by Ryan Walker

CincyReHackathon

CincyHackathon was a great experience.

I met some great new people, I grew as a developer, I got to stay up and program until 3am and didn’t feel like a bad person for doing it :)

I won the contest, but don’t feel “done.”

Not too many people participated, but I did pull down first place and $500 in gift cards, so my wife wasn’t too pissed when I came home bedraggled early Saturday evening.

But the app isn’t live on the internet, due to my lack of Oauth2 understanding and less-than-optimal guidance from Google on how to make it work.

So starts CincyReHackathon

I’ll give myself 12 hours, through Saturday afternoon, to do it better.

DirectedEdge over Google Prediction

Google Prediction let me down quite a bit during the event. Their API has become so complicated, and isn’t well supported or documented for Ruby. I think part of it is because they’ve gone for a discovery-based API, and Google Prediction isn’t very popular, so I was really struggling for working examples.

I did finally come across a sample sinatra app buried in their API library, and it formed the basis for my final presentation, but the Oauth2 concepts (especially the way I was doing it, using the 2-legged variation) was so poorly documented that I literally spent 1/2 my time trying to get that all working rather than on my chosen problem domain.

I did a little googling and have discovered DirectedEdge who I will try to use in my re-do, and were nice enough to give me a free developer account after explaining my intentions.

Test-driven vs. hackity hack

While I started out test-driven during the event, I quickly abandoned the practice when shit started hitting the fan, and things felt very “spiky.” Now that I know what I’m building, I’ll approach this re-do in a more disciplined way.

Live, working app on the internet vs. 1/2 working

My goal at the end of this is a working, snappy app on the web.

Should be fun… here we go!