New Site Concepts #2: Evolving Design
When I launched this site over 4 months ago, I started to write about some of the concepts I've been playing with while building the site. Until recently other commitments have got in the way of writing more, but I'm finally finding time to catch up on things.
The second concept I've been playing with is around the area of iterative design. It seems most websites, both professional and personal, are redesigned in one go. Lots of work is done behind the scenes on the structure, navigation and visual design, then the site is launched. Things are then kept fairly constant for months, or even years, until the next big redesign, where everything is thrown away and the process is repeated.
There's a couple of problems with this idea:
- By starting from scratch, it's easy to lose any of the benefits or learning from the existing design
- Redesigning everything requires a significant commitment, which isn't always available
- Any feedback you get after you put the site live is too late to inform the design process
- It's easy to create inflexible designs that cannot accommodate unforeseen features or functionality
Instead, it seems (to me at least) to make more sense to continually improve and tweak the site - each time making a small but measurable improvement. One of the unique attributes of a website is that it's never finished, and can be changed as often as you like. Using a process that doesn't reflect this seems somewhat foolish.
Some of the best sites on the web have been iteratively designed. For example, a year and a half ago Flickr was entirely focused around a flash based drag-and-drop chat interface, and now it's a HTML and javascript based browsable, bookmarkable website. The transition was gradual, and incredibly well managed - there wasn't a single point in time where the new site replaced the old.
There's also obvious comparisons that can be drawn with Stewart Brand's How Buildings Learn (reader notes here, here, and here). Also related is Dan's discussion of adaptive design from several years ago, which probably inspired some of these thoughts.
So the idea with this site was to start with a very plain website, and see if I could evolve it to something more interesting. Behind the scenes, I set things up to ensure that any changes to the templates could be quickly and easily rolled out across the site, and thought quite hard about where associated assets would live to avoid problems with caching (I'll write more on both subjects some other time).
But it's clear that over the last 4 months this idea of rapid iteration hasn't happened for this site.
To try and change this, I'm going to try making these iterations more public. From now on, most changes to the site will be accompanied with a short description and possibly a screenshot. Here's some from earlier today. They'll appear both in their own section and among the rest of the content (although I'll probably try to work out some way to reduce their prominence). A dedicated RSS feed is available, for those who are particularly interested.
In addition, behind the scenes I'll be saving copies of the templates, html and associated assets to allow progress over time to be visualised at some point in the future.
This is an experiment. I have no idea what will happen, although there are two obvious predictions I could make:
- The site will evolve slowly and in an uninteresting way, creating 6 months worth of dull one-line trivial changelogs before I give up and redesign everything.
- The site will evolve quickly, leading to an interesting design and generating lots of data and useful lessons on how to approach iterative design.
I imagine the end result will be somewhere in between. Either way, I'm sure I'll learn something.