Monday, November 26, 2007

The Web we Weave

The slow death of the web developer's technical spec

By Richard Banfield

The technical spec as we know it is dead. In a world of intensely visual design, we have to ask why we still need to write massive documents to describe web products that real people will use. Huge improvements in bandwidth, processing power and software efficiencies, combined with precipitously lower costs in design and development, demand a new approach to building web applications and websites. The time has come for the designers and developers of web products to follow the example of architects, product designers and manufactures and adopt the prototype first approach.

There is agreement among all of the teams I have worked with that the more difficult and complex a project, the more likely there will be problems. Conflicts arising from time, resource and budget mismanagement can be attributed to one of two problems -- not having the right team members onboard or using an overly complex process. This is not to say that conflict is to be avoided at all costs. Our goal is not to stop smart people from debating ideas, but rather to reduce the friction that prevents them from executing those ideas.

It's necessary to mention here that if you have a team of A-plus players assembled, you reduce many of these problems from the start. To borrow Jim Collins' "Good to Great" lexicon, it'll be far easier to get the bus heading in the right direction if you already have the right people on the bus. However, in the real world of designing web applications, you sometimes inherit resources from vendors or partners that make the requirements of a simple design and development process even more important. This simple process is prototyping.

Once the priority of a project is established, the team should immediately move toward visualizing that idea. This can take many forms, but we have found that whiteboards and large pieces of paper work wonders to get everyone on the same page. Nothing slows down the creative process like a 60-page document, complete with spreadsheets and appendices.

Written specs tend to become a territorial battle for the authors because of the time invested in writing them. In a workshop on this subject that I attended not too long ago, an audience member revolted with the idea of no spec: "If we didn't write specs, then there would be an entire department on the street looking for jobs," he said. That is probably true, but it's no reason to continue writing long specs.

Advocating for the abolishment of a technical spec seems so against the grain, it warrants further explanation. When software was originally designed, it was without the need for a user interface. Early computers and Internet connections could barely deal with text and basic images, never mind complex user interfaces and scripts. It made perfect sense to write down everything in a spec before you started developing. This habit of writing long, detailed specs became so entrenched in the design and development cycle, we assumed it was the best way to do things.

Arguments for detailed specs include the need for writing "use cases." But neither the person writing the use cases nor the person using the proposed functionality is part of the development process. Disenfranchising the people from the design process is as bad as having politicians develop legislation for communities they've never visited. Use cases are extremely artificial. They lack the authenticity of the real world.

Prototypes are also energy efficient. Almost all the work that goes into a prototype gets integrated into the final product. Recycling work is one of the best ways to simplify development. Large spec documents are one of the biggest drains on the team's resources and energy.

Simplifying the design and development process does not translate to being simplistic. Just because a project strives to be simple does not mean it can be achieved without critical thinking or hard work. As challenging as a no-spec world might seem, it grounds the entire process in reality. Try a low spec diet -- and watch the fat disappear.

Richard Banfield, CEO of the web design studio Fresh Tilled Soil, is a serial entrepreneur with a passion for the web and the people that drive its innovation. He can be reached at richard@freshtilledsoil.com.