00:08 <neuro_sys> "Rather than link each entry to its physical predecessor, link it to a predecessor in one of a number of chains. Scramble the word to determine which chain it belongs in, both when you enter it and when you search for it"

00:08 <neuro_sys> Does Chuck Moore talk about using a hash table here?

00:09 <Zarutian_HTC> assume nothing. How Chuck describes stuff might not have exact translation in usual comp sci terms

00:15 <Zarutian_HTC> many people are too prone to do what I call noospheric snap-to

00:18 <Zarutian_HTC> that is, they see a description of a concept and they think it is symonymus with a closely related but distinct concept they already know

00:18 <neuro_sys> I see what you mean

00:20 <Zarutian_HTC> what Chuck is describing might be something new or at least something that overlapps something well known

00:20 <neuro_sys> "noospheric snap-to" is a nice way to put it

00:22 <Zarutian_HTC> but looking at what he describes can be a source of inspiration of a diffrent view on these things

00:38 <neuro_sys> "It is a futile exercise to attempt to establish a universal set of nouns. Compiler languages have repeatedly stumbled by not providing enough, and no matter how many they provide, someone will want one more."

00:38 <neuro_sys> It makes me think of the latest C++ features and proposals.

13:13 <neuro_sys> I have a begin-while-repeat loop, and I'd like to do short-circuit evaluation in the while condition. I've read about Wil Baden's conditionals, but meanwhile I'm thinking to myself, maybe I should un-learn the short-circuit idea and think differently.

13:18 <cmtptr> put the condition in its own word which returns early to short-circuit?

13:20 <neuro_sys> "returns early", does this involve some return stack manipulation?

14:01 <cmtptr> no, just : condition? first-test if true exit then second-test if true exit then third-test if true exit then false ;

14:01 <cmtptr> (assuming what you want is a logical OR)

14:09 <neuro_sys> Ah I see

14:12 <neuro_sys> I'm doing Advent of Code 2020 to learn some Forth, here's my day 1.0 solution for a naive way to solve the problem of given N numbers, find the pair whose sum is equal to 2020, and multiply them: https://gist.github.com/neuro-sys/3197152e07b1dd95c85ad9245a6588d5

14:13 <neuro_sys> Any comments on it to improve?

14:14 <neuro_sys> I'll try doing it different ways until I get more comfortable

14:15 <cmtptr> i'll let someone more experienced respond. i would say it looks pretty much how most of my forth turns out (except i tend to have less newlines)

14:15 <cmtptr> (also isn't there a cell+ word?)

14:16 <neuro_sys> Ah right cell+ is better

15:00 <cmtptr> what forth are you using, by the way? gforth?

15:01 <neuro_sys> Yes, but I'm trying to keep it minimal for now

15:02 <neuro_sys> For instance, not resorting to local variables

15:02 <cmtptr> i should just install and play with gforth already

15:02 <neuro_sys> It is convenient and the documentation is nice. But I can't compare to any other.

15:08 <neuro_sys> Oh that stack juggling I did was not needed. Now I updated the gist with a simpler version.

15:45 <neuro_sys> Hmm, puzzled by >NUMBER in gforth. 0 S" 123" >NUMBER fails. Seems like it expects 4 cells in the stack.

15:57 <neuro_sys> I wrote this word definition to load a text file containing numbers delimited by new line, and store them in a dictionary entry like this: https://gist.github.com/neuro-sys/3f0cef186e881ea33525b7332d9d3512

15:59 <neuro_sys> I wonder how I can put the last line into a definition.

16:00 <neuro_sys> Ah I got it.

16:02 <hosewiejacke> Forgot the length after S" 123" ?

16:03 <neuro_sys> (I updated the gist for my previous question after I figured it out)

16:03 <neuro_sys> hosewiejacke: Hmm, let me check

16:04 <neuro_sys> S" 123" leaves ( c-addr u ) on the stack, so I'm not sure

16:05 <hosewiejacke> Oh yes, my bad.

16:06 <cmtptr> i think >number wants a base below the string

16:06 <cmtptr> nevermind i misread

16:07 <cmtptr> https://gforth.org/manual/Line-input-and-conversion.html#index-_003enumber--ud1-c_002daddr1-u1-_002d_002d-ud2-c_002daddr2-u2--core

16:07 <cmtptr> seems like it wants an initial value

16:07 <cmtptr> 0 S" 123" >NUMBER

16:08 <cmtptr> i don't know a double means in this context or whether 0 counts, sorry

16:09 <cmtptr> (i.e. does that mean dword, or does it mean two cells, or does it mean double-length floating point?)

16:09 <neuro_sys> Oh it's a double

16:10 <neuro_sys> I missed that. So the first value which is an accumulator should be 2cell

16:11 <neuro_sys> 0. S" 123" >NUMBER \ indeed works

16:12 <hosewiejacke> The Forth Programmer's Handbook says it wants an unsigned, double-precision integer.

22:41 <neuro_sys> Is there a nice way to prevent including the same file more than once? I think I should set a flag and check that.

22:45 <neuro_sys> I re-implemented my solution to use linked list instead: https://gist.github.com/neuro-sys/81628d981b229ae576976ae88538ff0b

23:32 <inode> neuro_sys: use REQUIRE instead of INCLUDE

23:33 <neuro_sys> inode: Thanks

23:34 <inode> check "5.17.1 Forth source files" for more info

23:34 <neuro_sys> Will do