Comment by jansan

I realized a while ago that Vector Networks (I call them path graphs) is actually an intermediate structure that I already have when performing a boolean path operation. I was skeptical of Vector Networks until recently, when I realized that they would be awesome as a result of bitmap vectorization, which has become more important with all the AI generated images. So yes, I want to implement them, but there is so much more mundane work to do that I cannot say why.

I am doing stress testing and test it on really crappy devices (an old 100€ Android tablet is the worst), but this post was quite spontaneous and I should have done a bit more testing before it. I used to deliberately work on an old ThinkPad to immediately notice any performance issues, but the IDEs have become so demanding that this is almost impossible. But I keep those devices around and try to make sure that it runs on older hardware.

Regarding your other suggestions: I am currently going through all suggestions in the comments and will make issues for them. The amount of helpful responses was just overwhelming. I was hoping for about 5 upvotes and maybe two comments :)

Edit: The decision on using SVG for rendering was done because I wanted to make sure that the rendering looks exactly like the saved SVG file. I have been thinking about using Canvas or WebGL for the overlay, but for now it runs fast enough with most images.