@MichaelHale How about using Image?

img = ColorConvert[
   ColorNegate@
    Binarize[
     ImageResize[
      Import["https://i.pinimg.com/originals/ab/e8/10/abe810d639b8715e294a294b34cf17f6.png"], 128]], "RGB"];

DynamicModule[{pos, i = Image[img, ImageSize -> Large]},
 Dynamic[
  pos = MousePosition["Graphics"];
  If[pos === None,
   i,
   i = ReplacePixelValue[i, Round[pos] -> {1, 0, 0}]
   ]
  ]
 ]

It's fluid on my machine..

@halirutan Works great. Yeah for just this that is probably best. But I was curious about the technique of partially updating a big list/array in a Dynamic display just in general. And thanks for reminding me to grab a nicer map projection too haha.

I got hooked for a few weeks on these fairly complicated historical strategy games from a company called Paradox Interactive. So I'm doodling around with maps and procedural generation and such haha. I'll get back to fancy 3D graphics soon enough, but I'm glad I'm feeling motivated to do hobby coding again after work. Didn't have the motivation since April, but it came back.

@MichaelHale Well, if you ever need motivation for Mathematica, just use Matlab for a while. I dare you :)

Haha, I don't doubt you. But I have no more excess capacity to experiment with software I don't enjoy. My coworkers' tendencies to burden me with excessive Docker/Kubernetes/AWS Cloudformation/continuous integration/database migration/testing framework tech is all I can bear.

I'm rather amazed at the number of things they've found to do instead of what we were asked.

@MichaelHale Ah, so you finally arrived in the new age where everything is web-first and stored on some Amazon server and each line of code is battle-tested with 100 unit-tests.

I feel a tiny bit sorry for you.

Lol. Maybe it's because my laptop is unintentionally connected through my phone, and coverage is a bit bad here (too much metal on windows?). Thankfully mobile data is truly unlimited around here.

Do plots like this mathematica.stackexchange.com/a/13548/37297 have a commonly used name?

In seaborn they're called "joint plots". Perhaps an indication, more than anything else, that there isn't a generally accepted term.

Is it just me or has Interpreter and friends gotten way slower recently?

@halirutan Haha, well then let me throw in the $600/month that now gets taken from my paycheck each month for strangers in Poland because of some kid I supposedly have that I wasn't told about from a drunken one night stand three years ago when two girls were traveling around the US. Other than work and that though, things are pretty great haha.

But yeah, some people have definitely jumped the shark with the whole unit testing ideology. I can't believe they trust the computer to add 2+3 when they've only written a test for 2+2. They write a test for each API endpoint to verify the server framework correctly sets an HTTP 200 response. One guy even angrily asked me "Yes, but how do you KNOW that line of code will get executed after the preceding line?"

And for those repeated tests that the server framework we are using successfully sets the HTTP status code, the only way it wouldn't is if some function in another "application layer" returned an error, but in the unit test they "mock re-implement" each function it calls to never return an error.

@halirutan Since the beginning of this year I find the CEOing and Physics sessions about combinators the most interesting/entertaining… In the Physics session episode Tuesday 20th 2020 S.W., Roman Maeder and others at 1:16:04 are digging to get Prof. J. Waldmann RX program (via wayback machine) to get some insight into S-Terms decidability.

I know that Prof. Waldmann works at Leipzigs HTWK. Do you have contact to him still? As far as I understood Roman was tasked by Stephen to mail him and investigate further.

@Xacobeo2002 I haven't spoken to him in years, but he was one of my tutors back at university, he taught me juggling, playing Go and we participated in quite some ICFPs (international contest for functional programming).

He was the one who turned my degree specialisation into "theoretical computer science" which was one of the things I thought would never happen. But boy, did he had an awesome teaching style. Who can say no when you learn graph-theory through game AIs?

@MichaelHale That story made my day. Really.

Every time I think to venture into the Graph functionality in WL I end up being turned away by how much pre-existing knowledge I feel I need to have. It's not like this in some other areas of WL, I practically learned machine learning from the WL docs, but Graph? Always tough.

@halirutan Ah, ok. My last contact was 2008 to get one of his students for a research project in industry. I also liked his style and sovereignty in theoretical matters. I am tempted to point Stephen/Roman to him or vice versa. Who knows, maybe it helps on Wolfram’s quest to find the „world formula“ (fundamental theory of physics)...

Like I've got this dataset. But turning it into a graph plot with edges being bolder the higher the value was seems so hard to me

I am more than happy I even managed to get a plot out of it at all, let alone getting the edges displayed like I would like

Not to be a bad workman blaming his tools but I can seriously never get what I want out of the graph docs

@CarlLange Check out EdgeStyle and look at the Details in the docs.

@halirutan Sadly I just can't seem to get it to work with directed graphs. All the examples use eg 1 [UndirectedEdge] 2->3 but I want 1 -> 2 -> Red. I'm absolutely sure it's a bug in my brain and not really the fault of the docs...

ah yes, under possible issues

Sorry, I shouldn't really use this chatroom as an outlet when I don't understand something, I can just never get my head around Graph stuff

@CarlLange I think that's a product of bad-design in Graph

@b3m2a1 Yeah, I dunno. That's definitely how I feel about it, but I've never so much as taken an introduction to graph theory class. I just feel that I quite often want to represent and plot graph-type data without knowing much beyond "I would like it to look like x", and in other parts of WL, that doesn't seem as difficult as it does with Graph

I mean, I can quite SPARQL pretty easily and do a fair bit of work with Wikidata and tripe stores, but that's kind of separate to what I really want from WL for this type of data. I literally just want a nice way to visualise a list of a<-w->b, and maybe do some elementary calculations on it

*triple stores, tripe store is what I think SW thinks of wikidata though