Reinvent the wheel or finding if it has been done?

Question

Quite often I find with the dilemma of building something that would take only about 10 hours to build, and maybe 30 hours to perfect. Most simple things like this have already been built elsewhere, it just takes time to find them, but there's no real guarantee that they exist. And even if they do exist, it takes maybe around 1-5 hours to realize that they're not really what I'm looking for.

A common example here is looking for a productivity app, when it's quite simple to just code one yourself. Or searching and sifting through questions before asking.

So, my question is, what is a good ratio of time saved to time spent?

I personally try to stick to around 10% of time searching for something, compared to building it. Like if I expect something to take 30 days to build, I'd spend around 3 days looking to see if it's already been done, before I give up and just work on it.

Answer 1

Searching

This seems like a good ratio. You can however try to improve your search to get more chances to find what you're searching for.

Create an organized list of search material to refer to when you need something. When you run in an intersting site with good references, mark it down in your favorites or in a file. Do the same with a book. Some are general enought to find a lot: Wikipedia, StackExchange, ... You don't have to search all by yourself: you can ask here, for example, or your friends. Ask before you need it, so you give enough time to get answer and avoid being in a hurry.

If you're searching for something that'll have a big impact, a good learning method, programming language or operating system, spend more time searching for it. I'd say about a few weeks for those.

Helping other people

If you eventually found something which you think was too "hidden", you can write a page on, i.e. your blog, or answer question sho ask about it if you see them.

Reinventing anyway

There are times when reinventing the wheel is a good thing, even if there are tons of wheels available. Image you're programming a game. You can either learn to use a physics engine and other libraries or write them yourself. You'll learn a lot, be able to use the other engines quickly and better that if you hadn't made yours, if you're not satisfied with it. You can also make something better, but it doesn't matter if you end up with the same code or API.

Answer 2

Reinventing the wheel seems highly unproductive to me. Unless the benefit of the app is HIGHLY transformative to my life, I would rather look for something that would satisfy the need.

Answer 3

Merlin Mann, of the 43 Folders fame, said somewhere that a geek is someone who, if a task takes 10 hours, and he can write an application in 9 hours to automate it, does it. You made me remember the quote :)

That said, 'reinventing the wheel' is an expression with not enough granularity when it comes down to productivity, which is a highly personal field: no matter how good an application out there is, it will never match 100% your personal tastes as much as something you've made for yourself. And, as a bonus, the process of building something will increase your understanding of the concept behind.

Also, poetically speaking if you will, it is nicer: in the past there was a time when a man needed to build things with his hands. A hatchet, a bucket, a bow, whatever. Now you don't have to, but the next best thing is making your own computer applications. The pleasure of having a tool that is useful and that was made by yourself, for yourself, is always a thrill...