Subject: on the horizon -- broadband filesystems v. distributions
From: Tom Lord <lord@emf.net>
Date: Tue, 29 Oct 2002 11:02:39 -0800 (PST)


What impact will broadband have on distributions?  or on the P.C., for
that matter?  Surely it will be larger than "I can get my ISO images
20 times faster."  Any bets on whether or not broadband deployment
becomes the new WPA?

Personally, I think file systems like AFS or Coda will change what
people buy and how they maintain their systems -- it'll be a big
economic shift.  It'll fix a lot of problems and there's a
decades-long history of this being a low-cost and featureful approach
to supporting lots and lots of seats.  Buying a computer will be 
a lot more like buying a telephone.

But: 

	* Will there be mom 'n pop providers or will it be just
	  cable-co's and AOL?  

	* Will the global filesystem namespace be territorial in the
	  manner of domain names?  Or will Plan-9ish directory views
	  bring namespace relativity to the rescue?  

	* Will we see a short-circuit of the path from maintainer
	  sources to user binaries when projects can get their own
	  user-visible $X/bin dirs?  In other words, what happens when
	  deployment to users becomes a matter of saying "Add this
	  directory to your $PATH".

	* What'll happen to the server market?  Whither the unix
          vendors?  

	* Will there be booths in every airport where, instead of 
	  just a fax machine and phone-line, there's a little client
	  you can use to log in to your universal account?  Or better: 
	  on every train?

	* How and how well will customer-specific customization be
          supported?

	* Will some universities, who have managed this kind of
          environment for quite a while, cut out the middleman and
	  go back to taking the R&D profit directly?

	* In short: What work will be left for publicly traded FSBs to
          do?


Conversation, contradiction, criticism,
-t