Matt's question is, to some degree, academic and as a result list members have made many valid points, some totally at odds with others. (Isn't that the point of academia? :) In practice, the questions are: What will structure do for *my* problems and what will it cost to implement?
I said the following recently in response to a similar question about XML over on techwr-l: For a long time my XML philosophy was, "If you need XML, you're already there." I've come off that belief as tools, best practices, and the like developed and evolved. That said, my current XML philosophy is "Don't deploy XML unless you have a clear set of goals and objectives and have a good understanding of the costs (effort, software, hardware, etc.) required to achieve those goals." In other words, XML for the sake of XML is a time and money pit. (My old philosophy was so much more catchy! ;( Find/replace XML with structure and you have my opinion on structure (less the catchy phrases). You better know why you're structuring your content and have a deep understanding of the obvious and hidden costs. Then it's just math ... is there ROI? -- Paul Nagai