Just completed newgle.co in my 24 hour coding challenge – this is a mashup of Google custom search results with some domain filtering and prioritising. Here is how the 24 hours went:
Hour 1
What the fuck is the idea? I want to do something with the search results but what could that be – what about setting up a custom search engine that just shows people instead of results? After a couple of rubbish ideas like that I came up with newgle!
Hour 2
Getting to grips with Google Custom Search is actually very easy. I had to make a decision pretty early on whether to use the API or the web interface. The API offers considerably more functionality but also requires significant investment time to get it up and running properly. The web interface provided by Google is great as a quick start wizard – in all honestly there is a surprising amount of functionality you can manipulate so unless you have a very specific need I’m sure you will get what you need out of the web interface.
Hour 3 to 8
Google custom search allows you to specify particular websites that you want to include in the search results, it also always for exclusion of websites – so I spent a lot of time filtering the custom search engines to get the right balance of websites for each of the summary, detail and person sections.
Hour 9 to 12
Visually, what should it look like? This wasn’t a difficult question because I was basing a lot of it on the current Google UI – interestingly there was no homepage for newgle until about the 15th hour everything was from the search page – I didn’t see the need for a homepage until I decided to give it it’s own website.
Hour 13 to 20
This was the witching hours of the project, the time when it gets to the middle of the night and you’ve been staring at code from the previous dusk-til-dawn, and you question WHAT THE HELL AM I DOING THIS FOR, IS IT ANY GOOD? sometimes the answer turns out to be ‘no’ and then I go to bed. But with Newgle I wasn’t trying to re-invent the wheel I was simply adding a ‘better’ UI to what is already a pretty perfect system.
This realisation drove me forward and these hours were spent adding the best user experience possible to the program. These UX things included a jquery persisant header, a tutorial tooltip and onclick highlighting!
Hour 21 to 23
Testing… testing… and more testing… I hate testing!!
Hour 24
Shit, how am I going to know if anyone is using it? The last hour was the tightest because I needed to add Analytics code and some event tracking to the Javascript. I wanted to track visitors, obviously, but I was also looking to track engagement through the system. So I setup some event tracking that in the future I will be able to use to measure which results: summary, detail, person, are the more used.
Here is the finished piece:
