Tom Erickson worked in the HIG group.
I remember I was chatting with a colleague when Joy (Mountford) came by and said, “There's this guy from Thinking Machines in the conference room, and he wants to talk to people about information retrieval. He wants to know what we are doing. "Why don't you go talk to him and see what he's up to?"And that was Brewster (Kahle).
Rosebud project lead Charlie Bedard recalls:
I had spoken to David Nagel about building search capability not just for our own Apple library but eventually for every library. As an ex-NASA engineer, he appreciated that we didn’t have the computing power on-site to go to the next level so I went looking for someone who had the horsepower to suck in all the raw data and compute an index. And, of course, nowadays, that's commonplace. Google has automated web crawlers that do all of that, but in 1988/89, there wasn't anything like that. So I reached out to Brewster Kahle, who was, and still is, a digital librarian, computer engineer, entrepreneur, and advocate of universal access to all knowledge and subsequent founder of the Internet Archive.
Erickson recalls, with a smile, decades later:
I walked into this conference room, and there was Charlie (Bedard), some others from ATG, and this guy, Brewster Kahle, sitting at the head of the table. And it was Halloween, and he had been told that people at Apple dressed up in costumes. Brewster had dressed up as a rooster, in a rooster costume, complete with a comb on the top of his head. And he was a bit, how can I say it? A bit chagrined to learn that there were areas of Apple where people did dress up, but our group was just wearing typical shirts and jeans and sandals. In the end, we managed to put all that aside, and Brewster started talking and got very excited.
After studying Artificial Intelligence at MIT , Kahle joined Thinking Machines, where he was the lead engineer for six years. Erickson continues:
He had a real vision of machines being connected over a network and a user being able to retrieve information from any of them. He had developed a system architecture involving the Z39.50 protocol.
Z39.50 is an international standard client-server communications protocol for searching and retrieving information from a database over a TCP/IP computer network.
He was putting together a consortium (WAIS) including his employer, Thinking Machines, and the idea was to run a big query retrieval engine. He knew that he needed an information source to test and refine his approach, so he had already forged a connection with the publishing firm, Dow Jones, who in turn offered the DowQuest information system to its clients, who were mainly stockbrokers.
DowQuest gave users access to over 350 news sources from the previous six months, with a full-text retrieval mechanism based on relevance feedback.
If you were searching for results, you'd say, “Oh, this is the right thing, find more like this, you know”. This was all interesting to us at HIG and ATG, as one of our concerns as we researched databases was the difficulty in retrieving results. Brewster believed that you needed a client to design for, so you didn't research indefinitely, so he had made a connection with the accounting company KPMG Peat Marwick. At the time interface design was in its infancy and Brewster was a pretty hardcore computer science hardware guy, so I don't think he believed yet in the value of design as such, however, what he wanted to do was something that HIG and ATG were exploring at length, and it was a great high-level conversation.
Months later, Apple was convinced to join the consortium, but at the time, 1989!, Brewster was in our boardroom talking about what we now know as Google search and OpenAI. I just remember watching the rooster comb on top of his head, flapping in time with his hands as he spoke.
