Google – Thousands of Wrong Answers
One of this week’s guest speakers, Chris Bunnett CTO of Adgistics, said an interestingly controversial thing. To paraphrase, “Google is awful, it brings you back so many wrong answers and leaves the user to wade through them.” An astute observation, but not one many people thing about, given how ubiquitous Google has become in our lives. Putting more hay on the haystack doesn’t make it any easier to find the needle. When will the web be able to find the single perfect answer, rather than thousands of wrong ones?
This was also the type of problem that a team of us had a go at solving in our “Makerthon” this week…
Our “Makerthon” project was Smart Twits, based on an idea I pitched to the cohort about having a solution to the problem of looking at Twitter’s trending list and not understanding why it’s trending – or as I described it, “is Stan Collymore dead?” which is usually the first reaction to seeing someone’s name trending.
With Richard, Marcin, Bibiana and Emily we created a single page web app (see image) that had two key components;
- a scheduled refresh of top trends and latest associated tweets that writes summarised information to a set of flat files, every 10 minutes, as Twitter updates it’s trends.
- a front end that reads the flat files and presents the information immediately to the user.
For a project that we started Tuesday lunchtime and presented back on Friday afternoon, it was a great piece of work, and a good trial for our final project (weeks 10 & 11) and a good introduction to team coding and project management (using a Trello kan ban).
From a technology perspective the core was Ruby, within a Sinatra framework, tested with RSpec, HTML 5 front end with a good measure of CSS (based on a grid model, but not Bootstrapped!), a D3 word cloud (with a bit of jQuery), plus a Unix based job scheduler and of course utilising the Twitter API.
Our other speaker this week was Charles Davies, CTO of TomTom. A couple of really interesting things stood out in his talk;
- he was part of the initial Psion team, which brought back fond memories of my first handheld computing device, the Series 3c
- the convergence and divergence of software / hardware. TomTom first gained prominence with selling their own hardware, but now sell their software for integration with others hardware – but as they push boundaries with new software (eg. fitness), there’s not necessarily the 3rd party hardware for it, which sounds like it could go along a similar cycle.