An Untold Story
The first search engine, according to all known sources, was called Archie. It was developed by Peter Deutsch, Alan Emtage and Bill Heelan in the early 1990s. This is the answer you should give if you are asked this question in an exam paper, or a pub quiz, and want your answer to be marked as correct.
However, the first search engine was invented in the second half of 1985, became the subject of a collaboration and confidentiality agreement with Hewlett Packard in 1986, and then slipped into oblivion in 1987.
This is the untold story of the first search engine.
A Small Development Team
In mid-1985 I was appointed End User Computing Strategy Manager for, what was then called, Royal Insurance Group. This sounds a grand title, but each of the major operating companies had their own autonomous IT department. The title should really have been 'End User Computing Strategy Manager - for those parts of the group that aren't big enough to have their own IT department'. This wouldn't fit on the business card, however, and I probably wouldn't have liked it as a job title.
At the time I worked for Derek Shute (I forget his exact title, but it was something like Policy and Strategy Manager for the Group) who in turn worked for Mike Sayers (the Group IT Manager). Derek and Mike both nurtured much of my early career and taught me a great deal. In retrospect, they took quite a risk by propelling me into positions of significant responsibility at an early age (I was in my mid 20s).
My team was small. Steve Fletcher looked after User Education or training. He was someone who was born to be - and now is - a Managing Director and he expanded an internal training operation into a very successful external training business. Gary Lewis, Yvonne Green and Jackie Crookes, the other members of the team, were all highly capable individuals with a great blend of technical, interpersonal and team skills.
We were located on the east side of the 7th floor of the 'sandcastle' - the Merseyside home of the Royal Group, and part of the skyline of Liverpool. It's a few hundred yards to the left of the Liver Building, and part of the same block as the Daily Post and Echo Building, so it doesn't always appear on skyline photos.
A Culture Shock
My previous job had been as the User Support Group Manager - a team of ten people who served our user-base (soon to be called customers) by answering queries and sorting out any problems that occured on any of our systems. It was non-stop from dawn to dusk, and sometimes from dusk to dawn as well.
When I sat behind my new desk in my new job - EUC Strategy Manager - I waited for something to happen. But nothing did. The phone didn't ring. No one asked me for anything. The team carried on with the projects they were working on.
The shift from support to strategy was not only a culture shock, it required a whole new way of thinking. Instead of responding to events after they had happened, I had to start thinking about what would be happening in 5 years time, and beyond.
A Wide Brief
My starting point was to start researching and thinking about what the strategic business issues would be over the next few years. Discussions with the internal senior management, who were all close to both business and technical strategic issues, were particularly informative - Mike, Derek, David Guerin (Development Manager) and the late Peter McMorran (Technical/Operations Manager).
The organisation was suffering from the same problem as many others. They had a mass of information spread across a number of disparate information systems, and couldn't get at it. The migration of paper systems to computer during the 60s, 70s and early 80s had introduced efficiency savings, but had effectively taken all the information away from business managers: it was locked inside a number of computer systems, none of which could share information with each other.
The challenge, therefore, was to give business information back to the business managers.
Integrated Information Management
To meet this challenge, we (the team - me, Steve, Jacqui, Gary and Yvonne) brainstormed some ideas.
I wrote up the results, and further adapted them, in a document called "Integrated Information Management". I can't recall where that term came from. It may have been a general industry phrase at the time, an expression used by a senior manager, or something that we came up with in the team. The phrase isn't particularly original, however, it was just the title chosen to express what we were trying to achieve.
What was original, however, was the idea that we had. We decided that we wanted to provide a single "system" that connected business managers to information, wherever it was, and without them having to have any technical knowledge.
For example, suppose a business manager wanted information on "shares". Our vision was that he/she:
- calls up the IIM system on his computer screen
- types in the word "shares"
- the IIM system then looks up, in an 'information reference', which systems, and which screens, contain information on shares
- the list of systems is presented to the business manager
- the business manager selects a system/screen to look at
- IIM then navigates the network, application logon and menus to take the business manager to that screen
In the context of today's internet and search engine, this doesn't sound particularly revolutionary. But this vision was produced in 1985.
