Email from Eric Loeb to Jon Garfunkel, August 9, 2006. Eric Loeb was a PhD student at MIT in 1992 when he volunteered to help the nascent Clinton-Gore Internet campaign. He later did the same for Senator Kerry's 1996 reelection campaign, and was the Chief Software Engineer for the Gore 2000 Presidential campaign. Today he is a political organizer in Philadelphia.

Hi Jon. Thank you for your note.

There were several, or even many, Clinton mailing lists in '92. The Marist list was a great one. Jock [Gill] was involved in all of those as the curator of the compuserve email address given to the Clinton campaign. I set up the MIT address (clinton-wins@ai.mit.edu). There was another big community group in Ohio... I forget the name. Natually the various Usenet groups and mailing lists were aware of each other and we cross-posted anouncements to them all to the extent we could. As I recall there was a volunteer in Chicago named Mary who took on that exact role.

I built the MIT address as a subject line processing system. It would respond to an unrecognized subject line with a canned email containing a list of valid subject line commands. It's main purpose was to email out Clinton position papers on various topics (there were about 25 of them). I built it precisely because when I met Jock (virtually) he was taking in emailed requests for position papers and mailing out floppy disks, via snailmail, containing the requested paper. In the process of building the position paper system I also added the ability to sign up for the clinton-volunteers email list, sign up for a discussion list, post suggestions & questions, ask for technical help, and so on. True to form -- and true to your theory -- I only consulted my own vision of what an online campaign system should do. I ended up with, basically, an email-based web site where one browsed the subject line commands one email/click at a time.

By August, as I recall, the Clinton lists had become prominent enough to attract the attention of the AI lab directors and the MIT lawyers. I proposed running them as a research project, with the same services being offered to the Bush & Perot campaigns. I needed a signature from the Bush campaign for that. I dogged some poor republican campaign staffer daily until I got that, and the PCIS -- Presidential Campaign Information System -- experiment was born.

Also around that time, John Mallory, a permanent grad student at the AI lab, got involved with the project. He convinced me to switch over to the LISP-based system he had been building. They were nearly ready to go and it was clear that they replicated the same functionality in a more robust fashion than the UNIX shell scripts I had thrown together on the fly. His system was almost ready, day after day, new exciting potential feature after potential feature, until three weeks before the election. During that time we brainstormed excitedly about the many possibilities of online campaigning and online governance. It was a lot of fun, but it would have been more useful to have cloned my scripts and move on to real campaigning.

Anyway, to make a boring story long, the LISP system that John built had a pretty long life as the internet systems for the transition team and then the Clinton White House. Some of my scripts survived as the USENET broadcast piece for them (and also later for Senator Kennedy's office). It was a great source of pride for me that the Clinton press conference transcripts made it out to the USENET world via my software for the 1st 4 years of the administration -- although obviously it was no great technical achievement. When we became aware of the web, John was very quick to show it to Jock, who showed it to Clinton-- who ordered all Federal agencies to create web portals before there were even 50,000 web pages in the world.