Everyone's OpenID is a bit different and I ended up with a lowest common denominator approach which was suboptimal in my opinion.