Could it be that the semantic standardization fundamental to web services reverses the roles of the institutions contrasted in the title of Raymonds' noted essay, where he compares the design of a splendid communal edifice with the chaotic but productive kludgery conducted across the commons, knee deep in squalor?
Tim seems to be concerned with the establishment of proprietary models for the back office provision of web services, but the essence of web service functions is their accessability and use out on the commons, on the net. The net is the OS, and the APIs to deep data bases provided by some vendors whose activities broadly correspond to areas of interest do not bestow any divine rights as essential resources (the broadest possible search, per se, being the unique and crucial function, at this stage).
(Maybe ten years ago I got a cattle-call email when ORA was making the great portage from Boston to Berkeley in which Tim introduced himself as a classics graduate, so I should make it clear in what follows that I have "no latin and less greek".)
The ideal of the semantic web can be boiled down, as Jon Bosak has put it citing Wittgensteins' famous line, "the meaning is the use", to a