|Density Discount [message #26196] Tue, 19 April 2011 07:48 UTC
I really like the density discount for WSF. However, there are somethings about it that are really weird.|
First off, the current calculation for density is just weird. The orientation of your model in the STL file can affect whether you get the discount or not. Also, I feel that spherical/circular objects (like pentagons, for example) are not really treated fairly. Taking the bounding box for those shapes is a bit weird since they can be stacked far more efficiently than just by bounding box.
But okay, it's probably a quite difficult problem to pick the correct bounding box. And perhaps the density discount is already a bit too powerful for you at Shapeways, I wouldn't want you to go bankrupt.
However, what annoys more more, is this:
If I have a 80cm^3 model in a 10*10*10cm box, the density is 0.8 so you don't get the discount. The price becomes 1.5*81=$120. However, if I just add a little extra material to my model and make it 100cm^3 (in the same box) then the price becomes 100*0.75+16.5=$90.
So I've gotten a $30 discount just by adding some dead weight to my model. This means I get paid $30 for taking 20cm^3 of extra powder. This seems weird and wrong to me.
Therefore, I suggest the following pricing algorithm for WSF:
IF density > 0.1 THEN price=min(price with discount, price without discount)
IF density < 0.1 THEN price=min(price without discount, price with density discount for a model with volume boundingbox/10 cm^3)
This seems fair fairer to me than the current situation. It would of course be even nicer to have it say "you can add X cm^3 to your model without increasing the price" but that's perhaps a bit of a strecth.
I'm not suggesting anything new or anything that makes it cheaper for me. My suggestion only has the effect that it's easier for me to the the optimum price, and it also discourages me from taking material from you where I don't need it. My suggestion alters the price graph like this, adding the green line:
Of course, it would be nicer still to have a pricing model for which the price always increases when the volume does, and the price always goes down when the density comes up, but that's going to be difficult to implement.
[Updated on: Tue, 19 April 2011 07:55 UTC]