In 1991, one of the first consulting jobs I had was for Chipcom Corporation. I was working with a consulting firm out of Southborough, Mass at the time and they wanted a way to distribution information to their resale chain around the world. The world wide web we think of today wasn't an option in those days. We built a solution using a Lotus Notes server, with an "Octopus" 8 port serial card connected to a bank of modems. The modems allowed partners to connect and replicate using their Lotus Notes version 2 client by dial-up. Partners in Europe, South, and Central America could connect to local telephone numbers and be routed back to the server over an x.25 network. We distributed the initial set of data on 3.5" diskettes in zip format. It was amazing that it worked, but it did work. The solution was called "OnCall". I didn't design OnCall. I just made it work.

Chipcom primarily made token ring equipment, and with the eventual dominance of ethernet, they were bought out by much larger rival 3Com Corporation. My work on OnCall got me a seat at the planning table for a new reseller Portal to be built for Web Browser access by 3Com. The original plan was a big J2EE server based solution but for a variety of reasons it didn't come to pass. It gave me the opportunity to design and build a "temporary" solution called "Partner Access" using Lotus Notes. I didn't design the look and feel of "Partner Access", and I didn't create the content. I just made it work. There have been professional design people to come up with good looking layouts and at least one other pretty good web developer who's added innovation and function to the tools we use. My focus with Partner Access was always the on technical architecture. Initially, the iNotes Web Publisher was used to export Notes documents to html files on disk, but that was quickly replaced with the add-in task for the Notes server code named "Domino". That task became part of the server in version 4.6 and eventually became the name of the server.

The Partner Access project followed me through two employers and eventually my own consulting firm. The project management was originally 3Com itself, and later was outsourced to a former 3Com Employee who started her own firm. I was working on Partner Access the week I announced the birth of my oldest daughter. She's nearly 17 now and was driving me some place when I last had a conference call about Partner Access. Laurie, the former 3Com person who has had her own firm for many years now was on the email list when Ari was born and we're still working together on other projects. Partner Access was recognized twice by IBM for Beacon awards - in 1999 as a finalist for "Best Web Solution" and in 2003 Laurie and I won for "Excellence in Partnering".

A few months back, 3Com was bought out by HP and the size difference between the two firms has pretty much meant that everything 3Com is quickly becoming purely HP. As of yesterday the Partner Access web site was taken down and is no longer available. The site lived through 17 years, sever version upgrades from 4.6 through 7.03. It was moved to new data centers twice -- once across country to new headquarters. The source of group membership and security profile data was changed more than 10 times over the years. More than 150,000 people accessed the site in several languages. The total amount of downtime spread across the years can be measured in just a few hours.

R.I.P. Partner Access -- and welcome what's new. On the same day that Partner Access went away, Laurie and I launched a new portal site for the Community College of Denver. ( http://www.ccd.edu ). It's not as big as Partner Access grew to become, but it's far larger than when that site started. We've had other projects over the years of course, including some surprising "household name" companies. Most of them are private portal sites so I can't share them. Still, there was something interesting about the combined shutdown of PA with the launch of the new CCD last night.

Anyone who tells me that pure "old school" Domino based development doesn't scale, or can't produce compelling sites can eat my shorts -- while we go on producing just those sorts of solutions at a tenth the cost of over complex, self important, "more scaleable" server tools.