« first day (837 days earlier)      last day (698 days later) » 

00:27
am I to infer that patenting (which discloses the method by which things are accomplished for future generations) is considered morally inferior to selling a black box
I mean either way you're trying to get money for your ideas
I hear certain things are intentionally not patented to keep the "secret sauce". Also, I guess it'd be pretty hard to prove someone used your patented idea in their software
patents expire, but trade secrets don't
01:19
@coltim exactly
 
4 hours later…
04:56
@ngn Using the pre-filled pattern for 2,3 and 5 is slightly slower than the original way.
\t (0<)#&1<{@[@[x;&n#~!p;1];p:*&~x;2]}/:2>!n:200000
1750
\t (0<)#&1<{@[@[x;&n#~!p;1];p:*&~x;2]}/:@[(n:200000)#010000010001010001010001000001b;2,3,5;2]
1810
Those tests were done with `mi 2021.06.15. The pre-filled pattern version doesn't work at all in `mi 2021.06.21`. I'm guessing that's because the `b` suffix does something different in the more recent version:

010000010001010001010001000001b
01:00:00.010000010
That'll teach me not to delete an old version of k before testing a newer one.
Old behavior:

mi 2021.06.15 4 32 (c)shakti 2.0
010000010001010001010001000001b
0 1 0 0 0 0 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 1
05:14
i think all integers are united and converted on demand. there are no more explicit booleans. but i cannot run recent versions anymore on wsl.
05:33
@rak1507 i how do i join
@ngn are we starting a new repo or pull request ngn/k
 
3 hours later…
09:04
for those who don't know - in the k9 web console you can press Ctrl-[ and see several nice examples. note it is not latest k9
can you? doesn't work for me
oh it's ctrl ]
ml.k in particular is quite cool and you can see the script with 0:"ml.k"
neat
@rak1507 oh whoops. yes, ctrl-[ and ctrl-] to go left and right
 
2 hours later…
ngn
ngn
11:33
@Razetime i think such a project would have to involve at least one person who understands both graphics and k, like JohnEarnest or ktye. i could contribute only a little (probably no more than 1 hour/day), regardless of whether it's built on ngn/k or not
@Razetime @ngn you also need to agree on what you want to do. 3d gpu access, 2d bitmaps with events, 2d canvas with vector api or use a full gui toolkit? should it work standalone as an executable or for the web/wasm only?
My mid-term plan is a data-oriented (html/js) ui interface for the online version.
ngn
ngn
@ktye if i understand correctly, razetime wants to use k as a platform for desktop games (mentioned sdl)
I was wondering if there is tooling that would make sense to share/make compatible with various k implementations (that said: not sure what that could be or look like)
@tosh when i started with k, i just wanted to right a "small lists/cheatsheet" with differences between the versions. i gave up quickly. i don't think a tool would work.
ngn
ngn
11:49
@tosh i don't think it's realistic, either. even k9 alone changes so often, it's hard to keep track.
yeah makes sense, it might also not be desirable to have converging/compatible implementations and also tooling for ides does not seem that important for k
ngn
ngn
ngn/k and oK are fairly compatible and i think that's a good thing. tooling - all you need is a text editor.
one half-baked idea I had was something like a data lake where you have stuff like taxi dataset, tick data, genome sequencing data and coding riddles like project euler and so on and an easy way to load data and select a runtime
then again a zip with files would probably also do the job
@ktye I can probably mess with oK for a bit and make a very slow electron app with a K game in it,
my main want is a K platform for writing 2D games and animations
which probably fits in the "2d canvas with vector API" segment
@Razetime is a web version ok? in my opinion it's more fun as you can share your games/animations and discuss with others. very few people would follow your "build instructions" for a native application/download examples and run them, vs clicking a link.
12:24
Well sure it is fun, but it's a lot slower and impractical for larger games
Maybe I'll try making something decent in ike and form an opinion
 
2 hours later…
14:50
last time I did any meaningful work with iKe I wrote a feature that can record and export GIFs, which can be a really fun way to share your work
esp. since it doesn't require the programs to run in realtime to render them
wow, looks like imgur won't even let you upload a GIF anymore without re-encoding it. the natural life cycle of image hosts continues to play out.
ngn
ngn
15:32
@JohnE a patent is more harmful as it monopolizes an idea in general, whereas ordinary copyright only protects implementations
also, "gagged" patents exist
 
3 hours later…
18:54
valid point with the "gagged patent" but I don't think shakti or kx have the authority to issue one. :)
and to clarify, I am in agreement that neither patents nor copyrights on software techniques are good for innovation
 
3 hours later…
22:14
@Razetime you can decide for yourself after using iKe for a bit, but I think the main thing I would have done differently if I remade iKe from scratch would be to add a symbol to the beginning of each of the drawing tuple types which identify them (b for bitmap, p for polygon, something like that) since it would make the set of drawing operations more extensible
there's an elegance to just having two drawing operations but expanding it slightly might be nice; having a proper line type, for example, or perhaps a text rendering tuple
it would make programs slightly longer, but in the scheme of things it doesn't matter
improving expressiveness often pays for itself

« first day (837 days earlier)      last day (698 days later) »