Re: bolder/lighter defintion

>> No, you've missed how bolder/lighter works, the inherited weight is
>> first mapped to an existing weight and *then* bolder/lighter is
>> applied.
> 
> I thought the point of your proposal was that you were making this no
> longer be the case (which I like).

The table of values is based on how the *existing* behavior of
bolder/lighter would work for a family with 100/400/700/900 weights. 
For that family '300 bolder lighter' would be a roundtrip using either
the old definition or the new one.

After the change, bolder/lighter would be a simple mapping independent
of a given family.