Show HN: Poxi – A Hackable Pixel Art Editor
github.com
github.com
But it looks really cool anyway! :)
https://buddhists.org/buddhist-art/the-symbolism-behind-the-...
@Simpliplant has a good point, people expect digital paint tools to save, but also made a blanket unqualified statement about creating anything without saving.
@kgwxd replied that some people don't want to save, the blanket statement doesn't apply to everyone. And that's absolutely true. This is a valid, though somewhat tangential, point of view about performance art.
There is a long history of artists using art tools that don't have save or undo. Traditional arts are still founded on this concept, the art world still largely values techniques that involve the risk of screwing it up, and are hard to copy. That's changing over time, digital arts are growing up, but the majority opinion in the art world is still that tools with undo and tools that can replicate without loss are for unserious artists.
Music is the same way, it's a performance art, and digital sequencing and digital performance have often been considered inferior. Robert Moog, one of the most famous creators of analog & digital synthesizers, even wanted to make his synth a performance instrument, and didn't like the idea of using a sequencer. Sound patches couldn't be saved either, and many famous musicians had enormous careers using Moogs.
(Unless the tangent leads toward a mind-killing subject like politics. I'm 100% behind you on applying careful consideration before making a tangent from a technical topic to an emotionally-charged one; people see that kind of subthread and never even make it to the rest of the subthreads in the same conversation.)
For me, the best part of having interesting conversations on HN is reading all the alternative points of view by smart people. Not much "productive" happens, from my point of view, by criticizing someone's show HN project, not much "productive" happens when you make blanket statements about what all people want. And not much "productive" happens by only agreeing with others.
So there might be another hidden motive behind what @kgwxd said. The comment he replied to literally insults the Poxi project, and perhaps @kgwxd was merely defending a Show HN project that's in a very early stage.
In my view, you may be defending the wrong person here. I want to see more show HN projects, and I want people to feel comfortable sharing their projects without fear of getting ripped to shreds, called a joke, and questioned why anyone would ever want to do such a thing. Let's be supportive.
So the point was to teach the commenter a lesson by responding unhelpfully?
>In my view, you may be defending the wrong person here.
So, because I was pointing out someone not responding constructively I am by necessity "defending" all points the other person makes? I think that is faulty reasoning.
But yeah - it is cool - great show for the initial offering.
I'm using Tor Browser 7.0a1-hardened (based on Mozilla Firefox 45.7.0).
It's also open-source: https://github.com/juliandescottes/piskel
I am also working on an infinite grid pixel art tool :)
In your case DataContainer would be some part of the canvas, maybe a 32x32 pixel area. And the int key would be the index of the x&y coordinates of that particular area. The image dimensions could then be a max size of the square root of an unsigned int, and to get the index of each tile you'd just do (y * dimension + x). You'd throw all this into a little wrapper so you have a "virtual" canvas that could be as big as you'd ever need it to be.
Also, don't actually initialize each piece of canvas until a pixel is actually placed in it, instead returning a static empty section of canvas if the key isn't in the hashtable yet. This allows for zooming way out smoothly.
https://en.wikipedia.org/wiki/Sparse_array
A cousin to this approach would be using some type of space partition with the assumption that most of the partitions represent large, uniform (empty or solid-filled) areas:
https://en.wikipedia.org/wiki/Space_partitioning
You can see that these are fundamentally related ideas- pay for the complexity of the data you store rather than its volume.
e: Another consideration would be keeping level-of-detail versions of each chunk of canvas as well, so zooming very far out loads/shows a lower resolution rendering. That would let you make a Google Maps style collaborative pixel art canvas or something.
Edit: still nice initiative.
I'm sure he didn't mean it personally.