In-Reply-To: <199504250237.AA07574@world.std.com>
Message-ID: 
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Status: O
X-Status: 



On Mon, 24 Apr 1995, Daniel P Dern wrote:

> Rex,
> 
> A fair amount of your "Internet history" doesn't gibe with what
> I know.  Before I go and debate you on stage, as it were, would
> you mind sharing with me a credential or two to help me assess
> where your info comes from?

I actually started working with internet back in 1982, as an employee for 
Computer Consoles Incorporated (CCI).  CCI was the leading producer of 
Directory Assistance Systems for the U.S., U.K., and Canada.  We were 
also one of the first commercial users of the UNIX operating system, and 
source code contributors/developers of the BSD operating system.  CCI 
consistantly allocated 20% of it's revenue to R&D, focused primarily on 
the development of the architecture now known as the Internet.

Because the company had almost gone bankrupt in 1974 when IBM switched
from BSC to SNA without warning (rendering their Cluster Controllers
useless) Herm Affel insisted that we always have access to the Source Code
of anything we developed from then on. Unfortunately, Roll your Own
was getting very expensive, costing upwards of $180 Million and 150 
staff-years to add 5 fields.  Several MIT graduates had been hired to do 
some hardware design and convinced management to try UNIX.  Since MIT
was releasing source code (Version 6) they tried it.  They bought/installed
version 7, conditional upon a source code license.  We sent AT&T fixes,
they sent us source code.  We agreed to do the 2ADA project (nationwide
directory assistance) until the Divestature.  As AT&T got less generous
with source code, we connected with Berkely, bought what was left of
Sperry, and put the Berkely Boys in Rochester and California.  Under
the GPL, we cooperated for over 8 years.  We got a 99.999% reliable system
that could be fixed from home at 3:00AM (8:00AM in England).

I was part of the SWAT team, called to put out the hottest fires with 
"quick and dirty" code.

To protect ourselves from possible liability, we used a ccivax!ritvax.edu 
connection for usenet and later NNTP.  We were also given anonymous 
logins (rexb@ccivax), and were advised not to use our full names or give 
company affiliation.

Some of the contributors from CCI included:
	Larry Early (Developed prototype RPCgen compiler)
	Marty McGowan (Contributor to rcs and imake prototypes)
	Bill Anderson (rcs, imake, and X11)
	Brad ???  (berkely unix domain sockets)
	Sifwat Ali (led team for rdbm, later created multi-media for Kodak)
	And many more.

Other contributions from CCI engineers included:
	LU6.2 boiler-plates.
	Anonymous Routing
	SMP (loosely and tightly coupled)
	Multiprocessing (CCI had a 24 CPU cage in 1974)
	Remote Procedure Calls (Called Virtual Processing Network)
	Distributed Computing Environment (VPN)
	RAID technology.
	Object Oriented Design (take a good look at a Unix Kernal Driver)
	Office Automation (OfficePower)
	Fault Tolerant Unix (predecessor to SMP) called PerPos.
	ISDN (Called BX.25)
	RISC Processing (6/32 or TAHOE architecture)
	SCSI Interface.

There were so many innovations, that British Telecomm Bought the company 
outright in 1987.  Many CCI Engineers "Graduated" to become pioneers in 
their own right.

I then went to work at Federal Express, they developed:
	The first network to "Kill" X.25/ISDN (ZapMail)
	The first PDA (the SuperTracker) (I received 3 awards)
	A comprehenseive survey of 12 LAN/WAN proposals, of which
	ONLY TCP/IP was capable of meeting the connectivity requirments.
	I was responsible for setting up the installations.  I also had
	quite a bit of direct contact with the folks at HP and DEC
	who lived in the area.  Although I left the company before the
	transition was complete, Federal Express ultimately converted
	the entire network to TCP/IP with generous UNIX.

