So I became curious on how the HN community approaches developing their weekend projects.
What are some of your individual approaches?
SHIP IT!
Most projects are just that, projects. They very rarely turn in to a business or even make a penny of profit. I tend to build something that either a) I will use or b) Something that will teach me. But, you need to make sure you don't go down the rabbit hole of development. Have a goal in mind to ship by the end of the weekend, even if it doesn't look great, work exactly right or behave just the way you want it to. It never will, there will always be something that you want to change. That's a good thing. Iterative design and development works.
I'm not advocating the release of useless MVPs with poor design or webapps that frankly don't work. What I am saying is that when you give yourself a deadline and stick by it, you can learn a huge amount about yourself. Try it.
I built http://www.livelystocks.com last weekend because I already used a very basic version of it on my local machine. I thought I would give a little back to the HN community. So, I published it. It hasn't gained much traction, I didn't expect it to, but it still got a few hundred people interested and many people use it every day. I managed to learn a new programming language, get to grips with Twitter Bootstrap and I'm pleased about that. More importantly, it taught me how to properly ship, in a very short timescale. And I enjoyed it.
After that I figure out the data model, create whatever database is needed, open up the docs for whatever new tech I'm learning (that's usually the point of a weekend project for me), and get to coding the actual app.
This past weekend I made http://bookmarkly.com to learn Backbone.js and a bunch of other related things. It's pure JavaScript on the server and client, and I had to build a screenshot generating server for it too, which bled into the week. I'll probably post a "Show HN" after I give all the code a second pass, I'd like to open source the site.
Know the ins and out's of auth and stuff that most of web app does have.
Know what the hell you are building -- development/programming is impossible to master so you will spend most of your time googling shit and fixing errors, try to minimize the amount of time spent.
Know how to emacs/vim/any text editor that makes you productive.
Know how to deploy fucking fast.