This past Sunday’s New York Times included an article about the looming confrontation between Google and Microsoft. The section of the article that described Google’s software development method caught my eye.
- New features and improvements are made and tested on Google’s computers and constantly sprinkled into the services users tap into online. In the last two months alone, eight new features or improvements have been added to Google’s e-mail system, Gmail, including a tweak to improve the processing speed and code to simplify the handling of e-mail on mobile phones. A similar number of enhancements have been made in the last two months to Google’s online spreadsheet, word processing and presentation software.
The rapid development software model is not new, but Google has implemented it to perfection. I would like to think that we have adopted some aspects of this model at school. Sometimes, we quickly develop a custom web script, install an open-source application, or adopt a new IT policy after a short conversation and see how well it plays in the field. At other times, we adopt a more conventional approach, quietly developing and testing an idea until it is fully mature before throwing it out to the community.
To some extent, the choice of which strategy to employ depends on the centrality of the system. When rolling out a new wireless security scheme or file server, much testing and gathering of feedback is required. When creating a new opportunity for teachers to post video on the web, one may proceed with abandon. Even Google appears to modify its core applications only to introduce new features in a test environment. Recently, Facebook took a big hit from failing to anticipate how users would react to yet another big feature change that impinged on their privacy.
We can stand to remember that the best feedback on an innovation is gained from everyday users giving it a try. Get user feedback early rather than working in isolation for long periods of time. Maybe we can benefit from more often adopting such a “perpetual beta” model.