In 1990, I worked for Great West Life where:

	We hired the Internet Consultants who eventually formed Colorado
	Supernet, as a "commercial internet"
	We were the first company in LOMA to develop a fully vertical
	"Open Systems" (Unix/TCP/IP) Insurance system.  Given that this 
	was the PENSION system, we were extra paranoid about hackers.
	I also pioneered a number of integration tools that allowed MVS
	 programs to pump entire databases through UNIX filters.  I led
	the development of RPC over APPC connections, and the development of
	"MiddleWare".

	Jim Rosborough (My boss) chaired the LOMA infrastructure committee
	and guided it (with help from a few little birds) through the
	from proprietary (IBM/SNA) systems to Open Systems.  This 
	included efforts lead by Jim Traeber to add secure authentication
	systems et. al. to SUN tools.  Through our consultants, we placed 
	this software under FSF-GPL.

In 1993, I lead the development of an "Internet for Idiots"
	project at SoftTronics.  I was also responsible for the the Internet
	"firewall/server" connected to CSN, and provided demonstrations
	commercial internet.  The product "softwin" or SoftWindows,
	was one of the earliest share-ware SLIP packages on the internet.
	The Colorado Supernet was actually a T1 connected from CU 
	Boulder to UC Colorado Springs, which connected to Salt Lake City
	through links provided by IBM(Boulder), HP(Colorado Springs) and
	to MASS via DEC (Colorado Springs).  No accident that MCI located
	in Colorado Springs, it was one of the "HUBs".

	While in a seminar in Denver, I chanced to meet the VP of 
	marketing responsible for sale of "Frame Relay" connectivity at
	MCI.  During an extended conversation with him, I created the
	possibility that the Internet Protocol, and providing secure access
	to the NSFNet would attract customers who could quickly be 
	connected via TCP/IP routers to Unix Systems and where necessary,
	to SNA gateways.  Two weeks later, he told me that not only did
	his boss like the idea, but that MCI offered to Carry NSFnet traffic
	on condition that it could use the same links for commercial traffic.
	CSN provided the Prototype strategy for PSI and CIX.

Due to economic shortfalls (revenue on another product hadn't 
materialized)  I proceded on to IBM/Kingston where I ported X11R5 to 
AIX and made infrastructure suggestions for AIX/MVS connectivity.
In 1994, I joined Dow Jones, and championed the use of TCP/IP and
UNIX/WEB technology.  When Greg Gerdy (marketing manager of DowVision)
said this "little start-up" called WAIS wanted to do business with us
again (the principle players had helped DJ develope DowQuest for NRS)
I stepped in and said that was a "Must Do" project.  I was given
very limited resources to support their effort, and they came through 
brilliantly with a WAIS/HTTP server, in about 3 months.  By the time the 
rest of the company realized that the internet had grown from 2million to
16 million in less than 12 months, Greg announced his coup.  He was a 
corporate Hero and I smiled quietly from the wings.

I was asked to lead a team to develop a protocol, and after 6 months of
reccomending TCP/IP, SGML/HTML and being told, by my boss, that we
couldn't use any of that "College boy software", the company accepted his
offer to work for Microsoft and moved aggressively into TCP/IP and Unix.

We also put lots of work into the MSN project.  Eventually, after 4 staff 
and 6 contractors were "Held at Bay" for almost 2 years due to delays in 
the delivery of software and tool-kits, we gave up and created a Lotus 
Notes product.  I know lot's more than I'm telling, but I'm bound by
nondisclosures.

Because I walked in the land of Giants, I looked a bit like a dwarf.
I chose to remain relatively anonymous for personal reasons.

> Me, I worked at BBN from 1983-1989, got to chat with a fair number
> of the folks there and read LOTS of related docs (I was doing tech
> doc on the DDN, then PR in general, which included writing about
> the ARPANET/DDN/MILNET/Internet relationships), have been writing
> about the Internet/NSFnet and commercialization activities
> since then, etc.  Some of what you say doesn't square with what
> I know.  I'm still happy to disagree with you in online-news,
> and find out which are the facts, but I thought I'd do a reality
> check first.
> DPD



From rballard@cnj.digex.net Thu May  4 02:33:23 1995