According to Howells, the site aims to provide the best of the web, and features over 2,500 sites, categorised by style, type and subject. He told .net that it’s a useful tool when doing design research or for helping clients get an idea of the sort of site they’d like. Howells doesn't use any specific formula to curate the site but, he explained, tends to favour clean, simple sites that make the “best use of the tools we have at our disposal, are free from gimmicks or decoration, or are just beautifully crafted”. At the very least, featured sites should be interesting and inspire designers.
However, Site Inspire itself hadn’t been touched since its March 2009 launch, and Howells decided he needed to practice what he preached: “The thing is, it was difficult to incrementally redesign and update the application at the same time, and so it felt like the right time to redevelop from scratch. Also, I wanted to make the site responsive, which can be hard to implement retroactively, and to move from PHP to Ruby on Rails, which is now my preferred backend language and framework.”
On his blog, Howells detailed the course of the development, which went through several radically different redesigns, each a “knee-jerk rebound” from the last. Owing to the sheer length of time on the project, Howells became disillusioned and “almost blind to what I was doing”. He therefore learned plenty of things, and now recommends not necessarily throwing the baby out with the bathwater, along with faster iteration. “It's become quite hackneyed in our community, but I believe in the ‘fuck it, ship it’ attitude, particularly after this long, drawn-out experience,” reasoned Howells. “A website will never be perfect, and not everyone will like it, and so it's best not to worry and just get it out there. That way you won't get bogged down in the process, and there is more urgency and motivation to iterate quickly since your product or site will be live and being used by real people.”