[funkaster]

I've been on both sides of the fence: being interviewed and interviewing. The times I've been interviewed I felt (most of the times) that either the person questioning me had no complete understand of what he/she was asking --it could be possible he/she was a junior engineer-- or it was a tricky algorithm question. The thing is that not everyone thinks or plans the same way. Analytical thinking is different for everybody. Most of the time I like to solve problems first either by brute force or the simplest solution and then try to improve it, but the author of the post also got a very good point: "why".

When asked questions without a context, you have no motivation and they seem weird. If you want to know if I know what a b-tree or a hash-table is, or what is database normalization, just ask me to explain you what it is. I also, wouldn't expect a candidate to implement one of those things right out of their head... that's just crazy. I would expect for him to be able to use them or at least how they work. If you need to see what I've done or how I perform against a particular issue your company might be having, code review of already code-reviewed code is a great, great way to test a candidate.

Github profiles are also a good indicator: it shows what you want, want motivates you, it can show real-world code that you have written.

Just talking of your experiences and how you've solved problems is also a good thing.