Great advice, all. The HTTP protocol skill is something I didn't even consider.

ATM Machine. PIN Number. HTTP Protocol.

You're playing with the capitalization.

Gopher protocol. DNS protocol. IRC protocol. HTTP protocol. HTTPS protocol. All fine.

For the less technical it sometimes helps to qualify what something is even if it is embedded in the abbreviation. But considering the audience here, you're right. I will remove the offending word.

RAS Syndrome.

http://en.wikipedia.org/wiki/RAS_syndrome

What are a few costly mistakes that can result from lack of HTTP knowledge?

I've seen a number of times when people don't understand the statelessness of HTTP, and try to put megabytes of data in a cookie-backed session--unable to understand why that doesn't work, etc.

This is especially prevalent in .NET land, where there are mechanisms for storing data in hidden forms and hiding the statelessness away from the user.

not designing with REST in mind

Why is that a costly mistake? REST is great but it's not the only way to do web development. Read up on continuations, for example. In fact, this very website is using continuations (check out how paging works) and it doesn't seem like a costly mistake.

Any recommended resources on getting a good foundation for HTTP?

Chris Shiflett wrote a good HTTP handbook, I recommend his blog too. (typing from phone, pasting url not easy)