Just to say, I didn't know the answer to this. My first assumption was that he wrote the code first (as the whitepaper has some excerpts of coding) but now looking online I see that v0.1 was released on Jan 8 2009... meaning to say the whitepaper was published at least two months before the first version release...

My guess is that the whitepaper still had to be completed after he compiled his first coding, so would still believe it had to be written first. He had to solve the problems Bitcoin aimed to do (direct, trustless p2p payments)... the whitepaper was not so much a proposal for a way to find the solution, but a proposal to use his creation as the solution.

If I'm wrong, and the whitepaper was, indeed, published first... then he did it to gauge the cypherpunk community's interest to decide if the code was worth completing.

Note: I use third person "he" but only figuratively.

P.S. Thanks for this initiative guys! I'm sure I'll learn from the responses.