Re: Thread 5: Defining the namespace

[Michael Dillon <michael@memra.com>]

On Thu, 26 Dec 1996, Simon Higgs wrote:

> > The namespace cannot be *DIVIDED* up in a meaningful way.

> Of course it can.
> 
> > Your statement
> > implies that the namespace is some sort of unitary thing and the only
> > problem is to find where the natural divisions lay.
> >
> 
> Uh... it's called library science. 

And I am saying that the namespace is not something that can be carved
up by library science.

> What do you
> think Yahoo, Lycos, Alta-Vista etc. do?

AltaVista doesn't classify anything at all. It's purely a text searching
tool. And Yahoo does not use library science to categorize things but
instead often creates sub-categories that are in multiple larger
categories. A single book needs to be in a single location on a single
shelf. There can be only one right classification for a book. This is a
problem well solved by library science. However the namespace is not 
a set of books, it is more like a set of words which, like many natural
words, have multiple meanings. Even professional librarians understand
that a single book may be referenced under several subject entries in
their card catalogues.

> Why was .COM and .EDU created to distinguish between commercial and
> educational organizations? What should be happening now is just further
> division, focus, and refinement of the name space.

.COM and .EDU were created by people who didn't fully understand the
implications of attempting to carve up the namespace. This can be clearly
seen when you understand that .COM contains registrations for educational
institions, non-commercial charitable organizations, military
organizations, etc. In effect, .COM has become synonymous with COMMON
not with COMMERCIAL.

> > But it's not like that at all. The IAHC will not be dividing up anything.
> > They will be CREATING a namespace. And if they do a good job there will
> > be significant overlap between the gTLD's in the sense that people will
> > not feel restricted to only one possible gTLD choice.

> The name space exists hierarchically under ".". Any new name space
> created will be a new division under root. This is just semantics.

Semantics are absolutely fundamental. If they were not, then the IAHC
would just create .X001, .X002 and so on, allocating them on a first-com
first-served basis to whoever applies. 

Michael Dillon                   -               Internet & ISP Consulting
Memra Software Inc.              -                  Fax: +1-604-546-3049
http://www.memra.com             -               E-mail: michael@memra.com