I’ve been struggling with plurality and ambiguity as intrinsic parts of the user-interface, but I think sliders and colour labels are now reasonably functional in this respect.
I just like to throw some ideas:
The colour label banner shows the mix of colours, but I don’t see (maybe there is) the logic of the mixing… this could be an opportunity for telling the user something. Maybe you can see the colours without clicking on it, see the attached orrible image.
The size of the ribbons could be the number of recurrences, or the % area cover.
Maybe when you move your mouse over on the ribbons you could see some statistic/info on it, like name of the colour, number of recurrences, or the % area cover.
I do love the colour average/dominant on the sides!
hope that helps somehow…
There’s no real logic. It’s just a random sampling of the unique colours in the plurality. So if there’s 200 yellow and 1 red, the noise will be 50-50 yellow-red. I did this to prevent some colours from being invisible in the noise. I do appreciate that therefore the noise is not a good representation of the actual plurality.
I did try almost exactly that actually, but for some reason it looked very strange. The noise seems to work no matter how similar or different the underlying colours are, whereas the hatch can look really disturbing if it contains conflicting colours (such as red and blue). Maybe if I change the order (sort by hue) I can reduce the side-effects. And maybe if I switch to just rectangles instead of diagonals I an use the width of blocks to convey frequency as well?
That wasn’t even part of the original design! I just had to do that to make the icons and text legible. Turns out it’s impossible to make grey text readable on a black-white noise background
I thought about the sorting by hue as well!
Maybe, you can mix the neighbours colors like this : https://goo.gl/images/xJXiBm
Does it mean that it will be possible to create icons directly in GH2?
The goal is to be able to create icons easily from a collection of bitmaps (using a Rhino or GH command), to create icons from scratch inside Visual Studio just by writing a new class which implements
IIcon, to create icons (relatively) easily in Visual Studio by creating a new instance of the
VectorIcon class and populating it with shapes, and to create icons easily by drawing them in Rhino and exporting them as either code or some sort of icon file.
The latter is what requires the most complicated UI, so I’m focusing on it now. It’s also probably the most common way in which people in the future will create icons. You only really need to use code if you don’t want your icon to look the same all the time.
I would like to make a suggestion about whether a series of libraries like numpy and scipy can be built into ghPython 23333