Ignoring the GPP, a specialist is a 0/0/3 tile which is utterly terrible. In a vacuum, a grassland cottage is much better.
Sure. But why ignore the

? Converting the

to

:
After 100

= 350

(assuming your first bulb is for CoL)
Means 1

= 3.5
So every specialist before the first GP and assuming CoL is 3

+ 3

= 13.5

/turn! (after a bulb)
Then, let's assume second GP (200

) is going to bulb Philosophy (800

):
800

/200

=4

/1

Meaning 3

/3

Specialist is a 15

Specialist (!?).
That seems OP. Is this right? I've always been bad at math. I know there's an article on it at hand, but thinking aloud in a discussion is going to make it stick in my brain more.
For your point on cottages,
@Araius, I know that cottages are OP in the late game - but for the early game with expansion I have always thought that it is better to forgo slow-growing cottages for specialists and just take the cottages of other civs. My principle is that first-mover advantage is more important.
And to your point on "at happy cap," I actually think that specialists are worth it to run prior to happy cap when these are true: a) growth is slow; b) new tiles are marginal (say, ocean tiles); c) specialists are strong due to buildings, civics, or golden age. -> The rationale again being that first-mover advantage coupled with aggression is the game-winner.