[3-51] See Robert M. Fano, "On the Social Role of Computer Communications," _Pro-_
_ceedings_of_the_IEEE_ 60 (September 1972): 1249.

[3-52] Berners-Lee, 46. See also James Gillies, _How_the_Web_Was_Born:_The_Story_of_the_
_World_Wide_Web_ (Oxford and New York: Oxford University Press, 2000); Hafner and
Lyon, _Internet_Dreams:_Archetypes,_Myths,_and_Metaphors,_ Mark J. Stefik and Vinton G.
Cerf, eds. (Cambridge, Mass.: MIT Press, 1997).

[3-53] Berners-Lee, 40 (describing Gopher and WAIS growing faster).

[3-54] Ibid. (interconnect).

[3-55] Ibid., 72-73.

[3-56] Ibid.

[3-57] Ibid., 74.

[3-58] Ibid., 99.

[3-59] See Barbara Esbin, "Internet over Cable: Defining the Future in Terms of the Past,"
_Communications_Law_Conspectus_ (Winter 1999): 37, 46 ("ARPA began to support the
development of communications protocols for transferring data and electronic mail [e-
mail] between different types of computer networks."); Needham J. Boddie II et al., "A
Review of Copyright and the Internet," _Campbell_Law_Review_ 20 (1998): 193, 196
("Funding for the Internet comes from five federal agencies, various universities and
states, and private companies such as IBM and MCI.").

[3-60] Naughton, 83-85.

[3-61] Ibid., 84.

[3-62] See, e.g., http://www.asiapoint.net/insight/asia/countries/myanmar/my_spedev.htm.

[3-63] Among other restrictions, AT&T was not permitted to get into the computer busi-
ness. This fact becomes quite important in explaining the birth of Unix. For a compre-
hensive and balanced account of the effect of these limitations, see Bickerstaff, 14-17.

[3-64] Bickerstaff, 25.

[3-65] For more on the history of the FCC's Computer I and Computer II decrees, see
ibid., 1-37; Huber, Kellogg, and Thorne, #5.4.

[3-66] National Research Council, 130-131, n. 18 (describing best efforts as consequences
of uniformity).

[3-67] "Humans can tolerate about 250 msec of latency before it has a noticeable effect,"
http://www.dialogic.com/solution/Internet/4070Web.htm.

[3-68] They are described in Mark Gaynor et al., "Theory of Service Architecture: How to
Encourage Innovation of Services in Networks" (working paper, 2000, on file with au-
thor), 14.

[3-69] Indeed, George Gilder believes the increasing capacity of optical fiber will render
moot this debate about QoS. "Today on the Internet, the consensus claims that QoS will
be indispensable for voice and video. But with true bandwidth abundance, QoS com-
plexities are irrelevant -- an ATM tax imposed on the vast bandwidth of fiber with its 10
to the minus 15 error rates, far better than the reliability of telephone circuits." Gilder,
_Telecosm,_ 80.

This point is made expressly with respect to quality of service and end-to-end in an im-
portant paper by Barbara van Schewick, "The End-to-End Principle in Network Design,
Its Impact on Innovation and Competition in the Internet, and Its Future in the Net-
work Architecture of the Next Generation Internet" (working paper, September 2000,
on file with author). Van Schewick distinguishes between two forms of QoS --
integrated services architecture and differentiated services architecture (6, 7) -- and ar-
gues that the former puts more pressure on the end-to-end principle than the latter. She

[[277]]

p276 _ -chap- _ toc-1 _ p277w _ toc-2 _ +chap+ _ p278

v?