Ernie Schlangen (10/31/2011)
Can't speak to other people's motivations but I (almost) always look up the answers, even when I'm sure that I know the answer. For me, not only do I get a full explanation of the answer by looking it up but I usually learn something else by reading the explanation/following the related links. It's nice getting the points for the questions, but the usable knowledge is better.
If the question is something I know I don't know and can't see any reasonable way of deducing the answer (with a decent probability or its being correct) from things I do know and what I believe to be how MS goes about things, I'll start reading chunks of BoL or other appropriate bits of MSDN (because if I learn things by following the chain through the documentation and getting the context I'm more likely to remember them than if I'm directed straight to the bottom level page that is what's usually referenced in the explanation.
If I find I'm getting nowhere like that, I might resort to google - that brings up a lot of unreliable stuff (which I can easily recognise when it includes statements that I know are incorrect, but it doesn't always include such statements) so it's not where I start from because sorting the wheat from the chaff is a pain, it's easier to start from something almost always trustworthy like BoL. But I definitely will resort to google when stuck.
A question like this one is difficult for most people because google is probably the only way to dig out stuff that old (I'm pretty sure it's not on any microsoft website) so Google is the only resort.
Having had a longish experience with RDBMS and SQL I could eliminate SQL95 (wrong date - I'm sure there's have been a big fuss if something codenamed SQL95 had slipped into 96, and I didn't remember any such fuss), and my experience with SQL Server itself allowed me to eliminate Shiloh, Yukon, and Katmai because I remembered those ones. I seemed to remember Sphinx from somewhere too - and as I was sure I had never encountered the codeword for SQL 6.5 that eliminated Sphinx too. So I was reasonably sure it had to be Hydra, but thought I'd check it and googled SQL Server Hydra and found a blog entry at SQLServerPedia (a blog that I feel I can trust more than wikipedia) which confirmed it.