Skip to main content

Tome Searcher - The world's first AI search agent

In the summer of 1986, I met a guy in a pub. At the time I was between jobs having left mainstream marketing at BMW and trying to make a go of Marketing 4 Motorsport (most of my start-up businesses subsequently contained a numeral... 2nd Byte, 10ACT, Calls2Account, R9 etc). The bloke I met, Tom, was a freelance systems engineer (a sort of software expert) who had formed a company called Tome Associates Ltd with another systems expert, Carl Mattocks (now living in New Jersey). Together with an academic duo - Alina Vickery and her husband Professor Brian Vickery, from the University of London, who were much published in the field of Information Science... or librarianship - they formed a company called Tome Associates. The Vickerys had developed some pioneering software with a small team of computer experts at UCL that used an expert system to formulate and then modify boolean search strategies (sets of words combined with AND, OR and NOT that you could use to search textual databases) from naively stated 'natural language' questions (ie ordinary English). Their early demonstration system was called Plexus about which sadly very little seems to remain on the web (searches for Vickery and Plexus or Tome Searcher will reveal several additional nuggets). Tome was founded to exploit this pioneering work.

The original research was funded by the British Library in 1983 to assess the potential for Expert Systems to replicate the skills of an 'Expert Intermediary' (librarian) to retrieve manageable volumes of relevant results from computerised databases. In 1986, when I met them, Plexus was being transformed into a search tool for online databases of information. The important point is to note the date. End-users (you and me) didn't search online in those days. So if you were an academic, for example, and you wanted to research something, you used a professional librarian to search for you. In those days they used what were called 'host services' which existed in places like California (Dialog), Italy (ESA-IRS) and Brussels (ECHO). Hosts were large computers who offered online access to a wide variety of databases of textual information. Mostly these databases were maintained by institutions such as the IEEE (electronics and computing) and the API (oil and gas). They contained carefully indexed references and summaries of articles, books and published papers in the subjects they covered. Typically you had to order the full text from somewhere else once you had decided from the abstracts that the article/book etc was what you were looking for.

Well Tom (who shortly after I met him decided not to take the risk in staying with the company), Carl, Alina and Brian had decided that the days of expert librarians getting in the way of ordinary people doing their own searches were numbered. They predicted we would all be using Search Engines, a phrase that had yet to enter popular usage, and they decided they needed someone to help them tell the world about it. Me! Probably not a totally daft idea to invite me to join them. I was an engineering graduate, coincidentally also from UCL, so vaguely numerate, and I was a sort of marketing 'expert' (having worked for several major consumer brands at a reasonably senior level). But the biggest challenge they faced was to communicate with naive people who didn't understand computers... so who better to do that, than one of them. Of course 5 years later, I knew a great deal about the fast evolving digital industry - and one of the things I learned was that it's all very well seeking the answers to peoples' problems, the trouble is convincing them they have problems in the first place.

The fact that you've probably never heard of Tome Searcher, the product we developed and tried valiantly to sell, doesn't exactly sing my praises, but we were a tiny lone voice in a digital wilderness, without a marketing budget. Imagine trying to describe Google to people who hadn't heard of the web or even the internet. These were digital dark ages indeed, populated only by geeks using bulletin boards and weird things called usenet groups (still don't know what they were). Browsers had yet to be invented and HTML (the 'hypertext' language of the web) was in its infancy. It would be another 3 years before Tim Berners Lee at CERN would even suggest a world-wide web, let alone experiment with it. Personal computing in those days had just started on expensive IBM machines with green screens operating in DOS. Our first IBM PC cost a staggering £7,000 - a fortune in 1986, let alone now. Our Toshiba laptop (no less) which we humped around the planet to demo our technology, had a 5Mb hard-drive. Massive! Sadly it finally failed in front of 100 senior executives of Phillips in Holland - leaving me to use a flip chart while one of their engineers managed to swap out our hard-drive into one of their own computers - and get it working before I ran out of paper... seat of the pants presentations. Business people these days don't know the terrifying knife-edge we walked every time we gave a presentation, especially a live one which went online. The risks we took!

But we did achieve a great deal of publicity. Here is a selection of press cuttings from the period:

We originally launched a PC product called Tome Searcher at an Online Information Conference in London in 1987. There was huge interest in our demonstrations. Our little stand was always three or four deep with people craning to watch our live demonstrations. Here's the headline in the trade press before the show:

(damn I've got old... Grombler is on the left)



I was also interviewed on the BBC Radio 4 Today programme after a press conference at the Royal Institution in London. I spoke at numerous industry conferences and travelled the world giving presentations at The World Bank, United Nations, IBM, AT&T, Exxon, the British Home Office, European Space Agency, European Commission and countless other governmental and corporate institutions. Drumming up interest about something so new was never a problem. The simple commercial reality was that we were way ahead of our time. End-users (as the information-using public were then called), didn't know there was a world of online information to be searched. Websites hadn't yet been invented and where we calmly talk (or moan) today about broadband being slow at a few megs/sec, in those days the best we got was 240 bits per second (8 bits to a byte) - that's 1/100,000th of the speed I get at home today. (Hurrah, by the way. Fast broadband arrived in our village last week!!). If you imagine that each alphameric character needs 1 to 4 bytes, that's a maximum download rate of only 30 letters a second. You would watch words slowly reveal themselves on a tiny green screen... and we thought WOW! This is cool.

