I finally finished revising the "How to Think Outside the Box" presentation I gave at the Technology Student Association conference over the summer. The original presentation mostly consisted of images and spoken content, so I prepared a text-based version for sharing. You can view or download the presentation at Slideshare and see an incomplete collection of related resources.

The presentation is centered on a model of the design process which, to me, represents the key to effective "design thinking" and problem-solving: the combination of divergent thinking and convergent thinking. Divergent thinking involves the exploration or generation of many ideas, possibilities, sources of inspiration, problems, solutions, and prototypes. Convergent thinking focuses on the reduction and refinement of those possibilities. To fit together in a useful way, both divergent and convergent thinking rely on analysis and synthesis. (However, it is important to note that divergent thinking is not limited to synthesis, nor convergent thinking to analysis—these processes work together in many ways, a point which Tim Brown briefly discusses in Change by Design.) In an effective design process, one alternates beween divergent and convergent ways of thinking.

One of my main arguments is that both divergent and convergent thinking are important throughout the entire design process, from start to finish. This idea is often lost because most discussions about divergent thinking or "creativity" are limited to strategies for brainstorming, which in turn is primarily used to generate solutions. This divergent/convergent alternation, however, strengthens the design process at every stage, from problem-definition to implementation and evaluation, and through many iterations. Divergent thinking, then, is not really about brainstorming solutions to a problem but about approaching a question—of any kind—in a flexible way. A core skill for a designer or problem-solver is the ability to bring an explorative approach to each new question and situation. This applies equally to technology, business, policy, and leadership.

For most people, divergent thinking is much more challenging than convergent thinking, but it is a skill that anyone can develop (and continue to develop for a lifetime). In my presentation, I briefly introduce a number of strategies and processes that can help with explorative thinking: brainstorming, visualization, prototyping, observation, research, and iteration. These strategies, depending on the context, can be used at many stages in the design process, when different types of questions and challenges appear.

The presentation, since most of the photos have been removed, can be shared freely, and you can re-use the content. I hope it's helpful—if it is, let me know what you've done with it!

Other News

Over the summer, I also put together a whitepaper on website optimization (posted at the ChiefDelphi forums). It was written primarily for FIRST Robotics Competition teams who are learning to produce websites, so my choice of content was based on the most common recommendations for FRC teams' websites in the past. The document, however, will be useful for anyone learning to build websites. More whitepapers on other aspects of web design/production are forthcoming.

If you're returning to IotB, you'll notice several small but important (and exciting!) improvements to the site. The biggest change is that comments now support text formatting using Textile, which allows visitors to format their comments without writing HTML. At the bottom of this page, you can see a short guide to using Textile and preview anything you've typed.

In addition, each article page in the blog comes with a "permalink" and "short link" on the right-hand side. The links are now accompanied by a small Flash widget that makes it easy to copy the link URL to your clipboard. The tool is a modified version of Philip Hallstrom's fork of Clippy.

Finally, for those interested in the behind-the-scenes details, the site's caching mechanism has undergone continuous development since the site was launched. Now, it is nearly finished. Each web page is dynamically generated by the website when it is requested; however, the cacher immediately stores a copy. The next time the same page is requested, the cached copy is sent to the visitor, which saves time because the page does not have to be generated again. Meanwhile, the cached copy is stored as a compressed GZIP file; it can be served from a flat file, pre-compressed, and without the need for a lookup table or any substantial PHP processing. On top of that, the site cache is organized so that a PHP program can automatically locate all versions of a given page, regardless of the GET variables originally used to access the page. When a new comment is posted on an article, the site takes advantage of this possibility by clearing all cached copies of the article. This means that there is no lag between the time the comment is posted and the time the caches expire. (The exception is the lag created by client-side caches, but that is controlled somewhat tightly by ETags and other headers.) The caching mechanism does not have to query the database to find out if comments have been posted, either. Each page is served up as quickly as possible, and caches are updated precisely when something has changed.

If you're interested in reading future articles, don't forget to subscribe!