <veltas> Me talking about forth early on and someone says "yeah... so locals and forth... er... hahaha generally a red flag." http://bespin.org/~nef/logs/forth/20.02.23

00:15 <veltas> Now I am that guy

00:15 <veltas> That proteusguy

00:15 <neuro_sys> Haha

00:16 <veltas> I could remove those conditional lines because they are just there to support standard forths

00:16 <veltas> I am defining stuff that gforth comes with

00:16 <veltas> basically

00:16 <veltas> I prefer working with stdin rather than opening a named file

00:18 <veltas> I 'get' forth state now but at the time apparently it made my 'head hurt', and that seemed to attract me to the language more for some reason...?

00:18 <neuro_sys> I think it's good use of time to write up a tutorial that specifically addresses the things that should be unlearned first by folk like me. Even though these things are written everywhere, it takes a while to figure out the why.

00:56 <neuro_sys> Just practicing. :) veltas: https://github.com/Veltas/aoc20/compare/master...neuro-sys:patch-1

01:03 <veltas> neuro_sys: If I want to check for end I probably would have rewritten to be a do/+loop and then just put the error message at the end of the loop

01:04 <veltas> I'm sure it's no surprise to you I would rather have shorter code than validate the input, a saner place to start would not be running EVALUATE on every line!

01:05 <veltas> https://archive.org/details/forthdimension

01:07 <veltas> neuro_sys: If you're interested by what I've written check out the way I handle arrays and looping in 1-2.fs

10:50 <neuro_sys> I tried day 1-1 with "classic" style: https://github.com/neuro-sys/advent-of-code-2020/blob/main/day1-1.fs

10:54 <veltas> Cool, I am reading it

10:54 <veltas> You don't need to close a file if you're just reading it, as far as I know

10:58 <veltas> I like : load-file open read close ;

11:00 <veltas> Instead of dup 0= you can do ?dup, but you will need to swap the if/else parts

11:01 <veltas> And then you can remove the drop when it's 0

11:06 <neuro_sys> Ah right forgot the ?dup

11:06 <veltas> Similarly in the loop you do "0<> while ... repeat drop", you could have "?dup while ... repeat" equivalently

11:06 <veltas> Quite often ?dup is useful when comparing against zero or checking a flag

11:08 <veltas> I think so anyway I am struggling to think seriously about your code because I am in the middle of writing merge sort for my big boy version

11:09 <veltas> I really like what you've done here neuro_sys, it's very clean as well

11:10 <veltas> I think I would actually prefer more comments reading it, I think Chuck Moore is wrong on that one, especially stack comments

11:11 <neuro_sys> Thanks. Yeah the two or three words in particular there would benefit a lot from stack comments.

11:11 <veltas> I don't mind no comments mid-definition if it's a short definition

11:13 <veltas> I think it helps the most to have them for the input and output of words, to avoid having to necessarily read the whole program to figure out what's happening

11:14 <veltas> Instead of leaving 0 at the end of the array, have you considered checking when the search address is HERE to end the inner loop?

11:14 <veltas> Oh and there is a forth standard word for "1 cells +": cell+

11:16 <neuro_sys> Nice, just pushed the changes

11:28 <veltas> Looking at your code has given me a good idea for my merge sort code

13:43 <neuro_sys> I added day 1-2: https://github.com/neuro-sys/advent-of-code-2020/blob/main/day1-2.fs

14:06 <veltas> Part 2 is not as easy to factor

14:09 <veltas> But it's worth trying to make that smaller I think

14:39 <veltas> neuro_sys: Is there anything you've made with these classic rewrites to put in your forth utils library?

14:41 <neuro_sys> veltas: That library is useless at this point. I'd have to re-write them all.

14:42 <neuro_sys> Maybe after I do day2, or possibly in day3 they may be needed, and reworked.

14:43 <veltas> Luckily I think that forth is either especially reworkable, or it requires the skill of constant reworking regardless

14:43 <neuro_sys> At least I can change them to use Dictionary instead of allocate, but so far I was imitating C-syntax, so they're pretty much moot at this point.

14:44 <veltas> I didn't notice you using allocate previously