Hacker News Reader

Good article, but the point is that you should roll your own when you gain a competitive advantage by doing so, and you have the talent to actually execute. I would argue that if your situation matches these criteria, you'll already know it. If you had to read a blog post to start thinking you should write your own big data stack, you probably shouldn't.

View on HN

Ha - fair enough point. Perhaps this will help young burgeoning startups give more thought to their options, though.

You can enjoy reading a blog post even if you've started thinking. I liked hearing his reasoning and comparing it to our own.

Of course, I enjoyed the post as well, and I'm always interested to hear how good companies are solving their problems. I just wanted to emphasize their point that the situations where it's appropriate to roll your own are usually pretty specific and obvious once you're down in the prolem you're trying to solve, and if you're a young startup that never thought about it before, there's probably no reason to spend the time and effort when there are a bunch of other things you need to be worried about.

Reply on news.ycombinator.com