Hacker News Reader

Comment by rmoriz | Hacker News Reader

Don't reinvent the wheel if you have no clear reason / vision to do so. Period.

Also most "re-inventions" are driven by a lack of understanding of the existing solutions or the non existing faith to stick to conventions. Many of the conventions are probably a legacy and should be "attacked" BUT if you do that, have a clear vision and reason to do so. Just don't go the road because it looks easier. It won't.

View on HN

Agreed, I think the OP is missing the real meaning of the phrase "don't reinvent the wheel". The phrase does NOT mean "don't improve on the wheel". Nor does it mean, "The stone wheel is and always was great, don't try new materials." It means, "Don't create this whole thing from scratch thinking its original and new, when some research will reveal that it isn't." In patent terms, it means, "Check for prior art."

Linus Torvalds (OP's example) had a deep knowledge of Unix before he began what became Linux. If he hadn't such knowledge, he would have not been successful, for one of the great draws of Linux was its familiarity. In other words, he didn't "reinvent the wheel".

Reinventing just about any wheel will at least teach you a lot about how that type of wheel works, even if your reinvention never makes it into production.

Exactly, exactly, exactly. :)

This is my standard MO for learning big new ideas. Say for example I want to use MVC for the first time. Once I think I understand the general idea, I will build my own. I might peak at Rails or Struts or whatever to see how they solve something, but in general I just go my own way. Once I have something to approximately proof of concept level, I feel ready to go out and evaluate existing options with a pretty high level of confidence in my conclusions.

Reply on news.ycombinator.com