Fewer NAs than DNS Hosts

Although the assumption that there will be many fewer NAs than DNS hosts is 
unprovable without the assistance of a time machine :), it would appear 
unlikely to me.

One vision of the future would have all named, Internet-accessible data 
possessing a URN.  Once the Web becomes a read-write medium, in this vision, 
not only would many (if not most) people produce works requiring the use of 
a URN, it could well be possible that many (if not most) people would set up 
their own NA (or multiple NAs) for their own works.  I see this in contrast 
to the current world of hardcopy publishing, as the current Web directions 
would seem to point to a world where many people "publish" directly; i.e. 
placing a document on the Web as something closer to shareware, except that 
Web mechanisms could assist in enforcing payment.

Part of the whole "re-engineering thang" is the elimination of 
middle-people, whether distributors, managers, or a select cadre of NA 
providers.  This (IMHO) is part of a whole trend towards flattening the 
hierarchies that served humanity during the periods of mass populations with 
low literacy rates.  I don't see a lot of people saying, once Web publishing 
becomes easier ("draft-ietf-html-fileupload-02.txt": are you listening, 
Netscape?), "Gee, I think I'll send all my documents to Fisher, Andropov, 
Kwanzaa, and Chin Publishing Ltd. Corp. to be placed on the Web!".  I see 
more publishers, NAs, etc., etc. in the future, not just a few.

The engineering of such distributed systems is a hard problem, but is being 
tackled right now (witness DCE among other systems).

We are going to have to handle the increasing decentralization of life.
======================================================================
Mark Fisher                            Thomson Consumer Electronics
fisherm@indy.tce.com                   Indianapolis, IN

Received on Tuesday, 20 June 1995 12:32:11 UTC