Hindery Fatwa on Portals
Originally uploaded by lyndon.
After the final demise of ExciteAtHome in 2002, I reflected a bit on my experience with the Internet Bubble, and among other things, read "The Internet's Coming of Age" a 2001 publication of the U.S. National Academies drafted by a committee including ExciteAtHome CTO Milo Medin and future Google CEO Eric Schmidt. This monograph outlined how the architects of the Internet foresaw the relationships between content (data), services (software applications) and distribution (network connectivity). One key attribute of the Internet architecture is the separation between the network layer and the application layer, deliberately putting the 'intelligence' of new services at the edges of the network (where servers executed software), and keeping the pipe 'dumb' (see CSTB 2001). The architects expected this would make innovation on the Internet easier, by eliminating any network dependencies when deploying a new application. If Internet application developers seem to be eating the lunch of the cable and telecom companies, it's in some respects a natural consequence of how the Internet was designed to work. Had the arrangements been different, the concept of "Internet-speed development" would not have arisen, and today's Web would have evolved for the worse.
Because each of these three legs - content, applications and distribution - requires substantial investment in capital, human or otherwise, I find rather comical the recent remarks by former @Home board member Leo Hindery on the impending death of the major Web portals. The economic role of portals is more accurately reflected by the Web applications they host than by the content they present. The expertise required to conceive of and deploy these software applications is quite distinct from the talent to create content (merely data to a software developer) or the logistical muscle to build physical networks (all those truck rolls for installing broadband access everywhere). The disparate nature of these talents contributed to the failure of convergence via vertical integration at AOL-TimeWarner and ExciteAtHome.
But there's no need to belabor the point further. The reaction of the blog-o-sphere (e.g. TechDirt, John Battelle) makes better reading. Twenty years from now, we'll all share a good laugh, because these bits will still be around for everyone to read with bemusement. In the meantime, let us hope that only a minimal amount of investment capital will get mis-allocated along the way.
Tags: convergence, lyndonwong