Beautiful Architecture

Data-Driven Applications

Once an organization has gone to the trouble of making its data addressable, there are additional benefits beyond enabling the backend systems to cache results and migrate to new technologies in unobtrusive ways. Specifically, we can introduce entirely new classes of data-driven applications and integration strategies. When we can name our data and ask for it in application-friendly ways, we facilitate a level of exploration, business intelligence, and knowledge management that will make most analysts drool when they see it. The Simile Project,^[24] a joint effort between the W3C and the MIT CSAIL group, has produced a tremendous body of work demonstrating these ideas and how much drool can actually be produced.

Consider the scenario of tracking the efficacy of various marketing strategies on website traffic and sales. We might need to pull information in from a spreadsheet, a database, and several log files or reports from web analytics software. Although tying these things together now is not exactly rocket science, it does require a nontrivial level of effort to find, request, convert, and republish the results. If we simply produce a spreadsheet summary and email it around, we effectively lose the ability to retrieve the results at some future point without searching our already clogged inboxes. Adopting a CMS or other document management system that we can link to will increase the amount of time necessary to produce the result. Whatever the frequency ...