[csallen]

Not too long ago I would have agreed with you, but I've recently changed my mind. Coding on a whiteboard is not completely useless:

1. It's a good indicator of how much you know for simple problems.

A few weeks ago I started helping a friend with a particular language and framework. He claimed to know it, having read several books on it. So I said great, go to the board and write a function that adds two numbers together. And he couldn't do it. Who knows, maybe he could have done it on a computer? Maybe he would have Google'd it quickly, or relied on code autocompletion, or the muscle memory from his fingers typing it, etc. But that's not what I wanted to know. Anyone who's written thousands of functions in a particular language could write a simple one like this on a wall, using paint, while hanging upside down. His inability to do so told me: he is not well-versed in this language. PERIOD.

2. You can practice.

"So what", you protest. "Google isn't going to ask simple problems. They'll ask hard ones." And maybe they will. And maybe having to solve it in a foreign way really will fuck you up. But my question to you is: If you know you're interviewing at Google soon, why the hell would you allow whiteboard coding to remain foreign to you? If you suck with whiteboards and you care at all about getting the job, then buy one and practice solving problems on it. Why would Google want someone who doesn't care enough to prepare for their interview? It's not that hard of a skill to pick up. And I daresay it will make you a better, easier-to-communicate-with programmer.

[zellyn]

"You can practice."

This.

If there was one thing I could tell all applicants at Google, it's to read up on the process, and practice. There is a huge amount of information out there[1]. I feel sad interviewing really smart, capable people who could have breezed through with just a few hours of review and practice.

[1] eg. http://courses.csail.mit.edu/iap/interview/materials.php

[jarek]

I imagine it must feel horrible interviewing smart people who would do great at work but whom you are unable to hire due to arbitrary hoops set up by your own company.

[bishnu]

This is true. However, the alternative is worse (hiring someone unqualified)

[thinkdevcode]

> 1. It's a good indicator of how much you know for simple problems.

Google doesn't solve simple problems.

> 2. You can practice.

Maybe if the interviewee was still in college they could "practice" coding on a whiteboard, but considering they may have other obligations like family and work they probably won't have time to practice such a pointless skill.

Personally, I wouldn't need to practice in the first place as I interview extremely well and could easily code on a whiteboard, I just find it antiquated and pointless to do so which was the reason for my comment - not because I am bad at it.

[csallen]

  Google doesn't solve simple problems.

Yes, but that doesn't mean you don't need to know how to solve them. If you can't solve simple stuff, you won't be able to solve hard stuff that depends on knowing the simple stuff cold.

  they may have other obligations like family
  and work they probably won't have time to
  practice such a pointless skill.

Yes, that's a possibility. But once again, if you're Google, you're getting 10s of thousands of applications a week. You can afford to set the bar as high as possible. Every Google employee I know regularly works 10+ hour days there, and sometimes does work at home, too. If you have two applicants of similar skill, but one has more free time to devote to your company, it's obvious who you'll prefer.