Doc updates for ChainRulesCore 0.5

[oxinabox]

New docs,
major change is about the pushforward now fused into the frule.

[nickrobinson251]

view at http://www.juliadiff.org/ChainRules.jl/preview-PR152/

[nickrobinson251]

This is great 👏

[willtebbutt]

Firstly, I agree with @nickrobinson251 that this is fantastic work. In no particular order:

• Looks like the Why does frule and rrule return the function evaluation section of the FAQ still requires updating.
• Also, how would you feel about changing all references to autodiff / automatic differentiation to AD / algorithmic differentiation?
• In the "On writing good rrule and frule " bit, we should consider modifying the recommendations for Zero() or One(). In particular, removing any mention of One because it's not generally appropriate or helpful. Additionally, the "Write Tests" section should probably be entitled "Write Tests using FiniteDifferences". Similarly, the "CAS Systems are your friends" should have some reference to them not being helpful when writing tests for derivatives.

[oxinabox]

Also, how would you feel about changing all references to autodiff / automatic differentiation to AD / algorithmic differentiation?
In the "On writing good rrule and frule " bit, we should consider modifying the recommendations for Zero() or One(). In particular, removing any mention of One because it's not generally appropriate or helpful. Additionally, the "Write Tests" section should probably be entitled "Write Tests using FiniteDifferences". Similarly, the "CAS Systems are your friends" should have some reference to them not being helpful when writing tests for derivatives.

These are valid points, but beyond he scope of this PR. Open an issue?