Developing a Prototype/Roadshow
Having defined the vision, the next stage was to produce a prototype and roadshow. The aim of this was to show business managers what was possible, and thereby get their business backing to further research and development.
At the time, Microsoft Windows was rather clunky. So, to build our prototype we chose to use a windows system from Quarterdeck - in retrospect, an unfortunate decision akin to buying a Betamax video. Also, to automate many of the processes, we used a product called Automator MI, a very powerful tool at the time.
The main focus of our effort was developing a front-end simulation, using the example I've described above (ie searching for shares). Once the simulation was complete, we then turned it into a roadshow to take around the organisation, which was dubbed (after a team vote on a number of options) the "Easy Demo", to show how easy computers would be to use in the future.
Getting Hewlett Packard Involved
The reactions to the roadshow were excellent, and Derek (my immediate boss) was very keen to show it to Hewlett Packard, who were our main computer supplier at the time (of minicomputers and Vectra PCs).
The head of office automation for HP in the UK was Andrew Kingsley. He was based in Pinewood (Berkshire) and we were based in Liverpool (Merseyside). He was not keen to make a long trip simply to see a system demo but, in the end, he came.
His body language at the start of the meeting made it clear that he was there under sufference. But as we started the demo - typing in "shares", getting a list of systems, and then IIM accessing the chosen screen - his demeanour quickly changed. He leaned forward towards the computer Gary was using for the demo, and became engrossed in what we were showing him.
"I want to fly our worldwide office automation manager out here" he said at the end of the demo. "He's got to see this".
New Wave Collaboration
Andrew was true to his word, and the US-based OA manager flew to the UK just to see our prototype system. As a result, they asked us to sign a confidentiality agreement and collaborate with them on the development of their New Wave product.
New Wave was an object-oriented GUI that HP eventually dropped in the 1990s, reputedly because they did not want to sour relationships with Microsoft, who were developing a similar product. At the time it was a highly innovative system and had the potential to revolutionise the way people used computers.
The Reward of Promotion
As a result of this project, I got a promotion and a payrise - roughly £3,000 which, in 1986, was a massive leap. I moved to take the role of IT Manager for Royal Reinsurance and handed over my old role to Martin Morris.
Martin continued to work with HP for some time, and the project eventually appeared in a Hewlett Packard magazine sometime in 1987 or 1988. Unfortunately, I don't have a copy and have not been able to track down a source of back issues.
Martin's emphasis on the project was different to mine. He saw the network side of IIM as being the most important, and dropped the Information Reference. At the time he was probably right - there is no point in having an information reference without being able to access the information through a network. And the most pressing need for the business was simply to get access to the computer systems.
But for me the Information Reference was the key and, if anything, in the project development I had neglected the networking side of the project. Perhaps if Martin and I had been working on the project together over a longer period of time we'd have ended up pushing both aspects forward.
The Information Reference
In some ways, the Information Reference was a less-sophisticated form of a modern search index. The concept was that it contained a list of business words and pointers to systems and screens that were relevant to those words. Visually, the use of it was very similar to a search engine - you typed in a word, and you got a list back (ie a SERP-equivalent).
However, in one respect the Information Reference was far more advanced. It didn't just contain pointers to those systems, it also contained a full automation script to take the user all the way to individual screens on internal application systems as well as external ("internet") sources.
Much of the information needed by a business manager, even today, is not publicly available on the internet or even on an intranet. It is sometimes on password protected workflow systems, internal to the organisation, where you have to navigate through various logon screens and menus. The modern use of cookies and Microsoft Passport can automate some of the logon protocols, but there is still no equivalent for IIM's automated access to internal systems, something that the New Wave Agent promised to achieve.
I don't fully understand the decision to drop the Information Reference from the collaborative project with Hewlett Packard. There may well have been good reasons at the time of which I'm not aware but, if the prototype had been developed, perhaps Hewlett Packard would have been the first company to develop a fully-functional search engine, rather than McGill University.
But, hey, if my Premium Bond numbers had come up, perhaps I'd be a millionaire.