In fact Tome Searcher still is! We had built an intelligent search agent. Instead of matching keywords, as Google and all the other search engines do, our system tried to understand what you were asking about and then deployed synonyms, broader terms, narrower terms and related terms (from thesauri managed by expert systems) in order to help the user achieve a manageable number of the most relevant 'hits'. We used semantic networks to reformat queries more accurately and more completely. Indeed Berners Lee's latest hobby-horse is the Semantic Web - which Tome Searcher pioneered 30 years ago! Google cleverly ranks them according to a) how much someone has paid to be at the top of the list, and b) the popularity of websites which contain the words you've used that others have looked at and talked about. In other words, today's search engines have little idea why you ask your questions or how better you might compile them to improve your chances of success. They just assume you know about your subject and at best correct your spelling. Whereas an intelligent search agent, which emulates a professional librarian (who are now largely extinct), would help the user retrieve the best possible results following a naively created question - naive, that is, with respect to its knowledge of the answers that exist online. For example, if you only search for 'flowers', the possibility is that you won't be shown florists to supply your floral requirements. There's also a chance you'll also be shown the best websites for 'rivers' (flow-ers?). An intelligent search agent like Tome Searcher would have said 'I assume you're also interested in florists, in which case here are the nearest to you that local people say are good, and I've also found a few that have special promotions today. Or are you actually asking about rivers and other things that flow?'

I have often wondered what would have happened to Tome Associates if we'd managed to survive into the age of search engines. We'd have been a decade or two ahead of Google! Perhaps Google would have made us an unfeasibly stupid offer for the rights to our technology (or perhaps we'd have made one for theirs!). We'll never know. My own career benefited massively from the incredible experiences I gained during the infancy of the digital age, despite the disappointment of Tome eventually running out of money in 1991. They say you should always bet on the gladiators with the most scars! Well we were certainly pioneering gladiators, and there were plenty of painful scars. The management of the company fell out painfully as we hurtled towards a penury chasm, and I will always regret the fact that the amazing Vickerys never earned a penny from the massive pioneering steps and advancement of human knowledge they achieved with their technical team for the 5 years Tome struggled along. But it was a huge privilege to have been part of that team who played such an important role in the evolution of the digital world we live in and take for granted today. I could never have achieved all my subsequent commercial successes had I not met Tom in that pub all those years ago and had my imagination not been tickled by the Vickerys.

Brian Vickery's life story, concluding with the demise of Tome Associates, can be read here. But it was his wife Alina (née Gralewska) who was the driving force behind the technology. Alina once told me a story about when at the age of 16 she was the leader of a partisan resistance group in Warsaw during the war. She told me about an occasion when she helped a Jewish family escape from the ghetto through a tunnel she was guarding. Many years later she visited Israel and was stopped by a customs guard who asked her if she'd lived in Warsaw. When she said yes, he broke down in tears and told her he recognised her as the girl who had saved his family. Alina was apparently feted by the press and government as a hero and stayed on in Israel to become chief librarian for the Israeli Atomic Energy Commission, no less. I can't find her name in the list of Righteous Among Nations compiled by Israel to recognise people who risked their lives to help Jews in the Holocaust, but I have no doubt she is there somewhere. She died in 2001 and there's very little about her on the web, sadly. So here's hoping this blog can contribute in some way to the preservation of her memory and as a salute to her work and imagination.

I have a small library of articles and clippings from the Tome era which I've scanned. Sadly very little seems to have survived online today about our work. If you are aware of any additional online information about Tome and our pioneering work, please include a link in your much appreciated comment below.

Comments

Popular posts from this blog

Phillips screws - yes I'm angry about them too

Don't get me wrong. They're a brilliant invention to assist automation and prevent screwdrivers from slipping off screw heads - damaging furniture, paintwork and fingers in the process. Interestingly they weren't invented by Mr Phillips at all, but by a John P Thompson who sold Mr P the idea after failing to commercialise it. Mr P, on the otherhand, quickly succeeded where Mr T had failed. Incredible isn't it. You don't just need a good idea, you need a great salesman and, more importantly, perfect timing to make a success out of something new. Actually, it would seem, he did two clever things (apart from buying the rights). He gave the invention to GM to trial. No-brainer #1. After it was adopted by the great GM, instead of trying to become their sole supplier of Phillips screws, he sold licenses to every other screw manufacturer in the world. A little of a lot is worth a great deal more than a lot of a little + vulnerability (watch out Apple!). My gromble is abo

Introducing Product Relationship Management - it's what customers want.

Most businesses these days have Customer Relationship Management (CRM) systems which store and process vasts amounts of information about us. They use this information to generate communications, amongst other things, which target us to buy their products and services. CRM is all about how a business relates to its customers: Past (keeping them loyal through aftersales and service), Present (helping them buy through bricks and clicks channels) and Future (prospecting). Most businesses will at some stage have declared themselves 'customer-centric'. They will probably have drawn diagrams on whiteboards that look something like these: But there's a problem with this whole approach of keeping the customer at the centre of your world and the focal point for everything you do. Is it what the customer wants ? Of course companies who ignore their customers eventually go out of business. And those who treat their customers well, tend to thrive. But is it really in the best inte

The Secrets of Hacker Golf

Social media is awash with professional golfers selling video training courses to help you perfect your swing, gain 50 yards on your drive and cut your handicap. They might help a few desperate souls, but the rest of us hackers already know everything we need to complete a round of golf without worrying the handicap committee or appearing on a competition winner's list. What those pros don't realise is that for us hacking golfers who very occasionally hit shots that if you hadn't seen how they were hit, end up where the pros might have put them, we already know everything we need to know - and more. Unlike pros who know how to time the perfect swing in order to caress a ball 350 yards down the centre of a fairway, we hackers need to assemble a far wider set of skills and know-how to complete 18 holes, about which pros have no comprehension, need, or desire to learn. Here are some of them: Never select your shot until after you've hit it. A variation on this is to alway