[306] in Commercialization & Privatization of the Internet

home help back first fref pref prev next nref lref last post

Re: one commercial site's interpretation of NSFNET restrictions

daemon@ATHENA.MIT.EDU (Bill Nowicki)
Tue Mar 5 17:00:47 1991

Date: Tue, 5 Mar 91 13:33:41 PST
From: nowicki@Legato.COM (Bill Nowicki)
To: bob@morningstar.com
Cc: com-priv@uu.psi.com

	Date: Tue, 5 Mar 91 11:50:29 -0500
	From: Bob Sutterfield <bob@morningstar.com>

	Sun has been distributing fixes via UUNET for several years,....

I was the one who started this practice at my former employer. Sun had
an interesting saying: "to ask permission is to seek denial".  This
came from the cynical view that if you ever ask a lawyer "Can I do
this?"  they will say "No." So avoid asking lawyers, but do things
that you trust are "right". I was so frustrated with Sun's slow
release process that I felt it was "right" to use the network, and
this was consistent with the spirit of the vague restrictions. I had
no "official" blessing.

The choice of UUNET was pragmatic, since Rick Adams happened to be at
the SUG conference when we talked about it. The vast majority of Sun's
(and UUNET's) customers were then NOT on the Internet, so UUCP was
presumably the most common method of obtaining the fixes, with FTP for
the small group of major US University C.S.  departments on the
Internet. Since then we have made some progress, with medium-sized
Universities now also on the Internet. :-)

The majority of TCP/IP users are still NOT on the Internet, which
might explain why LISTSERV-like services are still so popular.

As for the broader question of "Why no new application protocols?" in
the last ten years: In my opinion there have been only two important
new application protocols in the last ten years: NFS and the X Window
System. Another disclaimer: I have been involved with both of these to
some extent. It is interesting that although both of these did come
from the "Internet community" in the largest sense, neither one is an
IAB "standard". Despite useful tools like RPC compilers, the time
scale of an important new protocol every 3-5 years is about right.

Our company is working on other network applications, like
backup/restore and resource location/administration. But if I went
into details imagine the deluge of flames about improper use...

	-- Bill Nowicki
	   Legato Systems

home help back first fref pref prev next nref lref last post