Quote Originally Posted by Gerald C Koch View Post
If you look at the rule it says that the chain linking the two developer groups must begin and end with a double bond.
That clears it up quite a bit. Since vinyl alcohol changes its structure to acetaldehyde as soon as it gets a chance to, I naively assume that this would make the linear structure -C=C-OH highly unstable in general, while OH attached to a benzene ring is stable hence has a chance to work as dev before bad things happen to it. There is no reason to risk blow up our little dark rooms with alkadiene-diols after all.

It may all be obvious to real chemists, but for extra clarity we could rephrase the A-(C=C)n-A' rule as A-C=C-whatever-C=C-A', maybe with the hint that the C=C double bonds may(and in all likeliyhood will) come from an aromatic ring.