I'd wager there's a difference between what jemfinch is talking about and what you're talking about.
I used to unilaterally believe that coding exercises or algorithm quizzes on interviews were silly because if I had to solve some given problem I would just google it. And that's true, I would. However, as I've gotten a little more experienced I've realized that there is extreme value in expanding the domain of problems for which I don't need to google, because either A). I've happened to read about it on my own time or B). I've encountered a particular troublesome component before.
In other words, I think it is entirely reasonable for the hiring process to properly evaluate a programmer's toolset. Yes, I think researching via Google is or should be a piece of everyone's process, but it matters greatly what kind of things you have to Google.
Do programmers encounter all sorts of algorithms when engineering software? In my experience, the only things I really need to keep in mind what are the consequences of certain data structures and algorithm types rather than their specific implementation, because often the implementations are built-in to the language, but the choice of which implementation of data structure and algorithm to use are more important. I use sorting algorithms very frequently. Usually it is just a call like "list.sort()".
Yet in interviews, the questions I get asked are more like "Implement <My favourite sort> in 20 minutes, no googling.
I mean, really?
The other day I was in an interview where I had to implement a queue in 30 minutes. The unit tests were already written and the IDE was eclipse.
I wrote the "correct" implementation for the queue in about 10 minutes but there was a bug I couldn't find; I've never used eclipse before and took a lot of time figuring out everything. In the last minute I finally tracked down my bug. I forgot to increment a variable. I fixed it and all the tests as well as the performance test passed. I didn't get the job because I took too long. (It was a graduate position and the last time I used Java was 2 years ago)
I was distracted with all the buttons on eclipse... Otherwise I'd never have missed the variable increment! I swear. Anyway the point is, if a programmer is ever implementing their own simple Java queue in production software it is not a good sign anyway. Why are you testing for that?
I used to unilaterally believe that coding exercises or algorithm quizzes on interviews were silly because if I had to solve some given problem I would just google it. And that's true, I would. However, as I've gotten a little more experienced I've realized that there is extreme value in expanding the domain of problems for which I don't need to google, because either A). I've happened to read about it on my own time or B). I've encountered a particular troublesome component before.
In other words, I think it is entirely reasonable for the hiring process to properly evaluate a programmer's toolset. Yes, I think researching via Google is or should be a piece of everyone's process, but it matters greatly what kind of things you have to Google.