App in a month

I finished my contract at the beginning of December and as contract work had dried up due to christmas shutdown decided instead of wandering round the house in my underwear talking to the houseplants, I would create a new OS X app. It's a developer tool which I won't screen shot as I want it be a surprise. 

It's not a completely new idea for me as I had tried to make it on iPad earlier in the year but it didn't quite come to fruition as I didnt have the time and it didn't lend it itself to mobile at the time. Also it wasn't a start from fresh thing as some of the components already existed. 

The point is to go from near nothing to near publishing in about a month and I kind of achieved that goal. I created the meta-data on iTunes connect today and I'm just sort of crossing i's and dotting t's right now. All major features and use cases are complete. I'm expecting to wrap things up and hit the App Store about mid Jan 

Some lessons were...

Define an MVP and stick to it. 

It's a tool that could do a lot of things but for now just stick to one basic idea. If I found myself disappearing down a rabbit hole of functionality I stopped and got back to the core idea. It's OK to leave hooks in for future functionality particularly in the core data model but don't think that the first release needs to have everything. 

Do it right once

Use Clean Code (thanks, Uncle Bob) and the responder chain to create functionality and user journeys. A user journey would most likely start off as a item in the main menu attached to a First Responder target. Keeping refactoring and DRY principles in mind all the time meant that the code base is  readable and re-usable. An object has a single responsibility and it's time to create a new class once it needs to do more than one thing. The code as a result is mostly free of hacks and singletons.

It's a modern world

Auto layout yes yes yes! . It's the first project I have used it on for real and it's fantastic once you get used to it.   Collapsing menus and UI items come near free. No more springs and struts for me.

Mavericks support only. We are probably only 6 months out from the next OS X. Trying to support 10.8 or earlier just chews up time. 

Do something on the project every day 

No matter how demotivated, even it's just 30 minutes. Finish a feature. Fix a bug. 

Go outside everyday 

Go for a walk, bike or climbing or whatever gets some fresh air into your lungs. Put some fractals across those eyeballs.

Follow your own timetable

Im not a morning person. Working till late and waking up at 9.30 - 10.00 was the norm. As a consequence for the month I was mostly refreshed and motivated. When I'm employed I'm usually expected to be at work by 9.30 and I can honestly say that yes my body is physically present  but my brain not so much. 

So next time you find yourself with a month to kill give it go. I have no idea whether the app will sink or float but if it does float I'll take some more time to add some more features.