[336] in Commercialization & Privatization of the Internet
Re: A few questions re current discussions...
daemon@ATHENA.MIT.EDU (Jack Haverty)
Mon Mar 11 12:11:46 1991
From: Jack Haverty <jhaverty@us.oracle.com>
To: Stephen Wolff <steve@cise.nsf.gov>
Cc: Jack Haverty <jhaverty@us.oracle.com>, com-priv@psi.com,
In-Reply-To: Your message of Fri, 08 Mar 91 14:03:37 -0500.
Date: Mon, 11 Mar 91 08:56:06 PST
"... Don't you regard X Windows, NFS, and Z39.50 as new applications? I
sure do! And how about X.400(92) and X.500 - they sure (will) allow users
to do things they've never been able to do before...."
No, I don't think these are new applications. But maybe we have different
definitions of "application". To me, an application is a capability which
supports end-users' tasks in their job or personal activity, using computing and
communication technology to do so. And to be a real test, I'd like to require
that the end-users be involved in something not related per se to computing or
communications - i.e., not programmers, or network managers or in general people
working in the "networking/distributed computing" field.
And when I say "on the Internet", I mean something that has become part of the
*definition* of the Internet and available pervasively - i.e., part of the
infrastructure. These are capabilities that one would expect "The Internet" to
provide, support, and manage so that they are reliable enough to be considered
infrastructure.
Some of these "capabilities" may be high-level services - e.g., if there were a
pervasive NIC-like facility which was a library/repository for some large
fraction of academic literature accessible through the Internet using Z39.50,
I'd call that an application. But Z39.50 by itself is a technology.
If there were some service available using EDI, for example to order reports
from this hypothetical library, that would qualify. If there were some service
that provided a CASE environment for use in coordinating software projects
distributed geographically and organizationally, that would be in my list.
Other "capabilities" could be enabling technologies because they extend the
technical base. The DataComputer was like that. An Internet File Service,
which uses NFS (actually I don't think NFS is up to this) and provides a reliable
repository with tools for indexing, browsing, searching et al (NLS-of-the-90s)
would replace the "anonymousFTPserver" approach. But NFS itself is a technology.
Things like X.400 and X.500 are neither new nor applications. They are the
latest "release" of what began as the NIC WHOIS protocol and the original mail
protocol. They are important for the mail application, which has now evolved
from the small scale of the 70s to the global scale of the 90s and beyond.
Important stuff, but really targetted toward handling the scaling of an old
application rather than providing a new one (yet at least).
Similarly, X-windows (which I like a lot) is just the "latest release" of our
old friend Telnet - it provides a way for a human to interact with a remote
computer - and since the technology of computer interfaces has progressed from
Mod 33 ttys to screen/graphics, the "telnet" application must track that change.
More like sustaining engineering of the old application (remote access) than a
new application.
Anyway, you get the idea. Maybe I just have a different view of The Internet
(or NREN) as a distributed computing infrastructure, rather than a
communications infrastructure, and computers come with applications - the
interesting question being which applications are so useful that they should be
adopted as part of the definition of the infrastructure.
The Arpanet was created in 1970 or so to interconnect computers, and three
applications popped up (Remote Access, File Transfer, Electronic Mail) which
became so useful that they were defined to be part of the Internet (i.e., when
TCP was replacing NCP, Telnet, FTP, and Mail were explicitly, and painfully,
migrated.
As you point out, when the constituency widens to larger and larger audiences,
new applications pop out of the woodwork.
So, to make a long story longer, the question I have is "What applications have
surfaced during the Internet Era which have become so obviously useful that they
are now part of the definition of "The Internet", and should be part of the
foundation *provided* as part of the next era - the NREN? Or in a
commercialization of the Internet?"
Jack
