-
The most important rule of all: Use the products you build for at least a week. Try all edge cases in real-world environment.
-
Classic but true: Know where to focus. Build one product and make it shine. This applies to everything:
- Restaurants with narrow selection of dishes are usually better than the ones which offer everything from pizza to sushi
- Apple has succeeded with a small selection of devices instead of building something for everyone(..Nokia)
-
Perfect products surprise with diamond-like quality of details. Aim for perfection but know when to stop.
-
Delegate. If another product does its job very well, integrate it to your new product instead of making your own, shitty duplicate of the already proven quality product. Delegate small tasks to existing products and focus doing integrations well. This isn't always possible though - sometimes the "quality" products are actually not that good and you need to build your own.
-
If you're curious and think everything is crap by default, you'll still end up building duplicates of other products. It's not waste of time at all. You'll learn a lot by comparing your solutions to other peoples'
-
Never leave nasty work later, do it right from day one. If you don't have discipline, you'll end up with shitty, unmaintainable code. Spending hours on one "stupid" config file is not waste of time, you'll gain deeper knowledge of the tools you use and it'll save time later in your project/career. As a bonus, you'll improve as a developer continuously.
-
Automate all tasks you consider as "monkey's job". Don't just quickly hack something together but instead think how you could use the automation scripts in your next projects.