Journalism accuracy (for tech especially)

[Dave Farber <dave () farber net>]

------ Forwarded Message
From: Dave Burstein <dave () dslprime com>
Date: Sat, 26 Apr 2003 17:50:59 -0400
To: Dave Farber <dave () farber net>
Subject: Journalism accuracy (for tech especially)


> Dave

The news article you posted about the "selfish router problem" overplayed
what the Cornell researchers had demonstrated. I suspect the researcher's
original paper was stretching, but not unreasonable. To me, the problems
you were pointing to were bad journalism, not necessarily bad science.

   One of the first rules of journalism is to have a diversity of sources.
The reporter apparently relied only on the Cornell people, who  may well
have been over-enthusiastic about their results.

      Compare instead how Jonathan Krim of the Washington Post approached
the broadband debate and the possibilities of really fast networks. Besides
the usual interested parties, he interviewed a relevant academic (John
Cioffi of Stanford), engineers in the field, and independents (me, I
suspect among others.)

      I don't want to cast too many stones at a single article by a
generally excellent journalist. Rarely does a reporter have the time to do
everything the textbook recommends. Diverse sources is only one test, and
absolutely isn't sufficient. Consider the usual "one for, one against" type
article that doesn't give you enough info to know who's right. But it helps
considerably in a case like this, where a disinterested expert would
probably have steered the reporter away from what is obviously a
misunderstood report.

      My experience is that experts, even top academics, usually are happy
to give the time to a reporter trying to get a story right.

----------------------
Separate note on the original thoughts and Dave Reed's comments on
End-to-end.

     The largest networks in the world are constantly balancing spending
today's limited budgets on increased bandwidth (Dave Isenberg's "fat, dumb
pipes") or quality of service/traffic shaping. Contrary to rhetoric,
there's need for QOS improvements at many points. You don't want your VOIP
phone call degrading badly when your partner downloads a movie in the next
room. Cost of QOS improvements is going down with Moore's law and will
often make sense - it's about $10/user to add it to a current generation
DSLAM, for example, and will soon be less than $5 in silicon to add to a
broadband modem. The U.S. telcos have a strong team at the DSL Forum
driving forward that work.

      The same telcos, however, are not provisioning enough bandwidth for
the growth we all hope to see in internet use. (fat pipes) I see this as a
likely $B mistake on their part, and devastating to growth of technology.
The Third Internet is ready, fast enough to watch. U.S. telcos are building
networks around 20K or 30K per user while the Koreans and now Japanese are
proving the demand can explode and provisioning far more bandwidth. The
result: U.S. DSL networks will be hobbled if many users want full screen
video, which requires rates from 500K to 2 mbps. Mark Coblitz of Comcast
envisions carrying multiple 5 mbps HD video data streams, which should be
scaring the telcos.

      Larry Lessig's "Code", a great book if hard to read, makes really
clear how technical decisions (software design, network provisioning rules)
often determine crucial realities. For me, the ability of the internet to
deliver everyone's video is a crucial freedom of speech issue going
forward, and also smart business. Whether your desire is a Mormon Sunday
service, movies in URDU or Italian, Gaelic football or parliamentary
debates, $3,000 video cameras and inexpensive Macintoshes make the
production cost low. So I believe strongly in keeping internet channels
broad and open, e2e.

Dave Burstein

Editor, DSL Prime & Telecom Insider
Co-author DSL: A Wiley Tech Brief (February 2002), Journalism in the
Internet Age (forthcoming)
Special correspondent, The Personal Computer Show, WBAI-99.5FM, Three time
winner of Best Radio Show ,Computer Press Association



------ End of Forwarded Message

-------------------------------------
You are subscribed as interesting-people () lists elistx com
To manage your subscription, go to
  http://v2.listbox.com/member/?listname=ip

Archives at: http://www.interesting-people.org/archives/interesting-people/