Sigh. Oh, man, this drives me insane. I'm sure I'll think of more over the next couple of days.
But here's the biggest example of #FAIL: https://www.icgc.org/node/1083?destination=icgc%2Fcontacts
And although it is not entirely up-to-date (some examples don't work), I really like examples like the BioMart FAQ. The reason I like this is because it provides "solved examples" that you can load up yourself. We find that end users who want to solve problems could use stuff like this--with the basic query features--and then modify them for their favorite gene, region, SNP, whatever. That button click to load the query is particularly handy.
However, I don't think that's all that I would have expected in the FAQ. I would have wanted more foundational questions about the site/data/etc. But I like the sample queries.
I'll keep thinking of other examples and come back and edit as I think of good ones.
EDIT: Ok, here's what I think of as a typical example--PDB. They have one for web site users, and one for depositors. I think that's nice too, to separate out the uses. http://www.pdb.org/pdb/static.do?p=home/faq.html That's the User's one.
Here's one we like: http://www.ncbi.nlm.nih.gov/Structure/biosystems/docs/biosystems_faq.html This is clean, covers the things we expect in a FAQ, and gives quick answers but offers more depth or the appropriate links to answers.
Let me add that one of the things I want the most in an FAQ is which paper to cite when using the site/data/etc.