support fvar/avar userspace in variable font?

hi! first of all I am loving this library so much. THANK YOU.

RE: the subject. I am trying to figure out if this is a design decision, a TODO, or something wrong with my font! couldn’t find much in the docs. but it seems I can only access, for example, a wght axis from 0.0 to 1.0 (would expect 0-1000 for weight), and also whereas I might have an avar table with fvar (userspace) values like 100, 200, … , 900 which is what I’d expect to be able to use. (e.g.: https://simoncozens.github.io/userspace-and-designspace/)

[jpt]

Thanks for the tip on scaleVariations! I can see for the purpose of animation in particular, that normalizing everything between extremes at 0-1 is helpful, though. There’s a lot to think about RE: how people will want to use those values, ultimately. I personally like your dead-simple solution to this even if it isn’t in keeping with OT spec. have been having a blast with coldtype!

Barlow is not a good example for this as it ignores userspace entirely 😅 one day!

[jpt]

@stenson just to better illustrate this in python form in case it’s helpful (and really only thinking about the weight axis here), this is pretty much the problem I was working though. I provide a userspace value like 400 (which I know is regular weight), I get 0.3505 in return to provide to coldtype

def precise(number,precision=4):
	return float(f'{number:.{precision}f}')

def transformToRange(c,d,arr):
	values = []
	a = arr[0]
	b = arr[len(arr)-1]

	def affine(x,a,b,c,d):
		return (x-a)*((d-c)/(b-a))+c
		
	for x in range(0,len(arr)):
		values.append(precise(affine(arr[x],a,b,c,d)))
	return values

designspaceCoords = [18,38,62,86,104,130,152,180,212] # internal coordinates based on stem width - a pretty common type design pattern
userspaceCoords = [100,200,300,400,500,600,700,800,900] # maps to CSS values, also pretty common

def getzeroOneSpaceByUserCoord(userCoord):
    zeroOneSteps = transformToRange(0,1,designspaceCoords) 
    for i,coord in enumerate(userspaceCoords):
        if coord == userCoord:
            return zeroOneSteps[i]

print(getzeroOneSpaceByUserCoord(400))

of course, userspace inputs can be arbitrary, so it’s a bit more complicated, but suits my use case for the moment anyway 😃

looking forward to digging more into the rng and scaleVariations example you provided.

I really cannot stress enough how awesome coldtype is!