Furthermore, though I limited my analysis to just hypertext systems, since I
had sufficient evidence to substantiate this hypothesis, it seems to me that
the analysis easily extends to all types of Internet information systems.
Since adoption often equates to dollars on the Internet, analysis of
networks effects should be a critical aspect of many business plans.
Another avenue of future work is to analyze RPC-style systems for how they
do, or do not generate network effects. The Web generates network effects
through a growing corpus of information, yet this does not immediately
extend to cover RPC systems. It appears there is another mechanism at work
here. This has significance for the HTTP-NG effort, which is hoping to
generate network effects both via a corpus of documents, and also with RPC.