Sunday, October 22, 2006

I think one of the principle reasons there were so many people in the last 10 years
trying to do new network architectures is that there is a seriously broken cult of personality in
US research, and folks like Clark, Cheriton, Shenker, Cerf, etc don't do enough to dispell it
(actually Clark is probably the worst victim without being so guilty, but he could do more to undermine it) -

The fact is, like any Big Breakthroughs, the Internet was the product of a lot of peoples' efforts and most of the
papers people point at (e.g. Cerf&Kahn, or Clark's SIGCOMM 88 paper) were post hoc rationalisations at best
of stuff everyone working on stuff knew at the time but didn't have time to write a polished rant about.

The reason this is a problem is that it distracts people from the work that needs doing in exploring design space
and makes them write these "high level" land grabs, without much n the way of foundations. This is another reason I
think the FIND programme is misguided (and doomed:)

As I'll said in my CoNEXT talk, not only is architecture hard to do, most so-called architects didn't actually do
what they claimed anyway (thats the difference with people in the real world like folks that do processor
architecture&buildings, and network architects - there's tangible evidence the Power PC and the
Millau Viaduct are cool, as well as usable:)

oh, and there's also processors that fail & buildings that fall down:)
thats how we really learn what works, and what doesn't!

btw, one of the coolest websites for architecture (as in Le courbousier) is foster associates - see
real architecture and weep!

No comments:

Blog Archive

About Me

My photo
misery me, there is a floccipaucinihilipilification (*) of chronsynclastic infundibuli in these parts and I must therefore refer you to frank zappa instead, and go home