I'm looking for somebody to help me develop some of the backend for a major website project.
Imagine a website where you could type in the name of any major company, brand or product into a search box. The search would return not as a traditional Google-like list, but rather the results would be grouped into key issues (sort of like Nuggetize or Clusty (now called Yippy Search)). Results that did not relate to that product's social and/or environmental impact would be culled from the results. Results that were essentially duplicates (with the same content, even if the page was different) would be further culled. The remaining articles would then be further subdivided into those that were favourable and unfavourable towards the product/brand/company.
For instance (just to illustrate), if one searched for Cadbury, there are a few sites that say negative things about their use of Palm Oil in chocolate, and a few positive articles about how they changed in 2009 and stopped using palm oil. Those would be grouped under "Palm Oil". There'd be a whole bunch of other enviornmental issues too, such as their recent eco-egg campaign (grouped under "Carbon Emissions") and social issues like the Cadbury Coca Partnership (perhaps grouped under "Sustainability"). The whole point is to see at a glance what the good and bad points are about our consumption of various goods, without having to scan through dozens of pages of Google results.
The website you are imagining is only one medium-sized part of a much grander and more ambitious project that is being launched near the end of the year - and that's where you come in. Writing good, clean and understandable code is VITAL. What you'll be creating is only one part of a whole system. As well as your bid, if you are successful, your name/logo will be displayed on the "about" page of what is poised (we hope) to become one of the hottest start-ups of the year!
Here's what I expect you to deliver:
* A small, fast script that does the above reliably for any given product/brand/company
* Another script that determines whether a particular product/brand/company is relevant to a given issue (accessed via AJAX)
* Do heaps of testing
You don't need to worry about any UI/layout stuff. If you don't feel like writing a whole new spider (not sure that my database could handle this anyway), feel free to simply poll the other major search engines (especially news search engines).
Since the project is ultimately not-for-profit it does not have a huge budget attached to it, but at this stage I'm open to considering all reasonable offers. Thanks for your consideration and I look forward to working with (one of) you.