Back in 1996, when I first began applying semantic technology to the Web, my work was mainly motivated by two reasons. The first one was more superficial and focused on dissemination, while the second one was more about a matter of principles. Applying semantics to the web has proven to be quite a success in terms of acquired public funding, public awareness, scientific events, and an increasing number of commercial activities. Such a quick success and a fast growing community around a specific topic also carries a certain risk of the kind associated with any stampede. The first two or three bulls may know why they are heading in a particular direction; many of the others may follow with only a vague knowledge of these reasons based on lingering rumors they catch from their neighbors. Therefore, it may be worthwhile to reflect on the two reasons that initially drove us in the direction of the semantic web. In doing so, we will also reveal why the more superficial one is the more visible one, guiding a majority of current approaches, and which may nevertheless ultimately prove to be a dead end.