Recently my 2nd year University of Nottingham group project fell somewhat into the "public eye" by chance; a photo of myself and teammate Katie was shared on the university's official facebook page and was shared by a number of people including my previous school.
Because of this, I thought I'd write a quick blog post explaining what the project was, what use it had and how we went about doing it. Obviously I don't take all the credit for the project, the proposal was not our idea and was provided by the university. In hindsight I think we managed a decent approach to the problem and, as group leader, I'm proud of the way our team worked together more than anything else.
Serendipity Search is a simple concept: coming across a "happy accident" when searching the web. One important foreword is that this isn't, at present, a search engine. It takes the results of a Google search and provides them to the user. But here's the important part: because they are receiving the results by proxy of us (SS), they are getting completely straightforward results without any targeting.
By "targeting" it's important to understand how Google orders results at present. Google tracks you a lot - even if you aren't logged in, they have (allegedly) 13 different ways of targeting you. This gives you relevant results not only to your search query but also to everything else. For example your interests, location, friends and more. This is great and many people find this a good feature, however some people don't and that is where SS comes in.
The anonymity provided by us of course has the immediate benefit of unaltered Google results, ordered by what they deem to be the most relevant to the query rather than the most relevant to you. This can be helpful in areas such as research where someone might want to very quickly find information within a certain context which goes against their personal interests and beliefs. For example, if all of your Google results are being targeted towards your location, but you want to find things in a more general sense, that's quite difficult to do via Google. If you go through SS, you get completely non-contextual results.
Services like this already exist, a simple proxy would do the same thing, but SS provides a permanent solution to it, however gives back the control of some additional Google functionality which others might still want. We have an "interests" system which will highlight certain results if they match any of your interests. You, the user, specify all of your interests. You have complete control over a very simple list of things which interest you, a single textbox and a button. You can remove any interests, view all of them and even import them from facebook. The important thing to note about the last stage is that we don't store any of your facebook information, we simply import the "Likes" as text and save them in your account, the connection and permissions are then relinquished.
When searching on SS, you will get all search results relevant to your query but things which might interest you, essentially how Google might target you, are given a visual highlight. Most importantly, they stay in the same place in the results list and aren't shuffled around by Google to encourage more clicks.
Post will be edited in due course with technical information on the project, possibly eventually open sourced.
At present we have no future plans for the project but if anyone is interested we are likely to open source it.
